Parameter Optimization and Performance Analysis of Composite Substance That Can Prevent Burning Based on Machine Learning

Abstract

In the past 30 years, a substance that can prevent burning has played an important role in reducing the loss of life and property caused by fire. At present, the total amount of antibaking agent is second only to plasticizer and various plastic additives in the world. With an average annual growth rate of 0.5% from 2019 to 2021. The fire protection industry is a regulatory industry and a globally competitive industry. Therefore, the entry into force and gradual improvement of relevant laws and regulations at home and abroad will affect the pattern of the whole fire protection industry. China’s “Twelfth Five-Year Plan” will bring a substance that can prevent burning into key development industries and form a strategic alliance for technological innovation of green substance that can prevent burning industries. It provides a policy platform for the development of a substance that can prevent burning in industry. Firstly, we introduce several common a substance that can prevent burnings, then we use machine learning method to establish prediction model and to analyze the performance of composite combustibles. And draw the conclusion: composite material is better than the other two composites; among the six machine learning algorithms, the gradient boosting regression (GBR) model has the best prediction ability, followed by the extra tree regressor (ETR) model and the random forest regressor (RFR) model. Compared with the above three integrated algorithms, Ridge, Ada Boost regressor (ABR), and Lasso regression algorithms have relatively poor prediction results.