Research On K-medoids Clustering Algorithm Based On Dynamic Search Of Microparticles And Application Of Wind Speed Clustering

It has broad region,wide cross latitude and long coastline in China,the magnitude of the wind speed data every day,to deal with the plenty of historical wind data become one of the very important link.Wind speed prediction is one of essential Meteorological forecast.The history data of wind speed must be processed before forecasting and clustering analysis was carried out on the wind speed is a very important means of processing.Because the wind has sudden and randomness,this will bring great challenge to wind speed prediction,clustering of wind speed for different categories of specific analysis,dig up the influencing factors of wind speed,eventually be able to provide more accurate data for wind speed forecasting.Firstly,the optimization of the initialization process K-medoids clustering algorithm.As one of the classic algorithm partition algorithm,K-medoids algorithm has the disadvantage sensitive to initial center,clustering accuracy is not high degree of complexity and time.Based on this,this paper proposes an optimized algorithm K-medoids;the algorithm used in the initialization process of granular computing K-medoids clustering algorithm,and the method of selecting the initial centers has been improved: in the choice of K initial poly when class center to the similarity between objects as the basis for judgment,combined with the maximum and minimum cluster center initialization method can effectively obtain optimal or near-optimal cluster center,provide the basis for the next proposed search strategy.Secondly,putting forward the dynamic search strategy based on the microparticles.Under the precondition of optimization of granular computing,the initial center as a base,all objects within the particles to an average distance of the center of the radius,then form a microparticle;In microparticles inside,using the first after nearly far from the center of the principle of search,which can effectively narrow the search and increase the clustering accuracy.Tested on a number of standard data sets in UCI and compared with other improved K-medoids algorithm,the experimental results show that this new algorithm reduce the convergence time effectively and improves the accuracy of clustering.Finally,the improved algorithm is applied to practice,at different periods of the Winds cluster analysis,analysis of the factors affecting the changes in wind speed,to provide a reliable reference for wind speed prediction.