Text Analyze Studies Based on Customer Comments and Modeling of the Comments with Machine Learning Algorithms


Cesur E. , Işık M.

10th International Symposium on Intelligent Manufacturing and Service Systems, Sakarya, Turkey, 9 - 11 September 2019, pp.747-756

  • Publication Type: Conference Paper / Full Text
  • City: Sakarya
  • Country: Turkey
  • Page Numbers: pp.747-756

Abstract

Recently, text analytics technology has been utilized in many areas such as emotion analysis, categorization and text extraction. With this technology, the users are able to not only understand the customer's approach, preference and need but also revise their strategies according to the conditions. Within the scope of this study, firstly customer reviews will be analyzed by using various statistical methods. The analyses will be carried out in order to discover information pattern through the contents in the written text. The problem of text classification could be defined as an assigning classes process according to the content of the texts generated with certain techniques. Text classification applications could be seen in various fields such as cataloging web pages and news texts, optimizing search engines, extracting information, automatically processing e-mails. In this study, it is aimed to classify the customer comments with a specific procedure based on the several classification algorithms. At the end of the study, naive bayes, logistic regression and support vector machine algorithms were compared with the model performances on the interpretations and the success of the classification models was analyzed.