Abstract:

It is very difficult to estimate consumer behavior due to different variables. There are also differences between the online consumer and the traditional ones. While there are studies for the prediction of purchasing behavior of online consumers, there is need for further studies with larger data including different features. Large data is difficult to obtain due to restrictions on private information and causes the analysis systems run for a long time. So, in this study, it is aimed to create a meaningful rule by estimating the purchasing behavior of online consumers with fewer data. After selecting the Fisher Score feature in a current and open database, training and test data were determined with K fold and a rule was created with Decision Tree. As a result, it can be suggested that it is possible to determine the purchasing behavior of online consumers with high accuracy by using a single feature.

Keywords: