Forecast of Coal Price based on Random Forest and LASSO Regression

Yiwen Xu

《Frontiers in Economics and Management》 1卷11期 (2020/11) Pp. 260-266

https://doi.org/10.6981/FEM.202011_1(11).0036

摘要

In view of the coal price influencing factors, first select 7 factors and their data that may affect coal supply and demand, adopt a random forest model, Take the average of each influencer's contribution to each tree, and quantitatively study the degree of influence and value order of carbon prices. Then through the visual method and ten-fold cross-validation, the best penalty coefficient is obtained according to the minimum mean square error, and the LASSO regression model of coal price on influencing factors is established to predict the monthly coal price; Forecast coal prices. Secondly, taking the sudden factor of the new corona virus as an example, by performing multiple regression analysis on the data before the occurrence of the new corona virus, the new equation is introduced to modify the dummy variable, and the revised prediction equation is obtained. Finally, combined with today's social development trends, five points of policy recommendations are proposed to ensure the steady development of China's coal market.

關鍵字

Coal price ； Random forest ； LASSO regression ； Polynomial fitting ； Prediction

參考文獻

Xu Yuhang, Xu Yao. Analysis and prediction of factors affecting China's coal market price based on VAR model[J]. Coal Economic Research, 2012, 32(9): 55-58+67.

Cao Shenfu. An Empirical Study on the Change Trend of Coal Price in my country and Its Influencing Factors [D]. Xi'an University of Science and Technology, 2010.

Bai Zhongcai. China's coal demand analysis and price forecast [D]. Fudan University, 2008.

Jiang Lianhong. Correlation analysis and prediction of factors affecting coal prices [D]. Dongbei University of Finance and Economics, 2007.

Zhang Huan, Jiang Zuobin. The establishment and forecasting analysis of the ARIMA model of China's coal price[J]. Industrial Technology Economy, 2007(7): 102-105.