Target clustering method and system based on three c-means decisions

Abstract

The invention provides a target clustering method and system based on three c-means decisions, and belongs to the technical field of machine learning clustering. According to the method, a cluster ismodeled as a poise domain, a boundary domain, and a negative domain. According to the method, the target data is distributed to different domains of the cluster according to the relative relation between the center point of the cluster and the target data, the method can be applied to any problem that the clustering boundary is unclear, the application range is wide, and the clustering effect is good. Furthermore, in the calculation of the central point of the cluster, the weight of the target is determined according to the number of the pove domain and the number of the boundard domain to which the target belongs, and clustering analysis can be more effectively carried out on the target without using the empirical weight.

Description

technical field [0001] The present invention relates to the technical field of machine learning clustering, and more specifically, relates to a method and system for target clustering based on three-way c-means decision-making, and is especially suitable for educational resource clustering. Background technique [0002] With the development of data mining technology, more and more target clustering techniques have been applied to category prediction, common application scenarios such as image segmentation processing, biomedical recognition, educational resource classification and so on. Taking the classification of educational resources as an example, according to various characteristics of educational resources: such as type (video, text, exercises, etc.), duration of use (the average length of time the resource is used), and frequency of use (the number of times the resource is used within a semester) etc., several different types of educational resources can be clustered,...

AI Technical Summary

Problems solved by technology

The applicant of the present invention finds that these theories have internal unity, and can be summarized with the three-branch decision-making theory, but the current soft clustering method does not use the three-branch decision-making theory to model clusters; on the other hand, in When calculating the cluster center, different weights are applied to the targets in different domains, and these weights are determined based on experience. The consequence is that the cluster center is very sensitive to the weight value

Images