基于大数据的物资价格预测方法探索

Research on material price forecasting method for big data

通过分析物资价格相关的影响因子,并基于这些物资相关大数据,探索时序预测的方法、简单回归分析模型及复杂的分类回归模型等机器学习算法对电力物资价格的预测,并对模型的预测效果进行评估。结果表明,时序分析的方法效果差强人意,而将时序转换为回归问题,进而使用梯度提升树可以得到较理想的结果。

This paper analyzes the factors related to the impact of commodity prices, and based on these materials related to big data, which explores the method, time series prediction model of simple regression analysis and classification and regression model of complex machine learning algorithm, to forecast the power material price, and the prediction effect of the model is evaluated. The results show that the method of timing analysis is not satisfactory, and the time series is transformed into regression problem, and then the gradient lifting tree is used to obtain the ideal result.