An Offline Merchant Recommendation Method Combining Social Network and Location

Abstract

The invention relates to an offline commercial tenant recommendation method combining social networks and positions, wherein the method is used for telecommunication electronic commerce public servicecomprehensive support platforms. The method includes the steps: preprocessing data, and processing telecommunication data to obtain needed data table; building social relationship networks accordingto call detailed lists of users; performing hierarchical clustering on the social networks by a CNM (customer network management) community discovery algorithm; screening commercial tenants accordingto distance threshold valves based on user positions to obtain candidate commercial tenant lists; analyzing user internet surfing log information, and building two-dimensional preference matrixes of the users and the commercial tenants; recommending the commercial tenants by a collaborative filtering algorithm based on fusion social relationships of the users. According to the method, the social relationship networks are built by the aid of user call information, the social networks are excavated, closely associated user groups are found, the user positions are combined, commercial tenants arerecommended by the collaborative filtering algorithm based on the fusion social relationships of the users, computational complexity is reduced, and recommendation accuracy is improved.

Description

technical field [0001] The invention relates to the field of precision marketing, in particular to an offline merchant recommendation method combining social network and location. Background technique [0002] With the continuous expansion of mobile smart terminal users, the mobile Internet market has entered a stage of rapid development, and a large amount of user data has also been generated. Personalized recommendations for mobile terminal users have also attracted the attention of operators and merchants. Using mobile Internet user data to mine valuable information has become a popular commercial marketing method. The existing personalized recommendation methods of telecom operators mainly obtain user personal data, voice traffic package consumption data, user online log data and user trajectory data, and use these historical data to analyze user behavior, preference and consumption situation, and match marketing goals Features are matched, and the matching results are ...

AI Technical Summary

Problems solved by technology

[0004] The existing personalized recommendation methods of telecom operators only consider the user's personal historical preference and consumption information, but do not consider the influence of the user's social relationship behavior on the user's consumption preference

Moreover, when the traditional collaborative filtering algorithm is used for recommendation, in the case of a huge number of users and items, the scoring matrix is ​​very large and sparse, and the computational complexity is very high.

Images