Method for automatically identifying optimal K value of elbow rules

Abstract

The invention relates to a method for automatically identifying an optimal K value of elbow rules and belongs to the technical field of unsupervised clustering learning in machine learning. The methodcomprises steps that the to-be-searched optimal K value range [1, M] of the K-means clustering algorithm is set; the average distortion degree corresponding to the set range cluster number is calculated; 0-10 regularized conversion of the average distortion degree acquired through calculation is carried out; the average distortion degree after 0-10 regularized conversion and the range [1,M] are encapsulated into data pairs; the Cosine's theorem is utilized to solve angles among consecutive three encapsulated data pairs; the minimum angle is acquired; the minimum angle is utilized to acquire the optimal K value. The method is advantaged in that on the basis of K-means algorithm and the elbow rules, through 0-10 regularized conversion of the average distortion degree acquired through the elbow rules and further calculation carried out through utilizing the Cosine's theorem, the optimal K value in the designated range is lastly acquired.

Description

technical field [0001] The invention relates to a method for automatically identifying the optimal K value in the elbow rule, in particular to a method for combining and utilizing the K-Means clustering algorithm commonly used in unsupervised learning in machine learning, the elbow rule and the rule of passing 0-10 Regularize the average degree of distortion, package the 0-10 regularized average degree of distortion and the corresponding number of clusters into data pairs, and use the law of cosines to find the angle between three consecutive data points for automatic identification The method of the optimal K value estimated in the elbow rule belongs to the field of unsupervised clustering learning in machine learning. Background technique [0002] With the development of information technology and more and more data generated by people, human beings have gradually entered the era of big data. People use machine learning to study big data, and then acquire new knowledge or...

AI Technical Summary

Problems solved by technology

[0004] However, the K-means clustering algorithm also inevitably has shortcomings: the appropriate number of clusters cannot be determined in advance, resulting in low cluster quality

However, the elbow method often needs to roughly estimate the relatively reasonable number of clusters through observation, and the elbow method needs to use observation to find the optimal number of clusters, which affects its use in automated systems. and its promotion

Images