Research On Grid User Behavior Analysis Based On Data Mining

At present,power grid enterprises is in stepping up the critical period of "two transformations",innovation is still our persistent pursuit,To meet the challenges of big data,embrace big data,it is an integral part of innovation.Fully grasp the important node era of big data,and change ways of thinking,keep up the rhythm,so that we will all stand taller,and also will enable us to go further.With the development of market economy,our power enterprises gradually from the production-oriented enterprises into business enterprises,marketing concepts followed them into the power industry.How to use the electricity customer behavior analysis and prediction is becoming an important work of marketing in the power grid enterprises.On the other hand,with the recent power grid constantly improve the level of information technology,all types of information management systems of the power grid enterprises are constantly accumulate large amounts of user data.How to use data mining techniques and user data continuously generated to analyze and predict customer behavior is the main problem to be solved in this paper.Most of the existing information management systems can only add,delete,search,modify with a flood of user data,can not tap and utilize the data implicit in deep relationships and rules,but these relationships can not be based on trends and user behavior prediction rules.View of the above situation,this paper focuses on the user behavior analysis applies to grid companies,and realize some classical data mining algorithms.Through data mining techniques,we can analysisbe the large number of user data which left behind by the existing information management systems,dig out deep association rules,and type information into the decision-making.And these information can assist the power grid companies in marketing decisions and improve their customer service levels.The main contents include of three themes,electricity customer segmentation,customer credit rating and prediction of arrears high-risk customers.I have implemented many algorithms including naive bayes classifier,ID3 decision tree,analytic hierarchy process,K-Nearest Neighbor and K-means,which contribute to making detailed analysis and perditions.According to the purpose of maximize the benefits of the power grid companies,combining with behavioral characteristics of the user,we conducted a needs analysis.And after the learning of the related power marketing jargon,and in-depth understanding of the power marketing business processes,collate and analyze a series of formulas,to sort out the required data.Furthurly,data mining algorithms carry out electricity customer segmentation,customer credit rating and predict high-risk customers arrears and other related functions.Comparing to traditional customer behavior analysis of the power grid customers,the analysis of customer behavior based on data mining can improve the accuracy of the analysis of customer behavior,and achieve the quantitative description of the use of the electricity for customers.And also comparing with the analysis carried out by professional sectors,the analysis of customer behavior based on data mining pays more attention to the risk prediction of the customers and the mining of the benefits of large electricity customers.