Sensitivity Analysis Of Electricity Customer Outage

The Zhanjiang area for customers based on the actual investigation of electric business,starting from the customer behavior dimension, using decision tree, logistic regression and soon many kinds of data mining algorithm, using SAS software to the electricity customeroutage sensitivity analysis. Using the data selection, data modification, data conversionmethod for data pre-processing; using data filtering, data sampling, data appending data to filla vacancy, and the data partitioning and box transformation method for data processing in thefinal again; the use of logistic regression algorithm modeling, using ROC curve and Liftvalues, fit statistics, the coefficient of Gini evaluation the accuracy and precision. Aiming atfinding a power outage sensitivity high customer business needs and business goals, throughstatistical analysis, correlation analysis of customer groups, whether the process evaluation,and business requirements. The final results of the analysis, applied to the businessdepartment of the practical work and business processes, so as to guide the practical businesstargeted and accurate development.The main content is summarized in this research are as follows:(1) the use of correlation analysis, from the marketing system and electricity customersto collect all relevant information;(2) the use of statistical software preliminary analysis of electricity customers, exportrelated basic data table;(3) using SAS software to build the sensitivity analysis of electricity customer outagemodel of customer segmentation;(4) combined with the actual situation, the use of sensitivity analysis on electricitycustomer outage model tuning SAS software;(5) the use of statistical analysis of electricity customer outage group characteristicssensitivity characterization;(6) curing will power outage sensitivity analysis model, and applied in actual business.The results show that: the use of customer segmentation model in SAS software to build,can be well applied to the marketing system, to solve the marketing system outage sensitivityelectricity customers not originally identified the problem and look into each electricitycustomers blackout sensitivity to identify and effectively in the electricity customer outagesensitivity analysis and forecasting, high sensitivity for power customers in differentmarketing scenarios targeted to differentiated services, reduce the power outages may bringgreater economic loss or serious crisis of public opinion.