Fuzzy Clustering Algorithm Based On Hybrid Artificial Bee Colony

Clustering which is applied to data mining is one of the key techniques.With the application of fuzzy theory to fuzzy clustering analysis,it has provided fuzzy processing ability for many data processing with "intermediary" attributes in real life,so it can be widely applied to various fields.The FCM clustering algorithm is a typical representative of the fuzzy clustering algorithm,it has the advantages of simple and easy to implement,fast convergence and strong local search ability.At present,combining it with intelligent algorithm has become the main direction of its research and development.More and more hybrid algorithms have been studied by scholars,and have achieved relatively good clustering results.At the same time,it also accelerates the convergence speed of the algorithm.The swarm intelligence algorithm is a stochastic optimization algorithm which is constructed by simulating the group behavior and biological characteristics of the natural organisms.However,artificial bee colony algorithm as an optimization algorithm of younger swarm intelligence algorithm,because it has many advantages,such as easy implementation,less control parameters and strong robustness,and also it has excellent performance in dealing with optimization problems,it has been accepted and promoted by more and more scholars,and has been applied to many fields successfully.However,the FCM clustering algorithm is sensitive to the selection of the initial cluster center and the noise data,and the global search ability of the artificial swarm algorithm is not strong enough.In view of the above shortcomings,this paper proposed an artificial bee colony algorithm based on continuous distribution estimation by collecting the relevant data at home and abroad.Then,a fuzzy C-means clustering algorithm based on hybrid artificial bee colony algorithm was proposed by combining the improved artificial bee colony algorithm and FCM clustering algorithm.The main contents of this paper are as follows:(1)Research on artificial bee colony algorithm based on continuous distribution estimation:the main purpose of the study is to combine the local "microcosmic" information obtained by the artificial bee colony algorithm with the global "macro" information obtained by the distribution estimation algorithm,and put forward a new nectar generation strategy,so as to obtain a hybrid artificial bee colony algorithm which combined the continuous distribution estimation algorithm with the artificial bee colony algorithm.This made the algorithm improve obviously in the convergence speed,robustness and so on,so that the global exploration ability of the ABC algorithm was strengthened.(2)Research on fuzzy C-mean clustering algorithm based on hybrid artificial bee colony algorithm:In this paper,we use the artificial bee colony algorithm based on continuous distribution estimation to determine the initial clustering center of FCM,which overcame the shortcoming that it is very sensitive to the initial clustering center,so as to speed up the clustering speed of FCM algorithm and improve its accuracy.(3)Research on the application of fuzzy C-mean clustering algorithm based on hybrid artificial bee colony algorithm in Web log mining:By clustering users,we can divide users with similar browsing habits into groups.Using these information is of great significance in providing users with personalized services and e-commerce applications.By clustering the pages,we can classify the page groups associated with the content,and provide valuable information for the rectification of page structure and the design of online search engine etc..