Total Revenue Prediction of A Sports Management Application: Grook Using Machine Learning Models

Grook is a sports management app that provides a solution for people to book sports grounds, courts, and other facilities in their neighborhood for particular time slots to ensure a free space to play. It allows the owners to manage their facilities more efficiently and also aggregate all available sports facilities in the vicinity of the user's location. Every business runs to earn profit and in the case of Grook, the profit earned is the total revenue generated by any given facility on any particular day. Therefore, as indicated in this study, we have trained our model to predict the total revenue generated on a facility using ensemble Machine Learning Techniques in Regression and Classification that are suggested to enhance the existing algorithms. In addition to the existing features in the dataset like sports, days, months, etc., we have also added our polynomial features. We’ve tested algorithms including Simple, Polynomial, Ridge, Lasso, Elastic Net Regression Techniques, and Decision Tree, Support Vector, and Stochastic Gradient Descent Regressors making special models for each technique to get the best results. To make the best prediction of total revenue generated, we have also used the voting method of machine learning which includes a few of the algorithms mentioned. The polynomial features have been tested over several degrees to get an even better understanding of our model’s accuracy.