A new framework for cigarette market state prediction based on LightGBM
Taicheng Wei, Yanbing Liu, Tingbang Yang, Guiyu Zou · 2023
Cigarette market state prediction is of great importance in the tobacco industry. Accurate cigarette market state prediction can help enterprises make decision support, production optimization, supply chain management, and improve sales efficiency and market competitiveness. Traditional prediction methods are not effective in dealing with large-scale data and complex correlations, and the evaluation of the market state lacks representativeness and scientificity.In this paper, an integrated prediction framework is proposed to predict the cigarette market state. This prediction framework is based on a brand-new market state composite index and LightGBM prediction model. The new market state composite index considers the state from different dimensions of market changes. Then LightGBM is used as our prediction tool to predict the index. To evaluate the prediction effectiveness of the models, we compare the prediction results of various state-of-the-art machine learning models. The experimental results show that the new framework based on LightGBM can effectively predict the cigarette market state and obtain better prediction accuracy. This new framework provides a reference for the prediction and analysis of cigarette marketing analytics.