A Hybrid Approach for Automatic Customer Feedback Classification in CRM

  • Ruma Panda, DR. A. N. Nandakumar
Keywords: CRM, SVM, Random Forest, Logistic Regression.

Abstract

Customer Relationship Management indicates the entire strategies, techniques, tools and technologies used by organizations for improving, holding and getting customers. It stores the complete information about overall purchase history, personal information, purchasing behaviour pattern etc. Customer feedback is the information and opinions given by different customers about the product or service which are divided into positive, negative and neutral comments. It collects the customer feedback from multiple channels like social media, survey responses, online product reviews or text from chats with customer support team as customer is a central part of any business strategy. It is possible to find out which aspects of business are working well and which may require improvement by gathering and analysing the customer feedback. The aim of feedback classification is to categorize the feedbacks into a fixed number of predefined classes. We have used mobile feedback data for classification. We propose a hybrid classification approach which is the combination of SVM and Random Forest with Logistic Regression to overcome the difficulties of SVM and Random Forest classifier. It shows that the hybrid classifier is more accurate than the individual classification systems.We used keyword based clustering algorithm to separate the feedbacks into different category based on the features and recommend the best accessories based on the feedback.

Published
2021-09-03
How to Cite
DR. A. N. Nandakumar, R. P. (2021). A Hybrid Approach for Automatic Customer Feedback Classification in CRM. Design Engineering, 5400- 5413. Retrieved from http://thedesignengineering.com/index.php/DE/article/view/3971
Section
Articles