SEARCH RESULT

The use of Natural Language Processing (NLP) has become increasingly essential in enhancing the efficiency and responsiveness of customer support systems. This paper provides a comprehensive review of NLP techniques applied in this domain, focusing on research published between 2020 and 2024. It focuses on the comparative performance of widely adopted algorithms, including Support Vector Machines (SVM), Bidirectional Encoder Representations from Transformers (BERT), Term Frequency- Inverse Document Frequency (TF-IDF), Convolutional Neural Networks (CNN), Multilayer Perceptron (MLP), and K-Nearest Neighbors (KNN). These methods were evaluated across common tasks such as sentiment analysis, chatbot response generation, and customer review classification. The study highlights a significant performance advantage of deep learning models over traditional approaches. While traditional models such as TF-IDF combined with SVM exhibited varying accuracy (ranging from 40% to 87.41%) depending on dataset quality and feature engineering, deep learning models (architectures based on BERT and its variants) achieved remarkable accuracies, reaching as high as 99.21%. Furthermore, the review notes that most studies rely on static datasets, this may limit how well their outcomes apply to real-time customer service. This paper contributes to the field by presenting a comparative synthesis of state-of-the-art NLP techniques applied in customer support, emphasizing performance patterns and practical challenges. The findings provide useful guidance for researchers and practitioners aiming to develop or enhance NLP-based customer service systems.

International Conference on Advanced Technologies, Computer Engineering and Science
ICATCES

Nisreen BOUTA Mohamad ZUBI Bilal YOUSFI Ammar ALQADASI