68 results listed
A time series is a sequence collected at consecutive equally spaced points in time. The
basic idea behind the time series forecasting is the use of a model to estimate future values based
on previously observed ones. Traditionally, statistical methods are used to forecasting time series
however, Machine Learning (ML) algorithms have been also proposed as alternatives to statistical
methods in past decades. In this paper, we evaluate forecasting performance of different ML
algorithms and statistical methods on Turkey automobile sales. Recently, various of work has
claimed that traditional statistical methods dominate the ML solutions in terms of time series
forecasting. This study discusses different aspects of ML and statistical methods and compare their
performance on different time series.
International Data Science & Engineering Symposium
IDSES
Anıl Özdemir
Furkan COŞKUN
Selim BALCISOY
In this study, a nonparametric new test is proposed to test the hypothesis of equality of
locations against umbrella alternatives. The Shan test for ordered alternatives is adapted to the
umbrella alternatives. This test can be considered as an extension of the sign test and the Wilcoxon
signed rank test. By a comprehensive simulation study, the proposed test is compared with the
Mack-Wolfe and Hettmansperger and Norton tests in terms of type I error rate and power. The
simulation results showed that all tests ensured the Bradley's robustness criteria for type I error
rate. The power comparison results indicated that the proposed test gives better results than the
other tests.
International Data Science & Engineering Symposium
IDSES
Bülent ALTUNKAYNAK
Hamza GAMGAM
Merve BAĞÇACI
The main objective of the present research paper is to give an exploratory vision on a
new form of examining hotels’ performance of the Portuguese hotels’ market. This approach is
based on the experience of a consultant in this area and is made in terms of revenue and occupancy,
crossing then operational data with financial results. This analysis focuses on a recent period, since
2010 to 2017, involving the years since before and until after the big boom of tourism in Portugal.
This paper exposes available data about hotels operational results and compares these results with
hotels financial results. In terms of the evaluation of a hotel’s performance, the revenue per
available room (RevPAR) is regarded by the tourism industry and INE (the Portuguese National
Statistics Institute) as the most important measure, once it compares hotels among themselves and
also compares regions. An important aspect of this study is highlighting the limitations of
RevPAR.
International Data Science & Engineering Symposium
IDSES
Nuno Mendes FREITAS
José António FILIPE
This research involves Big Data and its definition. Also, it covers general
visualization of the problems in Turkish legal system and deal with a real case to create a
methodology for solving the problems in filing an indictment process with the help of Big Data.
The goal is to show that how can the finishing time of the process of filing indictments
significantly reducing when Big Data technology integrated to the process. This has been done
by creating a decision support system that can apply to all the cases in the phase of linking
evidences and filling the indictment. This research will provide valuable information regarding
the methodology of applying Big Data and address the role of Big Data in solving problems of
ineffective and complex organizations.
International Data Science & Engineering Symposium
IDSES
Zeynep ÇAĞLAR
Various kinds of medical treatment methods may be used to cure common types of
diseases. Experience based predictions are done to choose a treatment method among the choices
of cures to get better results for the patient. This condition sometimes may continue with trying
another treatment method unless a satisfactory result is reached and changing the treatment method
of the cure process is not a desired course for time and health. This study presents a confident way
to choose the treatment method for wart disease by using feedforward neural network. The study
uses two types of datasets, one for cryotherapy and other for immunotherapy treatment methods.
It was observed from the experimental results that, feedforward neural network achieved 94.4%
success and 85.6% success for cryotherapy and immunotherapy datasets, respectively. The results
are remarkable for both doctors and patients.
International Data Science & Engineering Symposium
IDSES
Ahmet ÇİFCİ
Mehmet ŞİMŞİR
Manual categorization of applications in software repositories such as SourceForge is
often time-consuming and error-prone. Automation of this process not only simplifies the daily
task of administrators but also helps project owners to add their projects into the corresponding
subcategory of the repository without any delay. In this study, we propose a cloudbased application
categorization system that applies machine learning algorithms to support the classification of
applications. The categorization system has a web-based client application to parse, process, and
submit the project source code, a web service which automatically performs classification of
applications into domain categories, and a cloud-computing platform which hosts the
categorization service. Several multi-class classification algorithms have been adopted including,
Artificial Neural Networks, Logistic Regression, Decision Jungle, and Decision Forest algorithms
to validate the effectiveness of the system in multiple case studies. The case studies were
performed on three public datasets generated based on 3286 Java applications of SourceForge
repository. Our study shows that the highest accuracy was achieved with Artificial Neural
Networks (ANN) algorithm. The resulting prediction model has been transformed into a web
service and then, deployed on the Azure cloud platform.
International Data Science & Engineering Symposium
IDSES
Çağatay ÇATAL
Besme ELNACCAR
Özge ÇOLAKOĞLU
Bedir TEKİNERDOĞAN1
Digital farming as we see it has the potential to revolutionize agriculture, and bring
significant benefits for farmers and the society overall, as we need new ways to grow more food
more sustainably. In this study, Sustainable Farming and Digital Age, Certification as a Model for
a Sound Precision Agriculture and Impacts of Precision Agriculture and Certification Needs are
described.
International Data Science & Engineering Symposium
IDSES
Massimo CANALICCHIO
It is often not easy to obtain results from complex processes multi variables. Additional
techniques and methods are needed to guide. In this study, after the detecting the out of control
and under control samples with Hotelling T2 control chart in a multivariate manufacturing process
then machine learning algorithms was used to predict the quality of future samples. Four machine
learning algorithms were trained and tested by shifts of different magnitude from the process
average. The performances of the algorithms were compared according to the accuracy and error
rates of the predictions and the most appropriate one was chosen as Multilayer Perceptron.
International Data Science & Engineering Symposium
IDSES
Deniz DEMİRCİOĞLU DİREN
Semra BORAN
Seda Hatice GÖKLER
Turkey and the world a lot of earthquakes that have occurred and will continue to
come. Due to the earthquake, many people have caused their lives and damage to their shelters.
Starting from a young age earthquake in Turkey to create awareness in every age group,
observed and loss of life and property damage in the earthquake is one of the lowest levels to
minimize the road. Therefore, it is very important to be able to use techniques and systems that
can analyze a large number of data sets. The process of converting these raw data into
information or meaning can be done by data mining. The aim of this study is to investigate the
effects of the education given to the secondary and high school students on the students and the
earthquake awareness. The research sample consisted of 14 middle school and 11 high school
students, who were randomly selected in Karabük and districts. The questionnaire developed
by the researcher was used as the data collection tool. In the scope of the research, 1165 students
were given questionnaires before and after 80 minutes of training. In this study, the seismic
awareness of the students' education on the students before and after the training was
investigated by clustering analysis.
International Data Science & Engineering Symposium
IDSES
Ömer KIVRAK
Filiz ERSÖZ
Risk is any possibility that affects the achievement of the intended objectives. In this
study, it was investigated which provinces did not contain any risks. In the calculation of risk
values of provinces; unemployment and traffic accidents were taken into consideration. Turkey
Statistical Institute (TURKSTAT), the General Directorate of Security, the Social Security
Institution (SSI) and Turkey Business Association (TBA) risk values obtained from benefiting
from the data of the years 2013 to 2017 were calculated. Risk values are evaluated between 1 and
5. The green-colored provinces are risk-free, while the red-colored region is cons idered to be of
high risk. Fine-Kinney method was used for risk analysis.
International Data Science & Engineering Symposium
IDSES
Taner ERSÖZ
Betül EBESOY
Banks are one of the most important actors of the financial system. Banks can
perform their services through alternative distribution channels such as branches, ATMs,
internet banking and telephone banking. Banks, which are affected by economic developments
and who are both financial market actors and profit-making enterprises and employing
employment, provide the most efficient services through their branches. The choice of the
banks' location is very important in terms of bank success. One of the factors affecting the
choice of establishment location of banks is the population of the place. The closest selection
criteria are the GDP per capita in the region and the activities of competing banks. It is important
to choose the place of operation of the banks as well as the widespread branch network in which
the customers will work. In the competitive environment, especially corporate customers work
with more than one bank. It is important that the banks where they work for the enterprises that
have active or widespread networks in the field have branches together in the same settlement.
The aim of this study is; a total of 960 settlements with Turkey's 81 provinces, the districts and
towns, bank examination for the presence of the branches of activity and which banks that
operate together on the same settlement "data mining" One of the suitable ones association rules
"Apriori Algorithm" and is to be determined. For this purpose, the "The Banks Association of
Turkey" in the number of audits and the existence of settlements in which they operate
according to their location by the banks operating in Turkey "IBM SPSS Modeler" has been
analyzed with the program. According to the results obtained, the banks operating together in
the same settlement area may determine the bank preferences of the financial managers.
