Predicting Sales Using Performance Comparison of Different Algorithms in Regression Algorithms

With the increasing pressure on businesses in today's society, using machine learning algorithms to predict sales has become a way for companies to tap into the market. Regression analysis is used to determine the relationship between two or more variables. When predicting sales, regression algorithms can be used to build a mathematical model that relates sales to other relevant variables (e.g., advertising inputs, seasonal variations, etc.). This model can then be used to predict future sales. Common regression algorithms include linear regression, polynomial regression, ridge regression, Lasso regression, ElasticNet regression, support vector machine (SVM) regression, and decision tree regression. When comparing the performance of these algorithms, techniques such as cross-validation can be used to ensure the generalization ability of the model. In this paper, experiments will be conducted with algorithm comparison functions, and the experimental results found that among dozens of algorithms, Adaboost regressor, Bayesian bidge, and Ridge Regression have the best performance with 86%, 83%, and 83%, respectively. Therefore, this paper chooses Adaboost regressor for the study, and finally verifies with data, and finds that it is basically close to the actual sales.