Research On Clustering Based On Swarm Intelligence Algorithm

Now, with the development of computer technology, such as Cloud computing and the Internet of things technology. There is produced a large amount of data in some fields, such as activities of business, the space of social, the production of business and so on. Because of this information age the technology of cluster analysis has become more and more popular. The fuzzy clustering because of its fuzzy set theory has become widely useful. Now, such as data mining, image processing, traffic congestion and other fields widely used that technology.There are many advantages of swarm intelligence algorithm, such as adaptability, randomness, parallelism and robustness. Because of its advantages, many scholars to conduct research and optimization swarm intelligence algorithm. And data mining, image processing, financial management, and many other fields are successfully applied that algorithm. Aiming at the problem of fuzzy C-means clustering algorithm that it is depending on the initial value of the number of clusters and easy to fall into the local optimization and sensitive to the initial clustering centers. An improved clustering algorithm that combine optimization of the Artificial Fish Swarm Algorithm with fuzzy C-means algorithm is proposed(NAFSA-FCM). The NAFSA-FCM is applied to the traffic congestion in the division of regional.Paper is focused on the research of fuzzy C-means algorithm and the Artificial Fish Swarm Algorithm, in the following is the main several aspects work:(1)Aiming at some defects of traditional AFSA, such as low optimization precision and long running time, proposed a new adaptive artificial fish swarm algorithm based on Log-Linear model and Gauss-Cauchy mutation(The New Artificial Fish Swarm Algorithm, NAFSA). Firstly, the new prey, swarm and follow behavior based on Log-Linear model are proposed to improve the performance of fish swarm algorithm. Secondly, in this new algorithm can adaptive visual and step of artificial fish. Thirdly, Gauss-Cauchy mutation used to keep the individuals diversity and escape to fall into the local extremum. Finally, compared with other algorithms, the results show that this new algorithm has a better effect on both the convergence speed and the stability aspect.(2)To avoid fuzzy C-means falling into local extremum of population, a new clustering algorithm that combine optimization of the improved Artificial Fish Swarm Algorithm with fuzzy C-means algorithm is proposed(NAFSA-FCM). The fuzzy C-means algorithm aimed to the target function optimization to the minimum value. Compared with other algorithms, the results show that the optimization algorithm has a better effect on both the convergence speed and search accuracy.(3)The NAFSA-FCM algorithm is applied to traffic congestion area, because of serious traffic congestion, and the popular of Intelligent Transportation Systems. Compared with other algorithms, the results show that this new algorithm can effectively avoid the noise points and falling into local extremum of population. The new algorithm has improved the classification accuracy of traffic congestion, and has a better job for the Intelligent Transportation Systems with other tasks.