Analytic hierarchy process based power consumer feature tag weight system construction method

Abstract

The invention discloses an analytic hierarchy process based power consumer feature tag weight system construction method. The analytic hierarchy process based power consumer feature tag weight systemconstruction method includes: constructing a user unified data model according to user basic data, using fields for reflecting features and motives of users in the data model as tags of the users to construct a tag system; using a 1-9 scaling method of Santy to give a determination matrix, and constructing the determination matrix according to the degree of reflecting the user motives, of each tag; verifying the consistence of the determination matrix, verifying the reasonability of the model if the determination matrix has the satisfied consistency, otherwise adjusting the determination matrix; using a square root method to solve a maximum feature vector of the determination matrix, performing normalization on the feature vector, and acquiring the weight of the tag. The power consumer feature tag weight system determined by the invention can acquire the optimal portraits of users, can provide a data foundation for clustering analysis of the users, and can achieve accurate marketing ofpower generation enterprises.

Description

technical field [0001] The invention belongs to the field of data mining, and in particular relates to a method for constructing a power user characteristic label weight system based on analytic hierarchy process. Background technique [0002] User portrait technology has been widely used in all walks of life. Securities, telecommunications, e-books and other industries analyze the historical data of user search and consumption, and classify users with similar attributes according to the clustering algorithm to realize insights and predict user needs. The purpose can improve the accuracy of industry services. [0003] In the context of the development of smart grids, big data has become a research hotspot. Power companies have massive data in marketing and management, and with the improvement of data refinement, these data will show exponential growth. The marketing data of electric power companies contains huge value. Many electric power companies have begun to mine effect...

AI Technical Summary

Problems solved by technology

[0004] However, there are some common problems in the realization of accurate positioning of users in the power industry: 1) There are many sources of power marketing data, and there are differences in naming rules and logical structures among the data of various users, and a unified user data model has not yet been established; 2 ) There is a lack of methods to combine user qualitative analysis with user quantitative analysis; 3) There are relatively few studies on user tag weights, and there is no established system

Images