International Data Science & Engineering Symposium
IDSES
Süleyman DÜNDAR
Seliha Seçil BAYRAM
Crime analysis has importance for the detection of crime regions, the prediction of
crimes before processing and the security forces to take necessary measures. By using
biclustering methods to detect crime regions, simultaneous clustering of the types of crimes and
regions where crime is committed to producing more comprehensive results than traditional
clustering methods. In this study, CC and Xmotif algorithms of biclustering methods were
applied to the real data set in order to detect the crime regions. “Crimes in Boston” data set was
used in real data set application. In order to measure the efficiency of the biclusters, the
performance of the algorithms was compared with Chia and Karuturi bicluster score (CCPS).
The results were obtained by using Matlab functions and it was observed that results of the CC
algorithm were better compared to Xmotif algorithm.
International Data Science & Engineering Symposium
IDSES
Nazan SARI
Sümeyye Gizem ÇAKAR
Olcay EYDEMİR
İbrahim ÇİL
A container is the one of the main components of transportation systems. Allocating
items into limited spaces, is a kind of combinatorial optimization problem and container loading
problems is a branch of knapsack problems, in which a set of items are loaded into capacitated
domains. Heuristic approaches are mostly applied to solve knapsack problems due to the
problem complexity. Artificial Bee Colony Algorithm and Genetic Algorithm are successful
for solving object placement issues. The first one can obtain sufficient results as well as the
second one. In this study, the performance of Artificial Bee Colony Algorithm, which is applied
on CLP rarely, is compared with Genetic Algorithm, which is applied on CLP widely, to see
the capability of proposed ABC algorithm for further studies.
International Data Science & Engineering Symposium
IDSES
Tuğrul BAYRAKTAR
Filiz ERSÖZ
Cemalettin KUBAT
Value-added is an important term which indicates the efficiency level of economic
activities. Valueadded term describes the difference between the value of produced goods and the
total cost of production. Some areas related to value-added manufacturing are pharmaceuticals,
automotive, computer and communications equipment manufacturing, aviation, etc. To develop a
value-added product some inputs are needed. Some of these inputs are energy, labor force and
research and development activities. The main aim of this study is to present a benchmarking
analysis of 29 OECD countries in views of value-added manufacturing values. To do so, an output
oriented BCC model is used to evaluate efficiency of countries and obtained results of the analysis
provide some improvement ways for countries and their position against the other countries.
International Data Science & Engineering Symposium
IDSES
Billur ECER
Ahmet AKTAŞ
Mehmet KABAK
Long Short Term Memory (LSTM) is one of the deepest learning methods capable of
learning along a chain. The method has a chain of modules able to repeating information and
transferring it to the next module. Due to this feature, it is a convenient method for data sets
consisting of time-dependent information such as finance. Bitcoin, using blockchain technology,
has become one of the most popular cryptocurrencies today. Bitcoin data is a time series. In this
study, price estimation model is proposed by using Long-Short Term Memory method for a Bitcoin
price estimation for multivariate time series consisting of opening price, closing price, highest
price, lowest price, Bitcoin volume, Purchasing volüme and weighted price variables. In addition,
the application has been developed in Python programming language.
International Data Science & Engineering Symposium
IDSES
Ali Osman ÇIBIKDİKEN
Ebru Şeyma KARAKOYUN
Bitcoin is the most popular cryptocurrency in the market. Satoshi Nakamoto has
created the Bitcoin in 2009. Bitcoin movements are a widely discussed topic nowadays. On the
other hand, Today machine learning is effectively deployed in an extensive variety of fields
from natural language processing to image processing, from medical applications to activity
recognition. In this research, we collected and used data about the market value of Bitcoin.
Also, Bitcoin is briefly described in this study. Artificial neural networks and time series
techniques used to estimate the market value of Bitcoin. Finally, The paper concludes with
critical considerations of recent developments and some recommendations for future
researches.
International Data Science & Engineering Symposium
IDSES
Derya SAĞ
Ufuk CEBECİ
Human-Computer Speech is gaining momentum as a technique of computer
interaction. A chatbot is a software, which can “chat" with a human user in natural language such
as English. Conversational artificial intelligence technology (CAI) enables learners to engage in
spoken conversations with the non-player characters. In this study, SAP's digital assistant can be
used to run business processes using SAP Conversational Artificial Intelligence. The benefits of
the digital assistant (chatBot) will be discussed. An example of the request for personnel’s leave
request, which is one of the areas of use, will be explained.
International Data Science & Engineering Symposium
IDSES
Metehan KOCAOĞLU
Kağan ÖZDEMİR
Alper KARABULUT
Rüştü Orkun KORKMAZ
Ufuk CEBECİ
With the development of technology from past to present, the types, properties and
product range of the materials used in the buildings are quite developed. Therefore, the effects of
climate, environmental conditions and energy consumption cannot be ignored for selecting these
materials used in the buildings. Usage of materials with the same characteristics for buildings to
be built on different climate may lead to adverse effects about energy-saving and green gasses.
Furthermore, the use of the same materials may not be a proper approach even in buildings with a
different purpose. In this study, forecasting of energy consumption and CO2 emission is analyzed
by utilizing artificial neural network structure according to different climate criteria and material
characteristics for public buildings built in recent years. The Effect levels to energy consumption
and CO2 emission of the building materials and the climate criteria are determined for buildings
serving the same using purpose in different climate characteristics. For the study, different pilot
regions where the public buildings are located are chosen according to climatic characteristics and
five different building materials used in these public buildings are taken into account. When the
results compare according to CO2 emission and energy consumption, it was observed that the
conditions which obtain most efficient results are different
International Data Science & Engineering Symposium
IDSES
Salih HİMMETOĞLU
Yılmaz DELİCE
Emel KIZILKAYA AYDOĞAN
The social communities created by social media, where people have shown great
interest, have led to the analysis of social network. In this sense, many techniques have been
developed and these techniques have been practised in various fields. By means of the softwares,
developed for investigation of complex network analysis, detailed surveys and research can be
made about social media. In this study, a social network analysis has conducted on the social
media. The study was carried out through the Nodexl program for two different social communities
via Twitter to draw the network graphs and to compare the network analysis results by using ,
“#EytHepBirlikteAnkaraya and #BirlikOlFenerbahçe” According to the results obtained from the
analysis, it is observed that although the EYT members are a more recent social group than
Fenerbahçe supporters, the ties in their network are stronger and the network density of EYT
members is four times more than in Fenerbahçe supporters. In addition, although the hashtags are
addressing different topics, the value of the network characteristics such as clustering and
centralization were found to be similar to each other.
International Data Science & Engineering Symposium
IDSES
Adem AKSAN
Ayşe OĞUZLAR
In the modern era, enterprises are facing a variety of difficulties because of today’s
emerging technologies. One of the numerous difficulties for a business is to fulfill its employees
in order to adapt to the consistently changing business processes and to make progress and stay in
competition. In order to build proficiency, viability, efficiency and occupation responsibility of
employee, the business must fulfill the requirements of its employee by giving great working
conditions. The target of this paper is to dissect the effect of workplace on employee work
fulfillment. This paper may profit society by urging individuals to contribute more to their
occupations and may help them in their daily work life. Consequently, it is fundamental for an
association to support their employee to snap down for accomplishing the hierarchical objectives
and goals. The investigation and the item are changing the advanced endeavor, expanding
representative commitment over the whole workforce, including forefront, field, remote and
outside laborers to enhance execution, efficiency and unwaveringness. Our stage is making
ground-breaking employee encounters, where every single representative feels some portion of an
option that is more noteworthy than themselves, are glad to be a piece of your image and effectively
advance the positive characteristics of the association.
International Data Science & Engineering Symposium
IDSES
Batuhan Burak ERSÖZLÜ
Ufuk CEBECİ
Şahen TOKATLIOĞLU
Sacrifice stands for the production of goods and function, constitutes the costs of
enterprises. Cost is also defined as the provision of consumed goods and functions by a
production enterprise. Accuracy of enterprise activity analysis is very important in order to
make appropriate decisions in enterprises. The consistency of the results ensures correct
decision-making; provides right marketing and competitive advantage. Various elements are
effective in the process of product costing. The items on the basis of product are examined one
by one and the analysis is carried out to obtain the unit costs that reflect the reality. The aim of
this study is to investigate the factors affecting the costs and to estimate the cost in the integrated
system, with data mining classifying models in the process of billet production, in an A
enterprise for the Iron and Steel sector. It is targeted to compare obtained estimation results
with the costs presented inside the enterprise.
