基于不同分布假设条件的自回归条件异方差族模型在评估日前电力市场风险价值中的应用比较  

Comparative Research on Application of Autoregressive Conditional Heteroskedasticity Family Model Based on Distributional Assumption Conditions in Evaluation of Risk at Value of Day-Ahead Electricity Market

作　　者：余帆[1] 沈炯[1] 刘西陲[1] 

机构地区：[1]东南大学能源与环境学院,江苏省南京市210096

出　　处：《电网技术》2008年第17期23-28,共6页Power System Technology

摘　　要：以美国宾夕法尼亚-新泽西-马里兰日前电力市场为例,对采用基于不同分布假设的自回归条件异方差族模型评估市场的风险价值进行了比较研究。首先利用自回归条件异方差族模型分析得到了边际电价收益序列的条件方差序列,然后利用风险价值计算公式得到了不同分布假设条件下的风险价值序列,并比较了不同分布假设条件下的风险估计效果。分析结果表明,在一天中的不同时刻,分别建立基于不同分布假设的风险价值模型具有较好的效果。Taking PJM day-ahead electricity market for example, a comparative research on evaluation of value at risk （VaR） of market by autoregressive conditional heteroskedasticity （ARCH） family model based on distributional assumption conditions is performed. Firstly, by use the analysis of ARCH family model the conditional variance series from marginal price return series is obtained; then by using computing formula of VaR, the VaR series under different distributional conditions is obtained; finally, the risk evaluation results under different distributional conditions are compared. Analysis results show that it is better to build VaR models based on different distributional assumption conditions at different times in a day.

关 键 词：电力市场 自回归条件异方差(ARCH)族 T分布 广义误差分布(GED) 风险价值(VaR) 

分 类 号：F407.2[经济管理—产业经济]