Machine learning classification analysis model community satisfaction with traditional market facilities as public service

Abstract

<p>Traditional markets are public service facilities that can be utilized by the<br />community. The market function is used place where sellers and buyers meet<br />in conducting transactions. This study aims to build a machine learning<br />classification analysis model in measuring community satisfaction with<br />traditional market facilities. The analytical methods used include Fuzzy.<br />multiple linear regression (MRL), artificial neural network (ANN), and<br />decision tree (DT). Fuzzy is used to generate a pattern of rules in determining<br />the level of satisfaction. MRL serves to measure and test the correlation of<br />rules that have been formed. The ANN method is used to carry out the<br />classification analysis process based on learning. In the final stage. DT is used<br />to describe the decision tree of the analysis process. This study presents the<br />results of machine learning analysis which is very good in determining<br />satisfaction with an accuracy rate of 99.99%. This result is influenced by fuzzy<br />logic which can develop a classification rule pattern of 32 patterns. MRL also<br />shows a significant correlation level of 81.1% based on the indicator variables.<br />Overall, the machine learning classification analysis model can provide<br />knowledge to be considered in the management of traditional markets as<br />public service facilities.</p>