International Data Science & Engineering Symposium
IDSES
Gizem KAPANŞAHİN
Filiz ERSÖZ
In today's urban life, most of the people's time is spent at home or work office. For this
reason, design and selecting of building materials that control heating (H), ventilation (V) and air
conditioning (AC) are essential. The building materials called, in short HVAC, must
simultaneously provide high comfort, low cost and high energy productivity. Furthermore, HVAC
must be appropriately designed to prevent adverse effects on the environment and climate. In this
study, nine different HVAC systems were examined according to nine different criteria over cost,
pollution, comfort and energy which are considered as four main factors in the selection of HVAC
systems. Since some of the criteria for HVAC systems are described as linguistic, it is not possible
to evaluate the systems with traditional methods using crisp values. Therefore, we propose
generalized intuitionistic fuzzy (IF) - rough set models which are a new and flexible method. IFneighborhoods
are formed by using IF-implicator and IF-t norms, and upper and lower
approximations in rough set theory are calculated according to the neighborhoods. Covering-based
generalized IF rough set models are generated by using the approximations and IF-TOPSIS
method. According to the obtained results, we can see that the proposed method is an appropriate
decision-making method which considers the uncertainties in the linguistic expressions for
selecting the most suitable HVAC system.
International Data Science & Engineering Symposium
IDSES
Salih HİMMETOĞLU
Emel KIZILKAYA AYDOĞAN
Yılmaz DELİCE
One of the most important steps in establishing a successful business is to do accurate
data analysis. The correct analysis of the data and the correct information as a result of the analysis
reveal the wishes, feelings, emotions, and demands of the users. With data analysis useful
information should be discovered, those who are useless should be cleaned and what should have
done in the next process. Several types of data analysis methods can be done based on the data
collected from Kansei survey. These analyses play an important role in the process of Kansei
Engineering. There are several types of statistical analysis that are developed to use in Kansei
studies such as variance analysis, linear regression analysis, flow data analysis, principal
component analysis, quantification theory I, factor analysis, cluster analysis, rough set theory.
The purpose of data analysis is to synthesis statistical data or Kansei words with the product
properties and therefore to be applied in the design context.
Kansei engineering is a method used to convert a customer’s ambiguous imagine product
into detail design. Kansei Engineering starts from decision of strategy as design domain as well
as target. It is collected the Kansei words related to the product domain. The word Kansei, which
is used in design and other research areas. It means the feeling of beauty and pleasant emotions
reflected by any object and its desire in Japanese. Kansei words form the basis of Kansei
engineering. In a way, Kansei engineering is a product development method which can measure a
customer’s feeling, values and affective needs and translate them into concrete product solutions.
Since 1980’s Kansei Engineering has expanded greatly and become a significant discipline
both in the industrial and the academic world. Furthermore, Kansei Engineering developed as an
efficient research discipline, providing many innovations and market success in the industrial
world. It is founded by Mitsou Nagamachi, a professor at Hiroshima University. He is considered
to be the father of Kansei Engineering. The term Kansei Engineering itself was used the first time
in 1986 by Yamanota, president of Mazda Automotive Corporation at Michigan University.
The main aim of this study is to explain Kansei engineering and model which is rarely seen
in Turkish literature and to reveal its relationship with data analysis. Besides, the future of
importance of data, Big data, data analysis and Kansei Engineering will be discussed.
International Data Science & Engineering Symposium
IDSES
Mustafa Umut Öztürk
Ahmet Öztürk
Nowadays, as a result of the fast advancement of technology, large and complex
databases are created. Since it may be hard to process the created complex data, there is a need for
software programs. The purpose of this study is to create a substructure for data mining to be used
in the health sector and to give a quick and different perspective in deciding to reach the desired
data. With this purpose, based on the data from the operating room of a private hospital, the data
mining method was used in this study.
International Data Science & Engineering Symposium
IDSES
Duygu KARABULUT
Taner ERSÖZ
In this paper, a deep learning model has been developed to detect whether
malware/spyware leaks data to command and control servers and a new dataset has been
obtained from real-time environment for test of the model. In addition, effect of the size of the
data set and hyperparameters such as the number of layers of the deep neural network on the
success rate have been investigated. In this study, real-time data for harmful and normal İnternet
traffic have been obtained in the application layer and 100 features have been selected. The
developed deep learning model has been applied to 16,000 sample obtained from real-time
Internet traffic. From the experimental results, accuracy rates of 90% to 94% were obtained
with various number of samples and various number of layers in the deep learning model. It has
been seen from the experimental results that increase the number of samples increases the
accuracy rate. As well as, it has been seen that as increase the number of layers in the deep
neural network the accuracy rate increased first, further increase the hidden layers did not affect
the success rate. In this study, more distinctive and important features have been investigated
than others in the literature and the results have been tested.
International Data Science & Engineering Symposium
IDSES
Emrullah ERGİNAY
M. Ali AKÇAYOL
With the rapid development of technology, competition in the markets is increasing.
In the conditions of rising competition, it is inevitable for companies to make predictions for
the future. The accuracy of the forecasts will also facilitate the decision-making of companies
in the future. Demand forecasts are very important for every company considering these
conditions. Demand forecasts play an active role in tactical and strategic decisions taken on the
administrative side of companies to achieve their near or far-term goals, bringing the company's
assets closer to optimal and thus ensuring the link between the targets set and the operations
implemented. According to the data obtained from the firm of Antik Atölye operating in Düzce,
the regression analysis method and weighted moving average method were compared. For both
methods, estimation modeling was established and implemented. As a result for the Antik
Atölye company operating in the wood industry regression analysis method was more effective
than weighted moving average method.
International Data Science & Engineering Symposium
IDSES
Sercan ODABAŞI
Taner ERSÖZ
Productivity is the relationship between the output produced by a production or
service system and the input used to create this output. High efficiency is to produce more with
the same amount of resources or more output with the same input. However, it is accepted that
efficiency and working life quality are closely linked. The efficiency of the enterprises is
important for the society and the country in which they carry out their activities as well as for
themselves. While the productivity of the enterprises is reflected in the decrease in the costs,
increase in the profit, the efficiency of the enterprises has reflections on the new investments,
increased added value, the welfare of the society and employment in terms of society and
country. It is also important to measure the efficiency that is so important in order to make it
sustainable in enterprises. The productivity measured continuously in enterprises will enable
managers to make correct and rational decisions, to see problems on time and to produce
solutions. The purpose of this study is to determine the factors affecting the productivity of the
employees in the steel industry. The study was examined under five titles. These are:
“Demoraphic Information”, “Economic Factors”, “Physical and Ergonomic Factors”, “Psycho-
Social Factors” and “Risk Factors”. According to the findings of the study, the effect of factors
on the employees will be determined and improved findings will be presented to the
management of the enterprise.
International Data Science & Engineering Symposium
IDSES
Taner ERSÖZ
Filiz ERSÖZ
Nurdan KALELİ
Burçin ATICI
The studies related to the production industry are limited in the world and in our country.
Especially in iron and steel sector, quality levels of different types of products need to be
monitored. Iron and steel products obtained from the studies have prolonged their use and price
and sales superiority has been achieved. At the same time, the market value of the products
increases and there is a minimum loss of product. Therefore, studies in this field should be focused
on. On the basis of quality, instead of debugging errors is the approach of not making mistakes.
Instead of using your earnings as a philosophy, we should adopt an understanding of gaining from
our losses. Understanding the importance of quality work and improvements, the primary purpose
of enterprises is to support quality production by preventing or reducing errors in production. Data
mining has started to be used effectively in enterprises. Data mining involves the process of
selecting, organizing and modeling the most necessary data for business executives. At this point,
it is possible to define data mining as a set of techniques and concepts that produce new
information for decision-making processes. In this study, firstly the VM process is defined and
then the VM studies which are selected from the literature covering 2010-2018 and applied to
certain quality improvement problems in the manufacturing sector are evaluated. The definition of
process and product quality, estimation of quality, classification of quality and optimization of
quality parameters are discussed. In addition, the application of decision trees, one of the most
widely used and effective VM techniques, in order to determine the variables and levels that cause
production errors in an industrial organization is also included.
