Comparison with Classification Algorithms in Data Mining of a Fuel Automation System's Sales Data

Abstract

This article deals with Otobil and pumps sales estimates at fuel stations. The fuel station data used in the study consists of 2384 data in total. Depending upon these data, classification procedures were performed on fuel station sales data using classification algorithms. In the study the classification algorithms that J48, Random Forest, KStar, Logistic Regression, IBk and Naive Bayes algorithms are used to compare the sales data estimations by using a software. The results obtained show that the accuracy rates of the J48 algorithm are more successful than others in general. It understands that these sales estimations shall encourage fuel station owners and association bodies to get more gainful.