本项目主要研究面向互联网数据的大宗商品深度学习预测方法。在大宗商品市场影响因素分析与互联网数据挖掘的基础上，搭建深度学习卷积神经网络（CNN）结构，提取互联网大宗商品市场特征。同时，将开源数据与传统统计数据有效结合，建立基于分解重构的长短时记忆（LSTM）大宗商品市场价格预测模型。在此基础上，有效融合CNN特征提取模型，建立大宗商品市场CNN-LSTM预测模型及多模型非线性集成预测模型。此项目对于大数据下新的预测方法的研究以及政府、企业等部门的科学决策具有重要意义。

This project focuses on the study on the commodity market prediction based on deep learning and Internet data. Based on the analysis of impact factors of commodity market and Internet data mining, the project constructs the convolution neural network (CNN) structure and extracts the characteristics of commodity market. At the same time, the open source data is effectively combined with the traditional statistical data, and the price forecast model using LSTM based on decomposition and reconstruction is established. On this basis, the integrated forecasting modesl like CNN-LSTM are established by combining the CNN feature extraction model and LSTM. This project is of great significance for the development of prediction method under lbig data and scientific decision-making of governments, enterprises and other departments.