International Data Science & Engineering Symposium
IDSES
İsmail Burak AKINCI
Filiz ERSÖZ
In general, it is seen that the fire data is presented as the amount of the area burned
and the number of fires in time. Flammable material loads in forests, behavior patterns and
models of flammable materials according to climatic conditions etc. works are carried out
rapidly. All these studies aim to manage the process after the fire. There is a need to work on
measures to be taken in order to prevent the occurrence of fire. For this purpose; socioeconomic
factors that cause fires should be determined in regions where forest fire is common.
The elimination of these factors will minimize the occurrence of forest fires. In Turkey;
considering that 89% of the forest fires are caused by human beings, the importance of socioeconomic
studies in these regions is increasing. In the studies to determine the socio-economic
factors that cause forest fires; In some period, some studies such as multiple regression,
correlation, factor analysis etc. were conducted among some socio-economic data determined
according to the conditions of the region and / or fire numbers in the given period. In all these
studies a time section / part was taken into consideration. Materials and Methods; First time in
Turkey with this work; the relationship between the number of forest areas and forest fire
counts and the socio-economic variables determined by considering a certain period of time
was analyzed together with time and space. In the forest fires in the years 1980-1990-2000 in
Antalya Forest Regional Directorate for twelve governmental forest enterprises, the
relationship between the number of forest areas and fire numbers and 25 socio-economic
variables were determined. These data were analyzed by panel data analysis or Time Series
Cross Section Regression (TSCSREG) analysis method.
Variables with no effect in analysis and at the same time the variables / criteria that were
derived from each other were taken into consideration and these criteria were eliminated by
multiple linkage analysis (multiple linear analysis, multiple collinearity analysis) and reduced to 12 variables. Analysis of burned forest areas and selected socio-economic variables; It was
tested by Fuller and Battese Methods in the scope of TSCSREG Analysis.
Conclusion: In the analysis of the amount of burned forest areas and selected socio-economic
variables; A strong relationship between forest fires and selected socio-economic variables
shows that the value of R2 is 90.18%. The value of R2 is 70.19%. It shows that there is a
relationship between the numbers of fires and the socio-economic variables selected.
International Data Science & Engineering Symposium
IDSES
Ufuk COŞGUN
Tourism in the natural world has an important place. Tourism revenues have a
significant share in the development and development of countries. The economic benefit to be
achieved; it is possible to maximize the natural, historical and cultural beauties. In this study,
which was conducted to determine the satisfaction of tourists, survey was applied to 169
tourists. As a result of analysis; It was concluded that variables such as age, gender, nationality
and income did not affect tourist satisfaction, variables related to purchased products, shops
and sellers had a significant relationship between tourist satisfaction.
International Data Science & Engineering Symposium
IDSES
Fatma ATEŞ
Nuray TÜRKER
Filiz ERSÖZ
Job evaluation; In order to create input for performance evaluation and wage
management in businesses, the skills of job, responsibility, job conditions etc. and formally and
systematically. Business valuation deals with the importance of the work and the added value
it provides to the business rather than the employee. This study was carried out in the iron and
steel industry. Firstly, two-way comparison matrices were formed by Analytical Hierarchy
Process which is one of the Multi Criteria Decision Making (MCDM) methods for the main
factors and sub-criteria to be used when conducting business evaluation. Then, the paired
comparison matrices formed by Fuzzy Analytic Hierarchy Process were reconstructed with
triangular fuzzy numbers and the weights of main factors and sub-criteria were calculated by
Chang Ground Analysis method.
International Data Science & Engineering Symposium
IDSES
Ataberk OLCAY
Muharrem DÜGENCİ
Mümtaz İPEK
Many indicators are used in the analysis of the financial status of companies. Indicators
are usually presented quarterly. However, some of the financial indicators of some companies are
not given data. In this study, it is examined whether there will be other indicators that Show the
same character instead of the data which is not present when the companies are evaluated as a
whole. A hierarchical clustering approach was used to evaluate whether the financial indicators of
10 randomly selected companies traded on the NASDAQ stock exchange to have similar
characteristics using distance methods. According to the results obtained, indicators with similar
characteristics are listed.
International Data Science & Engineering Symposium
IDSES
Ebru Şeyma KARAKOYUN
Bu çalışmada akademisyenlerin yayınları kullanılarak yayından yazara birliktelik
analizi yapılmış ve akademik çalışma grupları oluşturulmaya çalışılmıştır. Yayınların başlık,
anahtar kelime, özetlerinde geçen kelimeler morfolojik olarak köklerine ayrılmış, metinsel
temizleme yapılmış ve Empati-Çıkar yöntemi kullanılarak benzerlik/yakınlık katsayıları
hesaplanmıştır. Oluşan benzerlik/yakınlık matrisi kullanılarak yayında katkısı bulunan yazarlar
yakınlıklarına göre çalışma gruplarına dahil edilmeye çalışılmıştır.
International Data Science & Engineering Symposium
IDSES
Ömer Faruk ACAR
Burhan Selcuk
The wastewater treatment in the textile industry is of special importance due to the
intensive use of chemicals and dyes. However, wastewater treatment plants have impacts on
environmental and these environmental impacts should also be evaluated and minimized.
Neutralization is one of the processes with the huge chemical consumption in the wastewater
treatment plant. Therefore, the chemical alternatives used in the neutralization process should
be compared in terms of their environmental impacts. In this study, the performances of carbon
dioxide and sulfuric acid as two alternative chemicals used in the neutralization process applied
in a textile factory wastewater treatment plant are compared using the life cycle approach. The
neutralization process using carbon dioxide yielded better results in the categories of abiotic
depletion, fossil fuels, ozone layer depletion (ODP), fresh aquatic ecotoxicity, marine aquatic
ecotoxicity, terrestrial ecotoxicity, photochemical oxidation, acidification, and eutrophication.
International Data Science & Engineering Symposium
IDSES
Fatma Şener FİDAN
Emel KIZILKAYA AYDOĞAN
Niğmet UZAL
Blood is a vital product that is needed by thousands of people every day due to diseases,
surgeries or injuries. Blood banks should accurately determine the amount of blood they should
have in their stock to meet blood needs. Therefore, having less blood than necessary in hospitals
creates important problems such as not meet need for blood and loss of life. On the other hand,
storing large amounts of blood causes deteriorating the blood and causes stock out in other
hospitals. The aim of this study is to determine the criteria affecting blood demand and to forecast
the blood demand by the machine learning algorithm Adaptive Network Based Fuzzy Inference
System (ANFIS) method. However, since the number of impact criteria is high, principal
component analysis (PCA) method has been used in order to decrease criteria and eliminate the
dependencies between the criteria. The developed hybrid method was applied in a regional blood
center.
International Data Science & Engineering Symposium
IDSES
Seda Hatice GÖKLER
Semra BORAN
Industry 4.0 was first announced to the public by the German Federal Government
at the 2011 Hannover Fair. The concept of Industry 4.0 is taken into consideration especially in
our country, considering the rapidly developing countries. In this study the work of leading
organizations in Turkey is being examined and the approaches to Turkey’s Industry 4.0 are
reported. In the study, SWOT Analyzes are compiled from the results of civil society
organizations reports, academic publications, public institutions report, consulting companies
and related books. As a result of this review, a single SWOT analysis consisting of 9 items in
each item is presented with the text mining method. The SWOT Analysis obtained is linearly
scored between 1 and 9 points given from the sources using the R'WOT Analysis method.
According to the results of the R'WOT Analysis, our country's approach to Industry 4.0 is
considered as an opportunity of 31%.
International Data Science & Engineering Symposium
IDSES
Gizem ACAR
İlker ÖNALAN
This article sheds light on how games applications increased its popularity using social
network platform such as Facebook. What is more, this piece of writing reflects rapid evolution of
particular games such as “Farmville”, “Pet society”, which using Facebook API. In addition, the
article provides information what technologies, policies, protection measures; Facebook takes to
protect users’ personal information, “OAuth 2.0 protocol”, in particular. Additionally, the article
provides information concerning benefits Facebook and its users get from using particular games
apps, challenges they are facing. Finally, the article gives some recommendations how Facebook
and its followers can cope with these challenges.
