Multifaceted Regression Analysis on the Impact of Restaurant Payment Methods

The restaurant industry faces various challenges in maximizing revenue, with one significant factor being the impact of payment methods. The adoption of cashless transactions and digital payment platforms has changed transactions but also introduced complexities. Some customers may face issues with technology, leading to transaction failures or delays, potentially deterring them from returning. Additionally, transaction fees associated with certain payment methods can affect profit margins for both the restaurant and the customer. Moreover, the shift towards online ordering and delivery services further complicates revenue streams, as these platforms often charge substantial commissions even if applying technological tools. Meanwhile, traditional techniques provide a foundational framework, but there are limitations in capturing non-linear trends necessitating the integration of more sophisticated techniques. This research paper proposes a multi-faceted regression approach to the sales transaction's dataset of a fastfood restaurant. Initial exploration involves polynomial regression, neutralizing the basic relationship between payment methods and transaction amounts. To counter overfitting and ensure model stability, Ridge and Lasso regression with polynomial features are implemented as regularization techniques. Moreover, Random Forest Regressor and Stacking Regressor combine predictions from multiple base models using a final estimator. Analyzing the coefficients from the models can provide insights into how each payment method affects the transaction amounts. This amalgamation of methodologies aims to provide an interpretation of the impact of payment methods on revenue, considering complex relationships, potential nonlinearities, and interaction effects within the dataset. Furthermore, the synthesis of proposed techniques offers an effective method for tackling intricate relationships in similar business scenarios.