Classification of Customer Satisfaction Through Machine Learning: An Artificial Neural Network Approach

Abstract

This study aims to classify customer satisfaction data from Café Alvina using Machine Learning, specifically by implementing the Backpropagation Artificial Neural Network. The data used in this study consists of 70 training data and 30 testing data, with the input layer of the Artificial Neural Network having 5 neurons and the output layer having 2 neurons. The tested Artificial Neural Network models include the 5-5-2 model, 5-10-8-8-2 model, 5-5-10-2 model, and 5-8-10-2 model. Among the four models used in the testing process of the Backpropagation Artificial Neural Network system using Matlab, the 5-10-8-8-2 architecture model performed the best, achieving an MSE (Mean Squared Error) of 0.000999932 during training with 2920 epochs and a testing MSE of 0.000997829. After conducting the testing, the performance of the Artificial Neural Network models was as follows: the 5-5-2 model achieved 81%, the 5-10-8-8-2 model achieved 100%, the 5-5-10-2 model achieved 98%, and the 5-8-10-2 model achieved 96%. Through the implementation of Backpropagation Artificial Neural Network, the classification of customer satisfaction can be effectively performed. The trained and tested data demonstrate that the Artificial Neural Network can accurately recognize the input data in the system.