Link prediction method in social network

Abstract

The invention discloses a link prediction method in a social network, which belongs to the field of data mining, and comprises the following steps of: preprocessing, constructing context sensing embedding, constructing a link prediction candidate set, and combining the following text sensing embedding with the link prediction candidate set for similarity calculation. Context scenes with similar behavior modes of users are found by constructing context sensing embedding, so that information missed by a previous traditional method is captured, potential social influence is instantiated, link prediction accuracy in a social network is improved, accurate representation of the users in the social network is obtained by the link prediction accuracy and node structure information, the accuracy and performance of social network link prediction are improved, and the problems that only influence interaction between nodes and adjacent nodes is considered in a traditional method, and hidden socialinfluence information is not mined from the level of the whole social network are solved.

Description

technical field [0001] The invention relates to the technical field of data mining, in particular to a link prediction method in a social network. Background technique [0002] With the popularization and development of social networks, social networks such as Facebook, Twitter, and Weibo are playing an increasingly important role in people's lives, and people's social methods and communication concepts have also changed accordingly; social networks contain A large amount of information provides researchers with an unprecedented opportunity to explore human behavior patterns; research on social networks has always been a hot topic in the field of data mining; different from general data mining, data in social networks has its own Heterogeneous and diverse characteristics, for example, most people have only a few links, but a small number of people have far more links than others; another example is the community structure, there will be many small groups in the social networ...

AI Technical Summary

Problems solved by technology

[0004] The traditional link prediction method has a similarity-based method, which mainly calculates the similarity of each pair of nodes to determine whether they are connected; an important premise of this method is that the node pair with higher similarity is more inclined to There are links; similarity can be divided into two main types, one is measured by node attributes, such as location, age, and gender; the other is similarity measured by topological features, such as the number of common neighbors; these methods Based on the principle that node pairs with greater similarity tend to have potential relationships; the calculation of this method is relatively simple, and the potential connections can be quickly inferred; however, these methods have obvious defects; first , which treats node attributes and topological features as independent, and thus lacks consistency and a suitable method that can use both types of related features to define similarity; secondly, because the nodes are different, the predictive ability of the model may be there is a big difference

[0005] To overcome this shortcoming, some probability-based strategy methods have been proposed, which take node attributes and topological feature attributes as random variables to learn the probability distribution of links from observed networks, and use the assumed latent structure to estimate the link relationship. ; but the calculation of this method is complicated, and it is difficult to obtain the very necessary link distribution based on the probability model in real applications, so there is an urgent need for a link prediction method in social networks to solve the above problems

Images