TEXT CLASSIFICATION WITH FREQUENCY-BASED TEXT VECTORISATION METHODS FOR ENHANCING CALL CENTRE EFFICIENCY

Abstract

In today's business world, many transactions take place over the phone or online. Call centers play a significant role in dealing with different situations and solving problems that come with the large volume of global business. As an interface between companies/institutions and customers, call centers aim to eliminate problems, correct mistakes, resolve conflicts, and increase customer satisfaction. The traditional approach involves customer service agents handling inquiries and complaints, but human error can hinder effective problem resolution. Intelligent assistant applications have emerged to augment the skills of customer service agents, improve performance, and maximize customer satisfaction. This study focuses on addressing the challenges faced by the Republic of Turkiye Ministry of Trade Call Center in the (RTMTCC), which handles over 10,000 calls per day. For this purpose, it introduces an intelligent framework that uses AI-driven methods and frequency-based text vectorization techniques to efficiently route calls to relevant departments, with the aim of increasing customer satisfaction and reducing economic losses. Using historical call texts, Bag of Words (BoW), and Term Frequency-Inverse Document Frequency (TF-IDF), the study evaluates the performance of five different classifiers: Stochastic Gradient Descent (SGD), Logistic Regression (LR), Naive Bayes (NB), Adaptive Boosting (AdaBoost), Artificial Neural Networks (ANN). The results indicate that the AdaBoost classifier generally outperforms others in both text vectorization approaches by reaching higher precision, recall and f1-score values. The study provides new approaches to automate call routing, evaluates how to classify text effectively, and shows the strengths and weaknesses of different text analysis methods, helping us to understand call center operations better.

Keywords

• Text classification
• Text vectorization
• TF-IDF
• BoW
• Call centers

References

1. [1] V. Mehrotra, T. A. Grossman, and D. A. Samuelson, “Call Center Management,” Wiley Encycl. Oper. Res. Manag. Sci., no. January, 2011, doi: 10.1002/9780470400531.eorms0130.
2. [2] A. O. Adeyemi, M. A. Saouli, and B. Sinha, “Influence of Call Centers on Emerging Business Models and Practices,” Asian J. Bus. Manag., vol. 6, no. 5, pp. 44–52, 2018, doi: 10.24203/ajbm.v6i5.5467.
3. [3] S. Chaudhary, N. Nasir, S. Ur Rahman, and S. Masood Sheikh, “Impact of Work Load and Stress in Call Center Employees: Evidence from Call Center Employees,” Pakistan J. Humanit. Soc. Sci., vol. 11, no. 1, pp. 160–171, 2023, doi: 10.52131/pjhss.2023.1101.0338.
4. [4] S. Ananthram, M. J. Xerri, S. T. T. Teo, and J. Connell, “High-performance work systems and employee outcomes in Indian call centres: a mediation approach,” Pers. Rev., vol. 47, no. 4, pp. 931–950, 2018, doi: 10.1108/pr-09-2016-0239.
5. [5] A. Keser and G. Yilmaz, “Workload , Burnout , and Job Satisfaction Among Call Center Employees,” J. Soc. Policy Conf., no. 66–67, pp. 1–13, 2014.
6. [6] J. Chatterjee, A. Saxena, and G. Vyas, “An automatic and robust system for identification of problematic call centre conversations,” Proc. - 2016 Int. Conf. Micro-Electronics Telecommun. Eng. ICMETE 2016, pp. 325–330, 2016, doi: 10.1109/ICMETE.2016.48.
7. [7] G. Mishne, D. Carmel, R. Hoory, A. Roytman, and A. Soffer, “Automatic analysis of call-center conversations,” Int. Conf. Inf. Knowl. Manag. Proc., no. May 2014, pp. 453–459, 2005, doi: 10.1145/1099554.1099684.
8. [8] D. Galanis, S. Karabetsos, M. Koutsombogera, H. Papageorgiou, A. Esposito, and M. T. Riviello, “Classification of emotional speech units in call centre interactions,” 4th IEEE Int. Conf. Cogn. Infocommunications, CogInfoCom 2013 - Proc., pp. 403–406, 2013, doi: 10.1109/CogInfoCom.2013.6719279.

