Kernel Clustering Algorithm Based On Fuzzy Information And Its Applications

Clustering is an unsupervised learning algorithm that uses similar metrics to divide a data set into several subsets.Because many data contain fuzzy information in pattern recognition,image processing and data mining,the research and application of kernel clustering algorithm based on fuzzy information has attracted the attention of many researchers' attention.The main results of this paper are as follows:Firstly,for clustering problems with fuzzy information,some fuzzy rules are extracted from the fuzzy information involved in the data set.Then,the fuzzy rules are kernelized by the fuzzy equivalence relations.Namely,the kernel represented by the fuzzy information are obtained.Secondly,the kernel represented by the fuzzy information is applied to the kernel clustering algorithm.That is,the fuzzy information-based clustering algorithm is given.Furthermore,the process of algorithm is given.Finally,two practical problems are used to verify the effectiveness and superiority of the proposed kernel clustering algorithm.Specifically,(1)we use the car evaluation data set to verify the presented semi-positive definite kernel clustering algorithm.First,by sorting and characterizing the car data,the fuzzy information is extracted,and the number of clusters can be obtained by the fuzzy rule formulation.Second,select the appropriate membership function to quantitatively describe and simulate the fuzzy information.By applying the car evaluation data set to the fuzzy C-means clustering algorithm and the kernel clustering algorithm,respectively,the clustering results are evaluated by the effectiveness index.(2)we use the online shopping user behavior analysis data set to verify the presented clustering algorithm with positive definite kernel and negative definite kernel based on fuzzy information.Firstly,due to the diversification of the attributes of this data set,the indicators for the characteristics are redefined.Secondly,by using fuzzy rules,the fuzzy information of the index is formulated and the cluster number and membership function are obtained.Note that the parameters are obtained from the previously collated data.Finally,the comparable results are evaluated by applying the online shopping user behavior analysis data to the fuzzy C-means clustering algorithm and the kernel clustering algorithm respectively.