Strategic Feature Selection for Precision Augmentation in Cement Sales Forecasting

Predicting cement sales is of considerable significance to producers, distributors, and investors simultaneously. Accurate sales forecasting empowers businesses to optimize manufacturing, distribution, and marketing strategies. This work aims to delve into the effect of the feature selection technique in enhancing the precision of cement sales predictions. The author evaluates feature selection techniques with predictive models to determine how they affect prediction performance. This research integrated two classifiers, specifically Logistic Regression and Neural Network, and a feature selection technique, Gain Ratio. The performance of these classifiers was evaluated individually and in combination. The findings indicated that integrating the Feature selection technique with classifiers enhanced precision rates, signifying improved predictive accuracy.