Improving feedback analysis: Deep learning approach to college customer satisfaction assessments

Abstract

Establishing consumers' views via text-based feedback in a questionnaire is crucial for organizations, include education, since it gives a summary of significant areas that help administrators plan, regulations, and decision making. Through surveys, academic organizations have gathered huge quantities of textual data all over the years. For the organization, it is still difficult to analyse the vast quantities of unstructured feedback from customers to understand their concerns and opinions generally. In this study, we propose deep learning (DL) based technique called topic modelling that utilizing Naive Bayesian (NB) to automatically summarize text and retrieve ideas from this raw data. Additionally, it discusses the text mining procedure used to extract relevant information from the vast volume of text-based data. The most significant issues obtained through feedback from customers were subsequently identified. The findings showed particular issues for workplaces, including environment, staffing, IT infrastructure, and customer feedback system. The feedbacks also prominently highlight difficulties with the attitude of student assistance and security staff as well as the library's management and operations.