Research On The Algorithm Of Emotion Community Partition Based On SCAN

Community clustering algorithm,also known as graph clustering algorithm,has always been a hot topic in the field of data mining.For example,the maximum modularization algorithm and SCAN algorithm can effectively separate the closely related community structure from the complex structure diagram.At present,with the rapid development of big data technology,these algorithms can still solve the problem of community division.The research of this kind of algorithm is of great significance to shopping guide,public opinion and so on.However,people have different emotions for a certain object or service.The community structure obtained by this kind of algorithm is not the correct social relationship in real life.Therefore,based on the emotional factors in social networks,people began to study the algorithm of emotional community partition.However,the existing algorithms based on maximum modularity to solve this problem have the following problems:(1)Modular algorithm has a high time complexity,which has been proved to be a NP hard problem.Although the modular algorithm is transformed into a semidefinite programming problem which can be solved effectively after the emotional factor is introduced,its time complexity is still very high;(2)The non transitivity of emotion will lead to clustering deviation.In emotional community,if there are n points with close structure and small emotional difference between adjacent nodes,two points with different emotions may be clustered;(3)Modular partition results will be distributed discretely.Because the modular algorithm can not identify the connection points and outliers between communities,and these points have no close relationship with any sub communities in structure,so these points will be divided into separate emotional sub communities.In order to solve these problems,this paper proposes a new algorithm based on SCAN,which is called SPEC.SCAN is a community partition algorithm which can complete clustering in linear time complexity.When using this algorithm to solve the problem of emotional community partition,it can not only reduce the time complexity,but also effectively identify the special points between communities.SPEC algorithm has the following advantages:(1)By calculating the emotion interval,this algorithm can solve the non transitive problem of emotion value.Therefore,the points in each emotional sub community have the same emotional polarity.Thus,the clustering deviation is avoided;(2)In order to avoid the discreteness of the result,the connection points and outliers between communities are clustered by using the emotion interval;(3)The time cost of emotion community partition algorithm is greatly reduced.In this paper,a lot of experiments on SPEC algorithm are carried out.Experimental results show that the algorithm is correct and has better performance than the existing algorithm.