Relaxation fuzzy c-means clustering algorithm

Abstract

The invention discloses a relaxation fuzzy c-means clustering algorithm. The relaxation fuzzy c-means clustering algorithm includes the following steps: 1) performing optimized division on a sample set according to an RFCM object function minimizing principle; 2) initializing the position and the speed value of a plurality of particles; 3) implementing initialization of fuzzy grade of membership by matching the positional value of particles with the fuzzy grade of membership of samples; 4) according to an iteration formula of a particle swarm algorithm, obtaining an updated fuzzy grade of membership; 5) according to an iteration formula of a gradient method, obtaining a clustering center through calculation; and 6) obtaining an RFCM object function through calculation. The relaxation fuzzy c-means clustering algorithm abandons normalized constraint of the fuzzy grade of membership for a fuzzy c means clustering algorithm, increases the performance of the clustering algorithm for data inclusion and distinguishing, and can expand the fuzzy index m to a range greater than 0 so as to improve universality of the clustering algorithm.

Description

technical field [0001] The invention belongs to an algorithm for unsupervised data classification in the field of data mining, and specifically relates to a relaxed fuzzy c-means clustering algorithm adapted to noise data sets by relaxing the constraint conditions of sample fuzzy membership degrees. Background technique [0002] Fuzzy C-means clustering algorithm (fuzzy C-means clustering, FCM) is the most important clustering algorithm in the fuzzy clustering algorithm, which has a wide range of applications in the fields of pattern classification, machine learning and data mining. [0003] Compared with other clustering algorithms, the FCM algorithm has many advantages, such as the mathematical expression of the model is easy to understand and practical, the optimization solution method is diverse and the convergence theory is rigorous, the algorithm is easy to implement with the help of computer programming, and the fuzzy clustering effect is excellent. However, the FCM a...

AI Technical Summary

Problems solved by technology

[0008] In order to overcome the deficiencies in the prior art, the present invention ensures that the clustering algorithm has an effective ability to reject noise data and outliers, and proposes a relaxed fuzzy c-means clustering algorithm (RFCM), with the purpose of abandoning the FCM The normalization constraint of the fuzzy membership degree of a single sample in the algorithm is converted into a constraint that the sum of the fuzzy membership degrees of n samples is n, and the particle swarm optimization algorithm is used to optimize the estimation of the sample fuzzy membership degree, which not only enables the clustering algorithm to combine noise data and Outlier points can be effectively distinguished from normal data, and it can also maintain the strong clustering performance of the clustering algorithm. At the same time, it can also expand the fuzzy index to m>0, thereby improving the generality of the clustering algorithm for fuzzy index parameters.

Images