International Data Science & Engineering Symposium
IDSES
Babaev Vladimir Yandashevich
A popular risk measure, Conditional Value-at-Risk (CVaR), is called Expected
Shortfall (ES) in financial applications. The paper developed algorithms for implementation of
linear regression for estimating CVaR as a function of some factors. Such regression is called
CVaR (Superquantile) regression. The main statement of the paper: CVaR linear regression can
be reduced to minimizing the Rockafellar Error function with linear programming. The theoretical
basis for the analysis is established with the Quadrangle Theory of risk functions. We derived
relationships between elements of CVaR Quadrangle and Mixed-Quantile Quadrangle for discrete
distributions with equally probable atoms. The Deviation in CVaR Quadrangle is an integral. We
presented two equivalent variants of discretization of this integral, which resulted in two sets of
parameters for the Mixed-Quantile Quadrangle. For the first set of parameters, the minimization
of Error from CVaR Quadrangle is equivalent to the minimization of Rockafellar Error from the
Mixed-Quantile Quadrangle. Alternatively, a two-stage procedure based on Decomposition
Theorem can be used for CVaR linear regression with both sets of parameters. This procedure is
valid because the Deviation in the Mixed-Quantile Quadrangle (called Mixed CVaR Deviation)
coincides with the Deviation in CVaR Quadrangle for the both sets of parameters. We illustrated
theoretical results with a case study demonstrating the numerical efficiency of the suggested
approach. The case study codes, data and results are posted at the website. The case study was
done with the Portfolio Safeguard (PSG) optimization package which has precoded Risk,
Deviation, and Error functions for the considered Quadrangles.
International Data Science & Engineering Symposium
IDSES
Stan Uryasev
Inflation refers to an ongoing and overall comprehensive increase in the overall level
of goods and services price in the economy. Today; inflation, which is tried to be kept under control
by the central banks, is trying to ensure price stability, the continuous price changes that arise in
all the goods or services that consumers use includes. Undoubtedly in terms of economy, inflation
expectations are also ganing importance, except for rhe realized inflation. This situation makes it
necessary to predict the future vaules of inflation. In that case, a reliable estimate of the future
values of inflation in any country will create an entry in determining the policies that decisionmaker
units will implement on the economy.
The aim of this article is to predict inflation in the next period by using the Consumer Price Index
(CPI) data with two alternative techniques. It is also aimed to examine the prediction performances
of these two techniques in comparisons. Thus, the first of the two main objectives of the study is
to predict the future values of inflation with two alternative techniques. The second goal is to
determine which of these two techniques well compared to statistical and econometric criteria.
In this context, the estimated performance of both techniques was predicted by the 9-month
inflation, Box-Jenkins (ARIMA) and Artificial Neural Networks (ANN) in the April – December
2019 period, using CPI data consisting of 207 in the period of January 2002 – March 2019. In the
study, Eviews and Matlab programs were utilized.
International Data Science & Engineering Symposium
IDSES
Erkan IŞIĞIÇOK
Ramazan Öz
Savaş Tarkun
Acceptance sampling plans are a decision-making process on the basis of a randomly
selected sampling from a party, where it is not possible to completely scan the products for
reasons such as time and cost being limited or the formation of damaged products during the
inspection. For some products, the life span (time from beginning to failure) may be an
important quality characteristic. In this case, the quality control adequacy of the products can
be checked with an acceptance sampling plan based on the truncated life test with a censored
scheme for the lifetime of the products. Acceptance sampling plans based on life test of product
life in industry are called reliability plans. In this study, group acceptance sampling plans based
on life tests were studied under the type-I censored scheme for the compound Weibull-
Exponential distribution. Optimum sample size, optimum number of groups and acceptance
number were obtained.
International Data Science & Engineering Symposium
IDSES
Canan HAMURKAROĞLU
Ayten YİĞİTER
The interest in data science is increasing in recent years. Data science, including
mathematics, statistics, big data, machine learning, and deep learning; can be considered as the
intersection of statistics, mathematics, and computer science. Although the debate continues about
the core area of data science, the subject is a huge hit. Universities have a high demand for data
science. They are trying to live up to this demand by opening postgraduate and doctoral programs.
Since the subject is a new field, there are significant differences between the programs given by
universities in data science. Besides, since the subject is close to statistics, most of the time, data
science programs are opened in the statistics departments, and this also causes differences between
the programs.
Data science education has to be more project-based since up to now, there is no core knowledge
of data science like other sciences. It is probably the hypercorrect choice to learn this job in a
university which is intertwined with industry and provides plenty of opportunity for internships.
In this article, we will summarize the data science education developments and give curriculum
examples from the world at the undergraduate and graduate level. Regarding these examples, every
university thinks data science as he wants and the names and the contents of these programs really
differs.
International Data Science & Engineering Symposium
IDSES
Necmi GÜRSAKAL
Fırat Melih Yılmaz
Ecem Özkan
Deniz Oktay
Due to the new trends in global warming, environmental consciousness has become a
greater concern among the organizations and industries globally. Green Supply Chain
Management (GSCM) has received an increased attention from academia and industries in last
few years. Green Supply Chain Management (GSCM) has emerged as an innovative
organizational strategy to reduce environmental impacts of supply chain activities as well as
efficient usage of energy and material. In today’s developing economies, customers are becoming
more conscious of the environment and governments are making stringent environmental laws, so
the industries need to reduce the environmental impact of their supply chain activities. There are
many barriers which directly and indirectly affect the implementation of GSCM in an organization.
In this paper twenty barriers have been identified through extensive literature review and expert
opinion of academicians. These barriers are found to exist in all organizations irrespective of
industry domain. Due to the presence of various barriers, Indian organizations are struggling to
implement GSCM in their operations. By removing the barriers, Indian industries can focus on
cleaner production by adopting Green Supply Chain Management (GSCM) in their operations.
The Objective of the study was to identify the GSCM barriers in Indian context. The research
methodology was perusing literature in GSCM and validating by experts opinion. Literature was
perused irrespective of industry domain. The study concludes by narrowing on twenty barriers
which play a prominent role in the Indian context.
International Data Science & Engineering Symposium
IDSES
Mohd. Azmi Khan
Salma Ahmed
Sadia Samar Ali
With the increase in computational power and big data, studies on artificial
intelligence are increasing day by day. Especially deep learning applications are seen in almost
all areas of our lives. The most successful results of deep learning architectures are in image
processing. Different architectural approaches are tried to make image processing fast. Due to
the fact that video images consist of large capacity data, it is very important to achieve high
performance in these video images. In this study, size reduction function has been proposed
that can reduce the size of the high-quality and large-capacity file data and produce results with
a high accuracy rate. The results of the proposed method were compared in terms of
performance and speed with different architectures in image processing using CNN
(Convolutional Neural Network) algorithm. In addition, an application that uses the
recommended size reduction function has also been developed using the Python programming
language.
International Data Science & Engineering Symposium
IDSES
Ömer PİŞGİN
Ali Osman ÇIBIKDİKEN
processes in the business world have two fundamental disease, including deviation and
variability from the average (target). One of the statistical process control graphs used for
quantitative variables is to keep the average and the other to control variability. Apart from the
normal distribution or nonnormal of quantitative data, the average and variability are controlled or
not, and then the capability of process or machine is checked. The desired outcome is in addition
to the normal distribution of data, the process is under control and capable. On the other hand,
capability analysis is defined as the machine capability analysis when it is performed for the
machine, while the process capability analysis takes its name when it is done for the process. In
this study, machine capability analysis has been applied.
The aim of this article is to investigate the effects of the normal distribution or nonnormal of data
on machine capability analysis. For this purpose, data on the lengths measured by surface of the
shock absorber body pipe cut by a CNC machine in a company in the automative industry were
used. In the study, 50 observations values were used, and the lower specification limit was 124.5
and upper specification limit was 125.5, and the CNC (pipe cutting) machine was sufficient or not.
The analysis first started with the implementation of the normality test and the data was not
distributed normally. It is concluded that assuming this data, which does not have normal
distribution, is normally distributed, and the machine is under control and is also sufficient with
the I-MR control charts in the Minitab program. The same analysis was applied with the nonnormal
command under the assumption that the data was nor normally distributed, and even in this case
the machine was sufficient. In addition, the data that does not have normal distribution has been
transformed into normality, and I-MR control charts and machine are under control and also
sufficient. According to the findings, the average and specification limits of the values of I-MR
control charts are the same and the machine capability results are different. In this study, these similarities and differences were examined comparatively. Let us add it right away; these findings
are specific to the machine and cut pipes we take into consideration and should not be generalized.
