System and method for mining workplace-residence position tag based on mobile communication signaling big data

Abstract

The invention discloses a system and a method for mining a workplace-residence position tag based on mobile communication signaling big data. The key point of the technical scheme is as follows: the system comprises a user individual activity sequence establishment unit used for establishing user individual activity sequences, a grid generation unit used for dividing position information of a cell into uniform grids, a grid cluster analysis calculation unit used for calculating occupied dense grid clusters in appointed time frame by calling a grid cluster analysis calculation algorithm, a density cluster analysis calculation unit used for calculating mass centers of the occupied dense grid clusters by calling a density cluster analysis calculation algorithm, and a workplace-residence position tag generation unit for generating the workplace-residence position tag of the user. The system and the method have the advantages as follows: 1) any assumption on the user behavior sequence data distribution is not required and the obtained result through analysis is not related to data record processing sequence, so that relatively high objective degree is achieved; 2) sub spaces with cluster probability are found automatically; and 3) the complexity of data ETL is lowered.

Description

technical field [0001] The present invention relates to the technical field of mining location tags based on big data, and more specifically relates to a system and method for mining location tags based on mobile communication signaling big data. Background technique [0002] With the continuous development of user terminals, business applications and market environment, more accurate marketing, business development and enterprise applications need to know the user's work and residential location information (job and residence location), as a mobile phone with a penetration rate of 70% The mobile terminal, with its portability at any time and the regular / non-periodical interaction characteristics of network terminals, makes mobile communication signaling interaction data and user behavior data the first choice for analyzing user location tags, and because of the mobile communication signaling data The location information is marked with the wireless cell (CI), and the averag...

AI Technical Summary

Problems solved by technology

[0004] However, at present, the scale of mobile users of provincial operators has reached tens of millions, and the number of wireless base station cells has reached hundreds of thousands. With such a huge number of users and deep coverage level, the above method is used to analyze the location tags of users. , obviously cannot meet the requirements in terms of timeliness and economic cost

[0005] In addition, when mining and analyzing users' job and residence location tags based on mobile communication signaling big data, it is necessary to introduce time series and business behavior data to model and analyze user behavior and mobile location trajectories, so that the data has a larger dimension, while The TopN statistical analysis method cannot achieve ideal results in the analysis of large-dimensional data

[0006] The cluster analysis method can support multivariate statistical analysis of massive data, and is often used for data mining of the user's location track point entry and exit rules. However, for the mining of job and residence locations based on mobile communication signaling big data, Due to the lack of prior knowledge and predictable models for reference, only one user can collect thousands or even tens of thousands of data within a statistical period (two weeks or one month), thus forming a high-dimensional Data collection, as we all know, affected by the "dimensional effect", many clustering methods that perform well in low-dimensional data spaces often fail to achieve ideal clustering effects when applied to high-dimensional data collections

Images