Multi-model fusion based on Stacking: A predictive model for the price trend of natural rubber

Abstract

Natural rubber plays the most important role in manufacturing. It has an important position in the automobile tire industry, sealing industry and daily necessities industry. Therefore, its price trend has a major impact on industrial output and economy. For the manufacturing industry, a reasonable forecast of rubber prices will give enterprises a huge advantage in planning production, reducing costs and improving production efficiency. In this paper, in response to the demand for rubber price forecasts in the manufacturing industry, the relevant factors affecting rubber prices are analyzed, and the natural rubber price data sets of more than 500 sets of variables spanning 10 years are collected. The work related to data preprocessing and feature selection was completed. XGBoost was used as a super model to integrate the Random Forest, Adaboost and LSTM base models to obtain the final Stacking model. And provides an important reference for the development of natural rubber procurement strategy.