An Improved Decision Tree Model for Forecasting Consumer Decision in a Medium Groceries Store

This study aims to improve decision tree methods for forecasting consumer decision on the basis of known demographic and purchasing behavior in a medium grocery store. These characteristics include age, gender, spending, and shopping behavior. In this study, transaction data at grocery stores from August 2022 to October 2022 are used to optimize the decision tree algorithm. Different decision tree criteria are compared, and the model parameters, such as the maximum depth of the tree and the minimum number of samples required to split a node, are refined. Ultimately, the decision tree algorithm is optimized. The proposed method improves classification and prediction results for the consumption behaviors of customers, which enables businesses to better develop sales strategies. The accuracy of the model after it was optimized was 70.96 percent, and other performance measures saw improvements as well. After examining sales information spanning a period of three months, the decision tree model was able to forecast the purchasing patterns of Cooking Oil ‘S’ customers. This research can be utilized to learn about the shopping habits of customers in order to develop targeted advertising strategies.