K-means Hybird Clustering Algorithm Based On Improved Artificial Bee Colony Algorithm And Its Application

Along with the rapid development of science and technology, people’s life is filled with all sorts of information, clustering analysis on a large amount of data is necessary when facing the question that how to extract useful knowledge from the mass information. K-means is a classical clustering algorithm based on partition. The algorithm is widely used because it has the advantages of simple operation, easy to work well with large data set.At the same time, swarm intelligence technology develops increasingly and is applied to improve the performance of the clustering problem owing to its excellent characteristics. Artificial bee colony algorithm is a swarm intelligence optimization algorithm for its main feature such as simplicity and realizing easily, strong global searching ability, less control parameters required. So based on the previous research, firstly, improve the artificial bee colony algorithm, secondly,combined with the K-means algorithm effectively, and finally through the simulation experiment and application of community partition in complex network, show the effectiveness of the improved algorithm, the specific work is as follows:(1)initializing the population in artificial bee colony algorithm is extremely important,initialization is directly related to the iterative space-time late algorithm complexity. Aiming to the existing question about initializing random in artificial bee colony algorithm, this paper proposes a max-min distance algorithm to apply for population initialization, overcoming the blindness and randomness of the random initialization problem.(2)aiming to the problem of slow convergence in the later iterations, this paper proposed a kind of location update formula combined with the global influence factor. The improved formula not only has a strong mining capacity, but also enhance its ability to exploring. And in order to combine with K-means algorithm efficiently, a improved fitness function based on the process of K-means clustering algorithm is proposed. the new function can accurately guide the population,improve the robustness of the artificial bee colony algorithm.(3)the hybird clustering algorithm combine the improved artificial bee colony algorithm with K-means algorithm, it is used to overcome the defect of K-means algorithm global search ability and depend on the initial center point.(4)applying the hybird improved clustering algorithm to solve the problem of complex network community division, transforming the community partition problem into clustering problem. The simulation experiments were conducted on the three dataset including Karate,Dolphins and Football. The experimental result indicate that the application on complex network is rational..The experimental results on some data sets indicate that the K-means hybird clustering algorithm based on improved artificial bee colony algorithm has strong global searching ability and stability. The clustering accuracy and convergence speed has been greatly improved, and the improved algorithm can solve the problem of community division efficiently.