Factor Affecting Selecting Web or Facebook Channel for Online Purchasing

Abstract

this article presents a model to predict selecting channel for online purchasing of electronic commerce between web and facebook. There are 400 volunteers who live in Bangkok and nearest area. They all have an experience to purchase a goods from internet and also have willingness to give an answer for the research. By decision tree model that was created, the result is found that all decision rules are produced by occupation, purchase objective, gender, frequency of purchase, age, and income per month. A measure of the tree’s predictive accuracy is 71.5%. A model’s risk is .285. The cross-validation risk estimate for final tree is .327.