Sales trend analysis via regression tree algorithm (case study: Veruby Store Pontianak)

Abstract

^{This study aims to analyze sales trends in women’s fashion stores in Pontianak using the regression tree method to predict the optimal timing for selling popular products. The data utilized comprises 1,200 transaction entries that record transaction dates, item names, categories, and sales amounts. The research focuses on the “tops” category, employing temporal variables such as day, date, and month as predictors to forecast sales quantities. The regressiontree model was constructed using the fitrtree function, and the results were evaluated based on root means square error (RMSE) and normalized RMSE (NRMSE). The prediction results indicate an accuracy of 77.85%, with higher sales patterns observed on the 13th and specific days of the week, suggesting the influence of promotions or consumer shopping habits. This study provides important insights for inventory planning and promotional strategies in fashion stores, particularly in leveraging identified sales patterns to enhance efficiency and profitability. This analysis can be adapted for various product categories within the fashion industry to respond more timely to consumer demand.}