International Data Science & Engineering Symposium
IDSES
Erkan IŞIĞIÇOK
Gözde TÜRK
Teknolojinin ilerlemesi ile birlikte birçok veri dijital ortamlarda kayıt altına alınarak
büyük veri yığınları ortaya çıkmıştır. Veri madenciliği sayesinde bu büyük veri yığınlarının
içinden anlamlı ve yararlı bilgilerin ortaya çıkarılması için çalışmalar yapılmaktadır. Özellikle
büyük veri yığınlarını analiz etmede klasik analiz yöntemlerinin yetersiz kalması veri
madenciliği yöntemlerinin önemini arttırmıştır. Her dönemde olduğu gibi günümüzün en
önemli araştırma alanı olan tıp alanında da sürekli olarak hastalara ait veriler artarak kayıt altına
alınmaktadır. Kayıt altına alınan veriler bazen tek başına anlamsız gibi görünürken diğer
verilerle birlikte bütünsel olarak analiz edildiğinde gizli kalmış önemli bilgiler elde
edilebilmektedir. Bu değerli bilgiler, sağlık sektörünün gelişmesine ve doktorların daha doğru
bir şekilde teşhis verebilmesine yardımcı olmaktadır. Bu çalışmada, Kronik Böbrek Hastalığı
(KBH) veri seti üzerinde analiz yapılmıştır. Farklı modeller oluşturularak, bu modellerin veri
üzerindeki tahmin sonuçları karşılaştırılmış ve bu sonuçlara bağlı olarak veriler üzerinde hangi
modelin daha iyi sonuç verdiği belirtilmiştir.
International Data Science & Engineering Symposium
IDSES
Mustafa İlker ERDURSUN
Hasan ERBAY
Ömer Faruk AKMEŞE
İbrahim DOĞAN
Rapidly developing technology in the world endangers many companies we are in
today. The companies are now obliged to improve themselves in order to compete in both
domestic and international markets. In this sense, concepts such as quality, cost and customer
satisfaction are becoming more important. Lean production is a production method that is not
outdated at this point. In the 1950s, the lean production which was laid the foundation with the
Toyota Production Company, is still up to date. Lean production, which adopts zero error and
zero inventory principle, is applied in both the manufacturing sector and the service sector.
Today, it is still very important in terms of the continuity of lean production which is developed
and integrated with new technologies. Lean production is a gateway to the world from the
Japanese people. In this thesis,” Lean Manufacturing Application in Iron-Steel Production Line
Hatt is tried to be explained. Lean production and techniques are examined in detail. Then, in
light of this information, the results obtained in iron and steel industry were interpreted.
International Data Science & Engineering Symposium
IDSES
Taner ERSÖZ
Filiz ERSÖZ
Karanfil SARIZ
LiBerated Social Entrepreneurship in Developing and Emerging Countries consists of
a social entrepreneur using business metrics, to sustain social impact. We study differences
between developing and developed countries, introducing a new OR approach to development.
Commercial entrepreneurs are generally oriented to business metrics like profit, revenues and
return. Instead, social entrepreneurs are non-profits or a blend with for-profit goals, generating
Return to Society. In DCs, a social entrepreneurship has been uncommon. We introduce a midway
as LiBerated Social Entrepreneur, where social businesses should be sustainable. We apply
Game and Max-Flow - Min-Cut Theories, Schumpeter’s creative destruction and Adam Smith’s
diversification model for our business plan. As a result, B. Kjamili started Migport, formerly QZenobia:
a mobile application that runs as a “refugee portal”, supported by “Refugee Big-Data
Analytics”: refugees submit data to the application via “questionnaire” and search for
opportunities, verified news privatized based on their answers. The idea of both-sided help with
benefit generated by D. Czerkawski is an extension of B. Kjamili's conception. Nshareplatform
(NSP) will create a friendly public space for people with disabilities, understanding they needs. It
tries to facilitate better communication between “two worlds”- Ability and Disability and
personalizes an assistant (Special person helping people with disabilities). Multivariate Adaptive
Regression Splines (MARS), Conic MARS (CMARS) and its robust version RCMARS have
shown their potential for Big-Data and, recently, Small-Data. With that toolbox, we aim to further
support our joint and novel project.
International Data Science & Engineering Symposium
IDSES
Gerhard-Wilhelm WEBER
Berat KJAMILI
Dominik CZERKAWSKI
The map method works with a certain algorithm and the inputs which send to a value
list as a parameter. All values in the list converted to the intermediate result list. We sort all of the
data then obtain the map list. The proposed and developed map structure was tested with quicksort
approach. The sorting process depends on the byte situation of the each data. Small data can
do it easily side by side. Thus, small data do not need to applying of the reduce process. Sample
selection havbe to will be easier. The goal is to give an intermediate operationtothe map-reduce
structure. The more accurate is to get a ranking. In large data analysis, the data mapping sequence
and reduction works with a certain algorithm structure, introducing and sending inputs as a
parameter to a value list. An intermediate result list is created by converting all the values in the
list, which are included in the entered system. In the structure developed after the mapping (Map)
process, the mapping list is divided and obtained. The order depends on the byte value that is
generated by each data. In case of large volume data, the data will be used without using a single
line operation. In short, the data may be side-by-side, so there is no need to apply the reduction to
each data. Therefore, the process of selecting the sample will be easier.
International Data Science & Engineering Symposium
IDSES
Safiye TURGAY
Suat ERDOĞAN
At the moment we are entering a new era of knowledge, we are experiencing the first
stages of a digital transformation. Businesses should be restructured in a structure suitable for the
digital age when they are moving towards Industry 4.0, which is rising on cyber physical systems.
In this period when the digital enterprises are rising, each business unit of the enterprise should be
in the effort of using these innovations in accordance with this structure and using them in the most
effective way to reach their goals. With the understanding of marketing in the modern sense, one
of the most basic functions of marketing is to provide the information about the customer demands
and expectations in order to be used in the product design and planning stage. For this reason,
marketing is not only about the sales and marketing of the products or services produced, but also
for the decision of what to produce. In this period, where the effects of the digital age are becoming
more evident, the weight of digital applications in marketing is increasing and the concept of
Digital Marketing is becoming more prominent. Innovations on the transformation of Data (which
is called New oil) into information, give enterprises more competitive advantages and help to make
more accurate decisions. In parallel with the developments in technology, the change in the market
and the business environment necessitates the marketing to be prepared for these changes and to
transform itself with this change. In this study, the analytical concept is discussed with marketing.
In addition, the relationship between the analytical concepts of marketing is tried to be explained
and the opportunities awaiting marketing are put forward. At the end of the study, it has been
concluded that this digital transformation and developments all over the world increase the
importance of marketing.
International Data Science & Engineering Symposium
IDSES
Muhammet GİRGİN
Support vector machine is an effective machine learning method based on statistical
learning theory and used for classification problems. The optimization of the parameter is very
important in order to increase the classification accuracy. Meta-heuristic methods are one of the
main optimization approaches that can be applied in this context and have been used frequently
for parameter optimization in recent years. These methods are generally particle swarm
optimization, genetic algorithm, grid search method, differential evolution algorithm, ant colony
optimization. In this study, support vector machine parameter optimization studies between 2010-
2019 were investigated. According to the results of these studies, it was observed that parameter
optimization through meta-heuristic methods significantly increased the rate of classification
accuracy of classifier and significantly reduced the workload.
International Data Science & Engineering Symposium
IDSES
Zübeyir ÖZKORUCU
Turgut ÖZSEVEN
The revolutionary changes of the last decade in the banking sector have led to the
integration of information and computer networks into a single information and cybernetic space,
which has led to the creation of automated banking systems that have substantially expanded the
spectrum of electronic services of state and commercial banks of the world. As a result, threats to
such a national information resource of the state as the banking information resources under which
the banking information refers. Threats to the security of banking information resources have
become signs of hybridization. Manifestations of hybridity as a result of the simultaneous impact
of threats to information security, cybernetic security and information security on banking
information resources have led to the emergence of synergies, the negative manifestations of which
require a radical revision of the concepts of the construction of existing security systems. Thus, it
becomes clear that there is a need for a radical revision of the current methodological principles
for building a security system for banking information resources both for Ukraine and for the world
as a whole.
International Data Science & Engineering Symposium
IDSES
Serhii Yevseiev
Multi-objective optimization problems allow multiple purpose to be simultaneously
optimized. The nondominated sorting genetic algorithm II (NSGA-II), which is one of the most
effective multi-objective heuristic methods in the solution of multi-objective optimization
problems, is widely used in the literature. NSGA-II obtains a Pareto optimal solutions, known as
a set of dominant solutions without requiring any prior knowledge in one run. The NSGA-II is
more useful than the classical genetic algorithm, minimizing the computational complexity by
calculating the fast dominated sorting approach and the crowded distance without having to repeat
for each solution. In this study, NSGA-II method was used to optimize the cutting parameters of
hard materials turning. In the experimental studies, the regression models based on the cutting
velocity, feed rate and depth of cut parameters represent three different objective functions. This
optimization problem, which has five objective functions with three variables, has been discussed
by NSGA-II method. The optimal solution of these functions is to use the NSGA-II method to find
the most suitable set of Pareto solutions. The solutions obtained by using NSGAII method have
been found to be successful in multi-objective optimization problems. In addition, decision makers
from the optimal solutions can choose the most suitable solution according to their importance in
the objective functions.
