Combining Bootstrapping and Genetic Algorithm Based on Feature Selection for Franchise Location Prospects Prediction

Abstract

Location selection is crucial in the franchise fast-food industry. A thorough location selection model paired with a proper analytical technique can considerably improve the performance of placement decisions, attract more customers, and raise market share and profitability. Franchise location data sets have an imbalanced class nature. The franchise location prospect prediction performance decreased as a result of the dataset's noisy characteristics. We developed a hybrid approach to improve franchise location prospect prediction performance in this study. It combines Bootstrapping to address class imbalance problems and Genetic Algorithm (GA) to select relevant features in the franchise location prospect prediction. We experimented with four different classification methods (Naive Bayes, C4.5, Random Forest, ID3, Gradient Boosted Trees). The results show that almost all classifiers that use Bootstrapping and GA outperform the original technique. We employ the Confusion Matrix and Root Mean Squared Error (RMSE) to examine the proposed method's performance. The test results demonstrate that the proposed method considerably enhances the franchise location prospect's classification performance.