The Application of Tobacco Product Quality Prediction Using Ensemble Learning Method

With the rapid development of modern information technology, the tobacco industry has gradually achieved informatization and digitization transformation. A large amount of data is recorded in the production process and those data contains the information and knowledge which can determine the quality of the product. It's known that the traditional quality test method is not timely and with a large contingency, if the production data can be properly mined and analyzed, the product quality can be predicted in advance. Therefore, based on the mining and analysis of the tobacco production data recorded by the batch management system, this paper proposes a prediction model using ensembled learning algorithm to predict the product quality. The results compare random forest and XGboost method, which shows that the XGboost method has a better performance for quality prediction. This prediction model would be beneficial to guide the future production and improve tobacco product quality.