Method for constructing feature database and system and device thereof

Abstract

The invention provides a method for constructing a feature database. The method comprises the steps that third-party account information and network information released within preset time are acquired; interactive information between third-party accounts is analyzed so that an information transmission network based on the interactive information is generated; all maximal strongly connected components based on the information transmission network and all the maximal independent sets in each maximal strongly connected component are acquired; the least maximal independent sets are selected from all the maximal independent sets in each maximal strongly connected component; de-duplication is performed on points included in all the least maximal independent sets of all the maximal strongly connected components and then a vertex set is formed through merging; and the third-party account information and vertex set information are merged so that a feature information base is generated. The invention also provides a system and a device corresponding to the method. Marketing of different types of projects or activities can be flexibly supported so that marketing resource is saved and construction cost of a marketing model is reduced, and selection of new users of the operation level is also facilitated.

Description

technical field [0001] The present invention relates to the field of communication technology, and in particular to a method, system and device for identifying and analyzing social networks. Background technique [0002] In the entertainment performance industry, especially in the online ticketing industry, the search for target marketing customer groups is generally based on the determination of the project or event scene, and screening based on the history of viewing related projects or similar projects. This method does not take into account the limited number of ticket resources, nor does it consider the migration of user interests. At the same time, it cannot accurately acquire new users, which eventually leads to problems such as poor marketing effects and waste of resources. [0003] Existing technical solutions, such as application number 2015104847492, a method and system for constructing and analyzing social networks, disclose a technical solution for building soci...

AI Technical Summary

Problems solved by technology

This method does not take into account the limited number of ticket resources, nor does it consider the migration of user interests. At the same time, it cannot accurately acquire new users, which eventually leads to problems such as poor marketing effects and waste of resources.

[0003] Existing technical solutions, such as application number 2015104847492, a method and system for constructing and analyzing social networks, disclose a technical solution for building social networks based on stable communication, but this solution cannot effectively disseminate messages, such as:

[0004] Due to work reasons, users A and B often have phone calls and text messages. However, in real life, if A finds a traffic discount package, but B does not know, it is often because the users are in different regions and the relationship is not enough. Because of intimacy or not having the same interest, A will not take the initiative to inform B of this preferential policy, so stable communication cannot form the basis for effective communication, so the way of communication and marketing is often ineffective

[0005] Secondly, the technical solution based on the complete subgraph search algorithm disclosed in the above patent application is based on a simple undirected graph. The complete subgraph itself requires that any two users in the graph have a stable communication relationship. This requirement itself is too restrictive. In practical applications, most of them are two-point and three-point complete diagrams, resulting in no actual reduction in the marketing user base, failing to achieve the purpose of effectively utilizing limited marketing resources

[0006] Thirdly, the application of the AHP hierarchical analysis algorithm in the above-mentioned patent application is limited to the intercommunication of users in the telecommunications field, so the high-impact users screened out are only high-value users, not high-spreading users themselves, that is, factors considered by itself and stability There is no actual difference in the factors considered in the relationship

Images