Application Implementation Of K-means Algorithm Based On Wolf Pack Algorithm

Data mining is a hot topic especially in the field of artificial intelligence and database and many researchers pay much attention to it.Traditionally,data mining refers to the non-trivial process of researching and mining potential and unknown useful information from the massive data of existing database.With the continuous development of science and technology,in many fields such as machine learning,visualization technology,artificial intelligence,pattern recognition,statistics and database,people need to analyze and make decisions based on previous data cases.Providing decision support is one of the significance of data mining.Based on the existing enterprise data,data mining can effectively carry out automated analysis,induction,reasoning and mine potential laws or patterns from massive amounts of data.Thus,data mining can help enterprise decision-makers timely adjust market development planning,avoid risks,and then make correct decisions.In the various methods of computer data analysis,cluster analysis is a very important subject in data analysis,and also a key method in data mining.Clustering analysis is an important statistical method in the study of classification,which can dig out the hidden data distribution rules and data patterns in the massive data.It classifies the collection of data objects according to the similarity of the data,so that the data with high similarity can be divided into a kind of cluster,and the data with low similarity can be divided into different clusters.If different methods and means are adopted in cluster analysis,different results will usually be produced.Therefor,if the same set of data is analyzed by different researchers,the types of clusters may be different.K-means in non-hierarchical clustering analysis has become the most commonly used clustering algorithm due to its advantages of simple implementation and fast convergence speed.It takes Euclidean distance as a measurement criterion and on this basis divides experimental data into different categories.However,different selection of clustering center will greatly affect the clustering effect of k-means.If the clustering center is selected randomly,the algorithm will easily fall into the local optimal value and fail to achieve the optimal effect.Macroscopically,since researchers are unknown to the data,they can only set multiple different k values in the selection process of the number of clustering k,and determine the number of clustering through multiple experiments.In addition,different initial clustering centers may lead to different initial search ranges,which will lead to different search areas and different search areas will find different optimal solutions.Therefore,the uncertainty of initial clustering centers makes k-means algorithm lack of good stability.Therefore,this paper mainly studies the following four aspects:(1)When updating the location of fierce Wolf,add the historical optimal location of fierce Wolf;(2)In order to make the algorithm have better global search ability in the early stage and stronger local exploration ability in the later stage,the adaptive step formula is proposed;(3)A k-means clustering analysis algorithm based on the improved Wolf pack algorithm is proposed to optimize the parameters in the k-means clustering algorithm;(4)The proposed method is applied to cluster analysis and compared with other similar methods.