9. [9] O. Rashid, A. M. Qamar, S. Khan, and S. Ambreen, “Intelligent decision making and planning for call center,” 2019 Int. Conf. Comput. Inf. Sci. ICCIS 2019, pp. 1–6, 2019, doi: 10.1109/ICCISci.2019.8716472.
10. [10] I. O. Yigit, A. F. Ates, M. Guvercin, H. Ferhatosmanoglu, and B. Gedik, “Çaǧrı Merkezi Metin Madenciliǧi Yaklaşımı,” in 2017 25th Signal Processing and Communications Applications Conference, SIU 2017, Institute of Electrical and Electronics Engineers Inc., Jun. 2017. doi: 10.1109/SIU.2017.7960138.
11. [11] K. Fiok et al., “Text Guide: Improving the Quality of Long Text Classification by a Text Selection Method Based on Feature Importance,” IEEE Access, vol. 9, pp. 105439–105450, 2021, doi: 10.1109/ACCESS.2021.3099758.
12. [12] H. Chen, L. Wu, J. Chen, W. Lu, and J. Ding, “A comparative study of automated legal text classification using random forests and deep learning,” Inf. Process. Manag., vol. 59, no. 2, 2022, doi: 10.1016/j.ipm.2021.102798.
13. [13] M. Sarı, “Derin Öğrenme Yöntemleri Kullanılarak Türkçe Doküman Sınıflandırma,” TOBB University of Economics and Technology, 2018.
14. [14] K. Koruyan and A. EKERYILMAZ, “Makine Öğrenmesi ile Müşteri Şikayetlerinin Sınıflandırılması,” AJIT-e Acad. J. Inf. Technol., vol. 13, no. 50, pp. 168–183, 2022, doi: 10.5824/ajite.2022.03.004.x.
15. [15] O. Uslu and S. Akyol, “Türkçe Haber Metinlerinin Makine Öğrenmesi Yöntemleri Kullanılarak Sınıflandırılması,” Eskişehir Türk Dünyası Uygul. ve Araştırma Merk. Bilişim Derg., vol. 2, no. 1, pp. 15–20, Jan. 2021, Accessed: Dec. 16, 2022.
16. [16] B. Karakus, G. Aydin, and I. R. Hallac, “Distributed Readability Analysis of Turkish Elementary School Textbooks,” Proc. Int. Conf. Inf. Technol. Comput. Sci., pp. 80–87, 2015.
17. [17] R. S. Kuzu, A. Haznedaroglu, and M. Levent Arslan, “Topic identification for Turkish call center records,” pp. 1–4, 2012, doi: 10.1109/siu.2012.6204647.
18. [18] H. Saif, M. Fernandez, Y. He, and H. Alani, “On stopwords, filtering and data sparsity for sentiment analysis of twitter,” Proc. 9th Int. Conf. Lang. Resour. Eval. Lr. 2014, no. i, pp. 810–817, 2014.
19. [19] G. Gupta, “Text Document Tokenization for Word Frequency Count using Rapid Miner (Taking Resume as an Example),” Int. J. Comput. Appl., vol. 1, no. March 2015, pp. 60–768887, 2009.
20. [20] T. Korenius, J. Laurikkala, K. Järvelin, and M. Juhola, “Stemming and lemmatization in the clustering of finnish text documents,” Int. Conf. Inf. Knowl. Manag. Proc., no. May 2014, pp. 625–633, 2004, doi: 10.1145/1031171.1031285.
21. [21] A. Barbaresi, “simplemma,” 2022. http://doi.org/10.5281/zenodo.4673264
22. [22] A. K. Singh and M. Shashi, “Vectorization of text documents for identifying unifiable news articles,” Int. J. Adv. Comput. Sci. Appl., vol. 10, no. 7, pp. 305–310, 2019, doi: 10.14569/ijacsa.2019.0100742.
