Mobile phone user networking period prediction method

Abstract

The invention discloses a mobile phone user networking period prediction method. A periodicity calculating method includes that the time sequence data of the user networking state at a fixed time interval are regarded as discrete signals, and the periodicity of the user networking state sequence is analyzed by the Fourier; the difference calculation method includes that by means of the mobile phone networking data, and the preferences of mobile phone user groups for different time periods is explored, the number of people in 24 hours a day is hierarchically clustered to obtain three networking time periods, namely, a high-frequency period, a low-frequency period, and a transition period. The prediction method proposes two mixed Markov prediction models including a bidirectional Markov mixed model and a Lift-Markov mixed model based on the previous features, predicts the networking periods of the user from different angles, and combines under a probability framework. According to the invention, the data acquisition manner is simple, the cost is low, the model structure is simple, and the accuracy is high.

Description

technical field [0001] The invention belongs to the technical field of mobile phone surfing prediction technology, and relates to a mobile phone surfing behavior prediction method, in particular to a hybrid Markov prediction method for calculating the periodicity and difference of mobile phone users' surfing behavior and predicting the surfing period. [0002] technical background [0003] At present, there are relatively few technologies for mobile Internet access prediction. One method is to make statistics and predictions on the distribution and preference of wired Internet access time. However, in the current society, mobile Internet access has become one of the main ways for users to access cyberspace, and The time distribution and preferences of wired Internet access and mobile Internet access are very different, so this method cannot be fully applied to the prediction of mobile Internet access. The second method is to directly model and predict WAP access and web page ...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem that the existing mobile phone users' online behavior is difficult to predict, the present invention proposes a hybrid Markov prediction method to predict the user's online behavior

Images