International Data Science & Engineering Symposium
IDSES
Ahmet KOCATÜRK
Bülent ALTUNKAYNAK
Brass alloys are characterized by excellent workability, high thermal and electric
conductivity, corrosion resistance as well as exceptional antibacterial properties and are
therefore widely used in various industries, such as electric and electronics, automotive and
sanitary industry. However, power consumption should be eliminated for cleaner production in
terms of sustainable machining. Therefore, this study aims modelling of cutting forces in hard
turning of C23000 brass based on finite element method. The cutting parameters are chosen as
cutting speed, depth of cut and feed rate with three levels. The average of 4.66% difference is
achieved between experimental and simulated feed forces while 4.39% difference for main
cutting forces. The finite element modelling of cutting forces is quite compatible with the
experimental results and it can be performed by high accuracy without excessive machining
experiments of high machinability materials.
International Data Science & Engineering Symposium
IDSES
Mehmet Erdi KORKMAZ
Ball and beam system (BBS) is a benchmark hardware for designing control action. The
structure of the system is based on changing the angle of the beam so that the position of the ball
is changed. It is desired to move the ball to a reference position. In this paper Fuzzy Logic
Controller (FLC) is applied for this problem. Instead of conventional FLC, the derivative and
integral terms are integrated to the FLC, which is called as PID-like FLC. This controller has a
constant Fuzzy structure with variable parameters. The performance of the controller is based on
these parameters. Therefore, in this study, the parameters of PID-like FLC are optimized by using
three optimization algorithms; Genetic Algorithm, Particle Swarm Optimization, and Differential
Evolution. The performance of the controller is demonstrated on both simulation and hardware
environment. The performance of the optimization algorithm with respect to the obtained
performances are compared in this paper.
International Data Science & Engineering Symposium
IDSES
O. Tolga ALTINÖZ
A. Egemen YILMAZ
The possibilistic mean – variance (MV) model enables the practitioners to incorporate
the expert knowledge and robust statistics into the portfolio selection. Hence, it is a considerable
alternative in decision making under uncertainty. In this study, we will examine the possibilistic
mean – variance model theoretically under the assumption that the possibility distributions of the
asset returns are given with the triangular fuzzy numbers. Here, the triangular fuzzy numbers will
be determined based on the box plots. According to this, the possibilistic mean depends on the
data set’s median, interquartile range and skew. Furthermore, the possibilistic variance depends
only on the data set’s interquartile range. Then, we will illustrate this model based on the weekly
returns of ten sector indices in 2017. Moreover, we will compare the risk adjusted performance
and profitability of the possibilistic MV model and Markowitz’s traditional MV model where the
trading and testing periods cover the complete year of 2017 and 2018 respectively.
International Data Science & Engineering Symposium
IDSES
Furkan GÖKTAŞ
Süleyman DÜNDAR
Denim is a popular fabric among all of the age groups because of its good usage
performance and ability to provide convenience in adapting to changing trends in fashion. Apart
from the fashion and general performance properties, thermo-physiological comfort properties
such as air permeability are important for denim users. Fabric comfort depends on lots of factor
such as fabric structure and the types of fibers. Air permeability is a one of the comfort
properties of fabric is affected by many parameters of the fabric. A determination of the
relationship between the fabric parameters and the air permeability is highly complex and
difficult. For this reason, Artificial Neural Network model which has effective performance in
very complex problems was used. In the present study, an artificial neural network has been
used to predict air permeability amongst different denim production parameters. Finally, by
comparison with the experimental results, the efficacy of the proposed model was verified.
International Data Science & Engineering Symposium
IDSES
Esra AKGÜL
Emel AYDOĞAN
Yılmaz DELİCE
Drilling operation is one of the most frequently used process in manufacturing industry.
Especially in machine manufacturing sector, drilling operations take one third of the total
production time. Drilling operation cost a lot, not only because of taking some serious time of total
production time, but also because of the drill usage for this process. In this study, the cost
comparison of three different methods evaluated for drilling operation in CNC machines. These
methods are drilling cycle (G81), high-speed peck drilling cycle (G73) and deep hole peck drilling
cycle (G83). The experiments were performed according to the cutting parameters suggested by
the cutting tool company and the machining times measured in these three different methods. A
novel program coded on Microsoft Visual Studio 2017 C#, which is able to calculate from machine
amortization to workmanship, the whole process cost. Process costs can be calculated according
to the number of holes in these different methods, through this program. Furthermore, drilling
operation costs can be calculated for different cutting parameters too.
International Data Science & Engineering Symposium
IDSES
Tugay ÜSTÜN
Yakup TURGUT
Over the last decades, marketing scholars have often drawn attention to the value of
customers for businesses that aim to endure in a harsh competitive environment. Customer
Relationship Management (CRM) has been a prominent approach in marketing management that
aims to improve relationships with customers. A practical implication of the CRM approach is the
analysis of customer data to extract value for businesses, as well as customers. Segmentation has
been a useful task that helps to group customers with similar attributes and designate better-tailored
marketing strategies for customer groups. Among a variety of approaches for customer
segmentation, Recency Frequency Monetary (RFM) Model stands out as an easy-to-adopt and
effective technique. In this study, segmentation with RFM approach will be conducted over the
purchase records obtained from an e-retailer. The segments and relevant marketing strategies will
be presented in the findings. Moreover, a software implementation for the RFM model will be
introduced along with a case study.
International Data Science & Engineering Symposium
IDSES
İnanç KABASAKAL
Nowadays, with the developing technology, the amount of data that is owned and
processed is increasing day by day. It is very important to ensure the security of data, which is one
of the biggest assets for institutions and organizations. With traditional security methods, attacks
can be detected and prevented, but cybercriminals spend a lot of time and resources on advanced
and targeted attacks that can bypass these methods. The present methods are reactive because they
are generally updated with the information obtained from the analyzes performed after a successful
attack. More proactive approaches are needed to improve safety. Cyber threat intelligence
represents such a proactive approach and involves collecting and analyzing information for
potential threats from a wide variety of data sources. The purpose of cyber-threat intelligence is to
proactively adapt security controls to understand the methodology used by different attackers and
to detect and prevent such activities. In the world of technology, the defense against attacks is one
of the most important issues. Today, different approaches and effective methods have been used
to obtain intelligence. These include vital information about security threats, which are used by
hacker forums and other platforms as a means of communication between hackers. The amount of
data on such platforms is very large. The manual analysis of these data is time-consuming,
ineffective and requires a considerable amount of resources. In this sense, machine learning has
become one of the popular approaches used in the field of cyber-threat intelligence in terms of its
suitability to the subject, producing beneficial and effective results. In this study, information is
given about cyber threat intelligence and in the world of hackers, how to obtain intelligence by
using machine learning techniques is examined and evaluated in detail by supporting the studies
conducted in the literature.
International Data Science & Engineering Symposium
IDSES
Cemile SARICAOĞLU
Mehmet Demirci
Smart agriculture is the correct and economical use of resources, product control, and
production to be more ergonomic in order to increase production efficiency in agriculture. The
concept of internet of things that are widely used in conjunction with Industry 4.0 has started to be
applied in agriculture. The main objective is to apply the automation systems that provide
communication among themselves to agriculture in order to increase production efficiency. In this
study, the current state of the smart agricultural systems with IoT has been investigated. As a result,
the risks that may occur in production can be predicted in a short time with smart agriculture. Thus,
proper and wasteful use of the resources required for agriculture is foreseen.
International Data Science & Engineering Symposium
IDSES
Mervenur Sağlam
Turgut ÖZSEVEN
Increasing the flexibility and effecient of enterprises from procurement to sales
provides a great competitive advantage for meeting consumer demands. Providing competitive
advantage is possible through the effective implementation of innovative technologies of fourth
industrial revolution in the all stages of supply chain process. In this context, the technologies
related to industry 4.0 were mentioned in the study and the differences between the traditional
supply chain and the digital supply chain were determined. In addition, the industry 4.0
applications in the digital and green supply chain are mentioned and the steps that must be
followed in the process of transition to the digital supply chain are indicated.
