大数据环境下基于计算智能的预测建模技术及其在电力负荷预测中的应用

Big data has revolutionized the modern world, and is now the hottest research topic. The current trend is expected to continue into the foreseeable future. What is of importance is how organizations develop the tools and means necessary for reacting to, and exploiting the increasingly available big data for their advantage. In view of the fact that the well-established prediction techniques have been suffering the lack of integrated profile data view, specific modeling techniques, integrated platform for prediction modeling and whole understanding toward the prediction problems, it is challenging to re-establish computational intelligence based prediction modeling framework with specific considerations of the big data environment having large volumes of high dimensional data and real-time calls on computing. The aim of this project is, therefore, to create novel predictive models based on computational intelligence methods in large-scale and data pre-processing and data infusion analytics, deep support vector machine based forecasting models, and ensemble learning based computational framework for predictive analytics. Load forecasting will be the application domain to verify the proposed big data enabled computational intelligence based prediction systems. Generally speaking, this present study can make attributions in terms of innovative and theoretical development in field of big data driven prediction and provide implications for practitioners.

大数据革命性地推动了社会的发展并成为研究的热点，研发充分利用大数据优势的工具与方法有着重要意义。针对现有预测技术应用研究缺乏大数据环境下对预测数据全景性、预测建模技术针对性、预测建模平台化和预测问题全面性的研究思维，本项目以基于计算智能的预测建模技术和电力负荷预测为切入点，构建新颖的海量、高维、实时特征下的大数据驱动的基于计算智能的预测建模框架，重点研究构建预测建模导向的大数据预处理与分析、深度支持向量机预测模型、基于集成学习的预测模型集成计算框架等基础问题，同时结合电力负荷预测的应用特点与新要求开展应用研究。因此，本研究既具有创新性基础研究的理论意义，同时也表现出良好的应用研究的现实意义。

本项目聚焦“大数据环境下的预测”这一问题场景，从全景数据视角，研究数据的多源融合、多尺度分析等预处理与分析技术；从数据驱动建模视角，研究基于深度学习的预测模型；从大数据应用的平台化视角，设计基于深度学习和集成学习的预测建模集成框架。同时，携手来自国家电网公司华中分部的项目组成员，围绕电力负荷预测、电力电量交易优化、跨区省间清洁能源协同消纳等重要行业应用问题开展深入应用研究，推动预测技术研究的“大数据化”和提升我国清洁能源消纳。项目成果包括10余篇国际国内领域期刊论文、授权发明专利1项以及在湖北省、重庆市、江西省和河南省开展成果应用所取得的显著社会效益和经济效益。