Data Mining to Determine Correlation of Purchasing Cosmetics With A priori Method

Abstract

Abstract Data mining is the process of analyzing data using software to find patterns and rules in the data set. Data mining can analyze large data to find knowledge to support decision making. In this study the Rule Association will be discussed as one of the data mining functions implemented using the A priori Algorithm. There will also be analyzed two support calculation techniques in candidate generation in the A priori Algorithm, such as: K-way and 2 Group-By in three sample datasets with transaction attributes id and item. In this study the problem of support calculation in candidate generation is the bottleneck of the A priori Algorithm where the improvement of the A priori Algorithm was emphasized on the candidate generation and the effectiveness of the A priori Algorithm. This research was lead on the Oracle RDBMS by utilizing WEKA tools to determine maximum support and confidence and to find out the correlation between products. The results shows the highest confidence value at 93% if you buy DeepClensingMilk and DeepClensingToner then you will buy Whitening Soap with confidence = 93%.