Vertical Search Engine for Academic Publications
Asım Yüksel Muhammed Ali Karabıyık
AbstractWith the advancing technology, the storage of large amounts of data has become possible. Unstructured nature of data makes it difficult to access. Many sectors demand access to specific information within their area. Thus, it has emerged the concept of vertical search engine. In our study, a crawler was designed to filter reliable sites. The designed crawler only adds results related to academic publications to the database. Naive Bayes classifier algorithm was employed to identify the science branch of an academic publication by using its abstract. According to our experiments, the accuracy rate of developed vertical search engine was 70%. The application is designed in a way that it can self-learn so that the success rate can increase.