The Understanding of Customer Satisfaction on A Fintech Application Using A Machine Learning Approach

TL;DR: In conclusion, machine learning plays a significant role in analyzing and classifying customer satisfaction levels, with Random Forest outperforming other machine learning models.

Abstract: Fintech companies should prioritize assessing customer satisfaction and focus on specific market segments to attract and retain customers effectively. This study investigates customer satisfaction levels within the flip.id application, utilizing the PIECES framework and machine learning techniques. Among various models, the Random Forest classification emerges as the superior performer. With an 80% training data and 20% testing data split, the model achieves impressive metrics: accuracy of 92.7%, F1 score of 91.9%, precision of 91.8%, and recall of 92.7%. Detailed analysis using the Confusion Matrix reveals that 134 instances labeled as “Satisfied” were correctly classified, five instances labeled as “Uncertain” were accurately categorized, and one instance labeled as “Very Satisfied” was correctly identified. In conclusion, machine learning plays a significant role in analyzing and classifying customer satisfaction levels, with Random Forest outperforming other machine learning models.