Accurate Voice of the Customer Interpretation Via Hybrid Fuzzy – Decision Tree Classification

Managing, analyzing, and interpreting the results of student surveys is a complex process that involves analyzing a large amount of data. In this article, the authors propose to develop an intelligent system for classifying students’ responses to identify the “Voice of the Customer.” The study suggests addressing the issue of feedback analysis by considering the ambiguity of natural language and implementing an accurate and efficient method of classifying students’ responses.

The research is conducted using the BERT algorithm that utilizes lexical features enhanced with intensifiers and superlatives identified in the Spanish language to implement sentiment analysis. The authors also apply fuzzy logic and decision-making rules to introduce membership functions that estimate the feedback’s polarity and classify it. Finally, the article describes the Business Intelligence dashboard that allows monitoring the key metrics in real time.

As a result, the methodology identifies that using fuzzy sets and recognizing superlatives in natural language can improve the accuracy of classification by reducing the number of false discoveries and omissions.