项目名称: 数据特征驱动的清洁能源市场预测理论方法及其应用
项目编号: No.71301006
项目类型: 青年科学基金项目
立项/批准年度: 2014
项目学科: 管理科学
项目作者: 汤铃
作者单位: 北京化工大学
项目金额: 21.7万元
中文摘要: 鉴于清洁能源在解决我国能源安全与环境污染双重难题中的重要作用,本项目拟基于"数据特征驱动建模"思想,开展清洁能源市场预测理论与方法的创新研究,以提高其预测精度。具体的,本项目主要内容包括:1)在有序梳理各种数据特征(包括时间序列数据特征与多元数据特征)及其互相关系的基础上,构建一套全面、系统的清洁能源市场数据特征检验方案,以深入探索清洁能源市场运行机理;2)在此基础上,构建数据特征驱动的清洁能源市场预测模型,确保模型构建紧扣清洁能源市场各种数据特征,并有效体现其主要数据特征,不管在解释变量选择、函数形式设计,还是估参技术确定等方面。本项目创新点在于:理论上,引入"数据特征驱动建模"概念,构建一套系统的数据特征驱动的清洁能源市场分析与预测框架,是复杂系统分析与预测研究中理论与方法的创新;应用上,开发应用软件,研究成果可直接应用于我国及国际清洁能源市场分析与预测中,为相关管理部门提供决策支持。
中文关键词: 预测;能源市场;数据特征;多尺度分析;能源政策
英文摘要: From the perspective of energy safety and environmental pollution, particularly in China, this research tries to propose a novel forecasting methodology for clean energy market based on the concept of "data-characteristic- driven modeling", in order to improve the prediction performance. Specially, through carefully analyzing the interrelationship across different data characteristics for both unitary and multivariant data, a novel integrated data characteristics testing scheme will be built for clean energy market. Secondly, a novel data-characteristic- driven forecasting methodology will be proposed accordingly, where related technologies are carefully formulated based on data characteristics of clean energy data in terms of independent variable selection, function design, parameter evaluation. Generally, the innovations of this research can be summarized into two aspects: theoretically, based on the concept of "data characteristic" and "data-characteristic-driven modeling", this research tends to propose a novel prediction methodology for clean energy market, which makes innovations in the research fields of complex system analyses and prediction; from the perspective of application, the proposed methodology could be a useful tool for forecasting clean energy markets, providing helpful support for related de
英文关键词: Forecasting;Energy market;Data characteristic;Multi-scale analysis;Energy policy