Comparison of consumer comments on special day discounts with machine learning algorithm

With the acceleration of digitalization, traditional commerce has undergone a transformation, and online shopping systems have assumed a significant role in shaping consumer behavior. In this context, user reviews—especially during promotional periods such as special days—have become a critical factor influencing purchasing decisions. The increasing exchange of information among consumers has elevated user-generated content from a personal guide to a strategic data source for businesses. This study aims to analyze product-related reviews of university students who shop online during special occasions. The data, collected from the Trendyol platform, were processed using the Python programming language and natural language processing (NLP) techniques. The comments were classified as positive or negative, and seven supervised machine learning algorithms—Naive Bayes, Logistic Regression, Decision Tree, K-Nearest Neighbors (KNN), Gradient Boosting, XGBoost, and Random Forest—were used to develop classification models. These models were evaluated using performance metrics such as accuracy, ROC-AUC, and F1-score. Gradient Boosting and XGBoost achieved the highest performance among the tested algorithms. Furthermore, similar performance levels across different product categories supported the generalizability of the models. The findings indicate that sentiment analysis supported by machine learning can serve as an effective tool for improving consumer experience and guiding marketing strategies.