Analysis And Research On Power Outage Sensitivity Forecast Model Based On Big Data

Looking at the data trends for customers’ complained demand which related to power outages from 2015 to 2018(including channel data of 95598,electronic channels,social linkage and other channels),customers’ complained demand for frequent power outages,unreasonable arrangements for planned power outages,and the duration of power outages accounted for a relatively large proportion,and there was no obvious improvement.The number of power outages is mostly the point of dissatisfaction of customers for 5 years,and repairs and planned power outages for a long time are the last in the satisfaction ranking of power supply quality topics.The continuous development of information technology has provided power supply companies with certain technical support to meet the power demand of different users and realize the scientific power consumption of users.However,limited by the lack of big data analysis methods and tools for power supply companies,There are still some problems in terms of the accuracy,the performance and comprehensiveness of processing massive amounts of data.It is impossible to accurately and comprehensively explore the relationship between the characteristics of customer power outage appeals in various dimensions.This paper provides data and decision support for power supply companies to optimize their business environment and implement personalized power outage services.It conducts research by studying customer-side power big data,and uses random forest algorithms to build a big data platform based on distributed file systems HDFS and Spark.Training the power outage sensitivity prediction model,and finally build a power outage big data analysis model to predict the customer service risk of power outages,portray the profile of sensitive customers of power outages,and realize three hot scenarios to promote the improvement of customer service levels in the power industry,the hot scenarios are: "Frequent power outage complaint risk identification","Planned power outage arrangement optimization" and "Failure outage service remedy "Strategy".