The Improvement Study On Streaming Data Clustering Algorithm And Swallows Swarm Optimization Algorithm

With big data applications popular, more and more people pay attention on data mining. Streaming data mining is an important object in data mining field. Because of the amount of data is huge, it can not be stored all the time. Streaming data processing based on the outline of K-means algorithm is a species more effective flow of data clustering algorithm, it can efficiently processing concept drift problems in streaming data. but the k-means algorithm select the initial point by the way of random, which is a disadvantage of k-means algorithm. the paper improved K-means algorithm is the initial point selection method, Experimental results show that the improved K-means algorithm has a better understanding of the effect of data stream.Mehdi Neshat, who swallows swarm algorithm proposed in 2011, will put the Swallows into the role of leader Swallows,exploere swallows and aimless swallows. while the exploring swallows dynamically adjust their speed parameters according to their state at the time,which enhance the flexibility of population, convergence speed, a good optimization results, in this paper the optimal increase in exploration mode and levy flight mode, and in order to avoid falling into local optimum individual roaming, aimless individuals associated with increased weight factor, effectively adjust the proportion of social cognition and Cognitive roaming individuals, proved by experiments,This paper improves swallows swarm algorithm has a good effect.Particle swarm optimization is a common intelligent optimization algorithms, thinking it is simple, easy to understand, fast convergence, but there is a structural not strong, easy to fall into local optimum problem. this paper proposes a multi-mode of operation improved particle swarm optimization, it also coordinated local and global optimal running state, the performance of the proposed improved algorithm has been greatly improved.