Estimating Peak Sales Of Mechanical Store Item By Using Regression Tree Algorithm (Case Study at CV Kalbar Jaya Motor)

Abstract

This study aims to predict sales patterns for bearing items at CV Kalbar Jaya Motor using the regression tree method. The dataset consists of 4,029 bearing items from a total of 20,830 transactions. The primary objective is to identify the days and dates with the highest sales volumes to optimize stock management and inform marketing strategies. The analysis results demonstrate that the regression tree model predicts sales trends with an accuracy rate of 76.4%. The highest predicted sales occur on specific days and dates, influenced by factors like day, date, historical sales, and customer ordering patterns. For future research, it is recommended to explore ensemble learning method such as Random Forest or Gradient Boosting to enhance the accuracy and robustness of the predictions.