ANAGÜN Y., Bolel N. S., IŞIK Ş., Ozkan S. E.

JOURNAL OF ORGANIZATIONAL COMPUTING AND ELECTRONIC COMMERCE, vol.32, no.3-4, pp.217-231, 2022 (SCI-Expanded) identifier

  • Publication Type: Article / Article
  • Volume: 32 Issue: 3-4
  • Publication Date: 2022
  • Doi Number: 10.1080/10919392.2023.2210049
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, PASCAL, ABI/INFORM, Aerospace Database, Applied Science & Technology Source, Business Source Elite, Business Source Premier, Communication Abstracts, Computer & Applied Sciences, INSPEC, Metadex, Psycinfo, Civil Engineering Abstracts
  • Page Numbers: pp.217-231
  • Keywords: Customer complaint, text classification, complaint management, artificial intelligence, deep learning, agglutinative language, Turkish text classification, TEXT CLASSIFICATION, REVIEWS, PRODUCT, IMPACT
  • Eskisehir Osmangazi University Affiliated: Yes


In recent years, managing customer complaints poses a problem for companies due to the increasing market and customer base. One of the most effective ways to speed up the handling of complaints is to categorize customer issues and automatically forward complaints to relevant officers or departments. This reduces the response time to complaints and ensures that specific complaints are being handled by the people with the right expertise. Also, the companies can create a strategy exclusively for certain types of problems, which will hasten the problem resolution. In this article, we propose an intelligent customer complaint management system (CCMS) for financial services organizations. We described a pre-processing technique for Turkish agglutinative language using deep learning algorithms and it was not previously considered in the literature. Furthermore, the performance of the algorithm has been significantly increased by choosing the appropriate combinations of pre-processing tasks. The proposed method not only greatly increases text classification's utility for a broader range of customer complaints, but it also yields improved overall performance, recorded with a 96% accuracy score. The findings of the experiments show that the proposed approach is more effective than the other state-of-the-art strategies.