23. [23] R. Zhao and K. Mao, “Fuzzy Bag-of-Words Model for Document Representation,” IEEE Trans. Fuzzy Syst., vol. 26, no. 2, pp. 794–804, Apr. 2018, doi: 10.1109/TFUZZ.2017.2690222.
24. [24] W. Aljedaani et al., “Sentiment analysis on Twitter data integrating TextBlob and deep learning models: The case of US airline industry,” Knowledge-Based Syst., vol. 255, p. 109780, 2022, doi: 10.1016/j.knosys.2022.109780.
25. [25] S. Maleki, M. Musuvathi, and T. Mytkowicz, “Semantics-preserving parallelization of stochastic gradient descent,” Proc. - 2018 IEEE 32nd Int. Parallel Distrib. Process. Symp. IPDPS 2018, pp. 224–233, 2018, doi: 10.1109/IPDPS.2018.00032.
26. [26] M. Bhattacharya and D. Datta, “Diabetes Prediction using Logistic Regression and Rule Extraction from Decision Tree and Random Forest Classifiers,” 2023 4th Int. Conf. Emerg. Technol. INCET 2023, pp. 1–7, 2023, doi: 10.1109/INCET57972.2023.10170270.
27. [27] V. Vijay and P. Verma, “Variants of Naïve Bayes Algorithm for Hate Speech Detection in Text Documents,” 2023 Int. Conf. Artif. Intell. Smart Commun. AISC 2023, pp. 18–21, 2023, doi: 10.1109/AISC56616.2023.10085511.
28. [28] T. K. An and M. H. Kim, “A new Diverse AdaBoost classifier,” Proc. - Int. Conf. Artif. Intell. Comput. Intell. AICI 2010, vol. 1, pp. 359–363, 2010, doi: 10.1109/AICI.2010.82.
29. [29] M. A. Elgammal, H. Mostafa, K. N. Salama, and A. Nader Mohieldin, “A Comparison of Artificial Neural Network(ANN) and Support Vector Machine(SVM) Classifiers for Neural Seizure Detection,” Midwest Symp. Circuits Syst., vol. 2019-Augus, pp. 646–649, 2019, doi: 10.1109/MWSCAS.2019.8884989.
30. [30] S. Akuma, T. Lubem, and I. T. Adom, “Comparing Bag of Words and TF-IDF with different models for hate speech detection from live tweets,” Int. J. Inf. Technol., vol.14, no.7, pp. 3629–3635, 2022, doi: 10.1007/s41870-022-01096-4.
31. [31] P. Baldi, S. Brunak, Y. Chauvin, C. A. F. Andersen, and H. Nielsen, “Assessing the accuracy of prediction algorithms for classification: An overview,” Bioinformatics, vol. 16, no. 5, pp. 412–424, 2000, doi: 10.1093/bioinformatics/16.5.412.
32. [32] B. C. ÖĞE and F. KAYAALP, “Farklı Sınıflandırma Algoritmaları ve Metin Temsil Yöntemlerinin Duygu Analizinde Performans Karşılaştırılması,” Düzce Üniversitesi Bilim ve Teknol. Derg., vol. 9, pp. 406–416, 2021, doi: 10.29130/dubited.1015320.
33. [33] B. EKİCİ and H. TAKCI, “Spam Tespitinde Word2Vec ve TF-IDF Yöntemlerinin Karşılaştırılması ve Başarı Oranının Artırılması Üzerine Bir Çalışma,” Bilecik Şeyh Edebali Üniversitesi Fen Bilim. Derg., vol. 8, no. 2, pp. 646–655, 2021, doi: 10.35193/bseufbd.935247.
34. [34] Ö. ÇELİK and B. C. KOÇ, “TF-IDF, Word2vec ve Fasttext Vektör Model Yöntemleri ile Türkçe Haber Metinlerinin Sınıflandırılması,” Deu Muhendis. Fak. Fen ve Muhendis., vol. 23, no. 67, pp. 121–127, 2021, doi: 10.21205/deufmd.2021236710.