The New Algorithm For Community Detection And Its Application In Commodity Recommendation

Complex network is a graph with non-trivial topological features.As a new and active scientific research field,it has developed rapidly and has gathered researchers from many fields.Community detection of complex networks is meaningful,we can discover the structure and information hidden in the network.Most of algorithms for community detection are based on the topological structure and characteristics of social networks,without considering the relationship between nodes.Based on the study of the connection between two nodes,in this paper,we propose a new algorithm scheme.In this paper,the current research status at home and abroad divided by complex network communities is the starting point,and some existing classic algorithms are introduced.Finally,the new algorithm proposed in this paper is introduced.The new algorithm defines a new indicator for the connection relationship between two nodes.This indicator describes the strength of the relationship between the two nodes.The larger the value of the indicator,the two nodes(two small groups)corresponding to the edge are more closely related,which means that they are more likely to be in the same community.If the value of this indicator is relatively small,the less likely it is to be in the same community.Through this indicator,it is possible to sequentially test whether the two nodes(two small groups)should be divided into the same community.When the modularity gain is calculated to be negative,theoretically,the subsequent division can be stopped.We numerically simulate the new algorithm on scale-free networks and random networks,and conduct experiments on real networks,Comparing the final result with the result of Louvain algorithm division in terms of modularity,the results show the effectiveness of the new algorithm.Later,the new algorithm is applied to product recommendation,and the productcommodity network is constructed for analysis and research on the sales data of a certain platform.The algorithm proposed in this paper can effectively separate the commodity network,divide the commodities with relatively strong relationship together,and extract the more important commodities,which can make a more valuable supplement to the association rules that are now widely used.