发电商滥用市场力行为识别方法研究——基于朴素贝叶斯方法的分析

Research on the Identification Method of Power Generation Abuse of Market Power——Analysis based on Naive Bayes Method

随着国内电力现货市场建设的推进,及时并精准地识别发电商滥用市场力的违规行为,是电力市场平稳运行的有力保障。本文提出一种基于朴素贝叶斯识别发电商利用机组滥用市场力的方法。首先,本文从机组的结构、行为及其影响等三个角度构建滥用市场力行为的评价指标;然后,使用卡方检验与基尼不纯度联合分析指标关联性对评价指标进行筛选,并计算各个指标的权重;接着,基于贝叶斯原理构建朴素贝叶斯分类器,将待测机组归类为所属概率最高的机组类别中,从而识别出滥用市场力的机组;最后,使用国内某区域的电力现货市场数据进行计算,并分析机组的滥用市场力行为。结果表明:该方法具有较高的准确率,有效地减少专家判断的工作量,提高了识别的准确性,对建设有序竞争的电力市场具有积极意义。

Timely and accurately identifying violations of power generation companies’abuse of market power is a powerful guarantee for the smooth operation of the electricity market.This paper proposes a method based on naive Bayes to identify power generation companies’abuse of market power by generating units.First,construct evaluation indicators for abuse of market power from the three perspectives of unit structure,behavior and influence.Secondly,the chi-square test and Gini impurity are used to jointly analyze the correlation of indicators and screen the evaluation indicators,and calculate the weight of each indicator.Next,a naive Bayes classifier is constructed based on the Bayesian principle,and the unit to be tested is classified into the unit category with the highest probability,so as to identify the unit that abuses market power.Finally,using the electricity spot market data of a certain area in China to calculate and analyze the unit’s abuse of market power,the results show that the method has a high accuracy rate.