International Data Science & Engineering Symposium
IDSES
Deniz MERDİN
Filiz ERSÖZ
Abstract: In this study, a buffer stock between two machines which are working at the same
speed, is considered. It is assumed that the stock level between two machines alters in the
interval 0 , 2α. In the case that only the first machine is broken, the stock level can decrease
to zero if repairing time of the first machine is extended. This causes the second machine is
required to be halted. After fixing the first machine, the system begins to work again. If the
second machine is broken and the first machine is working, then the process proceeds until the
stock level reaches the maximum level 2α and then the first machine will be halt compulsory.
In order to re-work the system, the repairing of the second machine must be completed. Under
these assumptions, the buffer stock level will be stochastically fluctuated in the interval 0 , 2α. Thus, the buffer stock between two machines can be expressed by a stochastic process Y t and
it is observed that this process is a random walk with two barriers. Some problems of queuing
theory, stock control, reliability, insurance models and risk management can be expressed by
random walk and its modifications. In literature, there are several considerable studies (e.g.,
Aliyev and Khaniyev, Feller, Janseen and Leeuwarden, etc.). In this study, the stationary
characteristics of the random walk process Y t which represent the buffer stock level, are
investigated. Especially, for all moments of the ergodic distribution of the process Y t, the exact
expressions are obtained under the assumption that the random walk Y t generated by bilateral
exponential distributed summands. Moreover, the exact and approximate expressions for
variance, standard deviation, coefficient of variation, skewness and kurtosis of the process, are
obtained. It is also observed that the ergodic distribution of the standardized stochastic process
W t = Y t / a weakly converges to a triangular distribution in the interval 0, 2.
International Data Science & Engineering Symposium
IDSES
Zülfiye HANALİOGLU
Tahir HANALİOGLU
Selection and evaluation problems have uncertainties due to concept and its
perception discrepancies. Fuzzy set theory is one of the widely used methodology to cope with
these uncertainties. There is a growing interest in evaluations using intuitionistic fuzzy sets.
Differently from fuzzy sets, intuitionistic sets include both belonging degrees and nonbelonging
degrees. Thus, the evaluation using intuitionistic fuzzy sets gives more realistic results. In this
study, an intuitionistic fuzzy set-based evaluation system is proposed for supplier selection
problem of a construction company. This system has qualitative- and quantitative evaluation
parts. As qualitative part, decision makers of the company evaluate suppliers by the supplier
selection criteria: i) quality, ii) price, iii) delivery, iv) productivity, v) service, vi) flexibility.
The quantitative part includes supplier scores calculated through the current evaluation system
of the company. A supplier-evaluation database was created by the abovementioned parts. The
database was structured by α-cut representation of the evaluations. The calculations using
intuitionistic fuzzy sets were done for this database, which was occurred by lower and upper
bounds. After defuzzifying the database, suppliers were classified using the well-known fuzzy
clustering algorithm, fuzzy c-means. The classification using fuzzy clustering algorithm has
95.0% accuracy.
International Data Science & Engineering Symposium
IDSES
Ayşenur AKIN
M. Bahar BAŞKIR
Hamza GAMGAM
Bayesian networks are probabilistic graphical representations which are used to build
models from data and/or expert opinion. They can be utilized for a wide range of tasks including
prediction, anomaly detection, diagnostics, automated insight, reasoning, decision making, etc.
Dynamic Bayesian Networks (DBN) are extensions of Bayesian networks with temporal
support, which can be used to model systems that dynamically change by the time. Nowadays,
DBNs are utilized in a wide range of applications including robotics, data mining, speech
recognition, digital forensics, protein sequencing, and bioinformatics. Several software tools
exist in the public as well as commercial domains that support modelling and simulation of
DBNs. However, these DBN software tools differ in terms of features support, ease of use,
documentation, user’s community, etc. Therefore, it has become important to establish various
metrics for selecting the proper software tools for creating and simulating DBNs, such as cost,
licensing, GUI, built-in support for inference algorithms, structural learning, data types, etc.
The goal of this survey is the evaluation and comparison of existing software tools for building
DBNs based on a set of users centered criteria.
International Data Science & Engineering Symposium
IDSES
Hüseyin ÇAMBAŞI
Özgür KURU
Mehmet Fatih AMASYALI
Sofiene TAHAR
Amazon Prime, Hulu, Apple TV, Puhutv, BluTV, Turkcell TV Plus vb. gibi digital
platformlar insanlar üzerindeki etkisini artırmaya devam ediyor ve Netflix bunlardan birisidir.
Netflix insanların seyretmek istediği film ve dizileri belirleyip onlara en kısa yoldan ulaştırmayı
hedefleyen bir platformdur. Bu platform günümüzde çok popüler hale gelmiştir ve sinema
sektörüyle yarışır seviyede bulunmaktadır. Netflix, veri madenciliğini etkili bir şekilde kullanarak
insanların neyi izlemeyi sevdiğini bilmektedir. Ayrıca, Netflix aşırı ve gereksiz verilerden
kurtularak büyük verilerin değerli bilgilere dönüştürülürken daha net olmasını sağladı. Bu bilgiler
ışığında ülkeler için ayrı dizi ve filmlerin çekilmesine öncülük etmiş, izleyicilerin sevdiği aktör ve
yönetmenleri bir araya getirerek, çektiği filmlerin izlenilirliğini arttırmış, böylece de Netflix
markasını dünyaya tanıtarak izleyicileri kendine bağlamayı başarmıştır. Bizim bu çalışmadaki
amacımız Netflix kullanıcılarının davranışlarını incelemek ve açıklamaktır. Bunu sağlamak için
anket çalışması yapılmış olup, Netflix platformunun kullanma maksadının belirleyicilerini
araştırmak için teknoloji kabul modeli kullanılmıştır.
International Data Science & Engineering Symposium
IDSES
Ufuk CEBECİ
Oğuzhan İNCE
Teknoloji kavramının insanoğlunun hayatına girmesiyle teknolojinin kullanım
alanları gün geçtikçe artmıştır. Gelişen teknik beceriler ile üretimin odağı değişmiştir. Zamanla
üretimin verimliliği tartışılmaya başlamıştır. Bu tartışmalar sonucunda endüstriyel hayata
kalite, optimizasyon, verim gibi yeni kavramlar girmiştir. Üretim süreçleri üzerine yapılan
araştırmalarda daha kısa sürede, daha az maliyetli ve daha kaliteli üretim için yeni fikirler ortaya
çıkmıştır. İkinci Dünya Savaşı’nın her alanda olduğu gibi üretim yöntemleri üzerinde de etkisi
olmuştur. Artan teknoloji arayışları yeni üretim ve kalite anlayışlarına olan ihtiyacı
doğurmuştur. Bu çalışmada yalın üretim ve altı sigma kavramları incelenmiş, yöntemler
gösterilmiş ve seçilen bir tersanede uygulaması yapılmıştır. Yalın üretim felsefesi, üretim süreci
boyunca oluşan artık ve katma değersiz durumları elimine etme üzerine oturtulmuştur. Altı
sigma metodoloji ise üretim çıktısındaki hataların azaltılmasını ve üretimin standartlaştırarak
hızlanmasını hedeflemektedir. Yapılan uygulama sonucunda teorik olarak süreçlerdeki
kayıpların büyük oranda azaldığı ve firmanın iyileştirme çalışmaları doğrultusunda kazanç
sağlayacağı görülmüştür.
International Data Science & Engineering Symposium
IDSES
Emre GÜVEN
Ufuk CEBECİ
Today, with digitalization, it is possible to read digital data and make the right decisions
based on analytical results. Along with big data, the science of data management and analysis is
evolving to enable organizations to transform their knowledge into information that will help them
achieve their goals. In this study, it is given as an example to increase awareness of big data, data
mining, data mining and its applications in various sectors in Turkey.
International Data Science & Engineering Symposium
IDSES
Filiz ERSÖZ
The success of an Ad Mediation Software’s decision on from which ad network to
request ads and in which order depends on the ability to estimate eCPM (effective Cost Per Mille)
value used to measure ad revenue. This value varies for different applications depending on
different external factors. It is not possible for domain experts to make successful predictions by
analyzing different sets of external factors for a large number of applications and to keep these
estimates constantly up to date. Therefore, eCPM values were automatically estimated separately
for each application based on different advertising spaces and different countries using time series
analysis. The ARIMA model was used to estimate and the hyper-parameters of the model were
optimized by using grid search method. For most of the values obtained, it was found that the
values obtained with the intuition of domain experts were closer to the actual values.
International Data Science & Engineering Symposium
IDSES
Görkem GİRAY
Murat Osman ÜNALIR
Şeyma TAHMAZ