User layering method and prediction method based on power consumption complaint behaviors of users
A forecasting method and user technology, applied in forecasting, neural learning methods, data processing applications, etc., can solve problems such as difficult to discover business rules, difficult to abstract mathematical models, etc., to achieve the effect of active response, easy convergence, and fast training
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0045] Such as figure 1 As shown, this embodiment provides a user stratification method based on user complaints about electricity consumption, including the following steps:
[0046]S1, obtain the data set D of user behavior, carry out the initial layering, and obtain the index matrix A; user behavior includes electricity consumption behavior and complaint behavior;
[0047] In this embodiment, an algorithm with low time complexity such as Max-Min Distance (longest and shortest distance) is used to generate each partition of the user's power consumption behavior and complaint behavior, which can save a lot of time when generating the initial label result.
[0048] Carry out the initial hierarchical stratification, specifically:
[0049] S101, set the initial clustering number H, randomly select a point from the data set D as the first center point p 1 ;
[0050] S102, for each point, find the nearest neighbor of the point from the selected center point, and record the dist...
Embodiment 2
[0070] On the basis of the first embodiment, this embodiment provides a prediction method based on user complaints about electricity consumption. Embodiment 1 Extracts and classifies various behaviors of electricity users through a hierarchical label generation method, completes the classification of electricity user complaint behavior labels, and establishes a user label system that can form a power marketing business portrait. Such as image 3 As shown, the principle of this embodiment is based on the classification of electricity users in Embodiment 1, and the user portrait can be constructed according to the classification, and based on the user's electricity complaint behavior, a joint graph convolutional network is proposed to complete the prediction of the user's next complaint content , to achieve active response to user demands, thereby improving customer satisfaction. Such as Figure 4 Shown is a schematic diagram of the architecture of the joint graph convolutiona...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More