Individual Travel Behavior Prediction Model Considering Mobile APP Usage Characterization

With the rapid development of information and communication technology,mobile phones have become an indispensable part of human daily life,and human activities have gradually extended from physical space to cyberspace.The mobile phone tracking data and Internet traffic data generated by mobile phone users provide a data foundation for researching human activities in physical space and cyberspace.The paper integrates individuals' mobile phone tracking data and Internet traffic data,analyzes the relationship between mobile users' online behaviors and travel behaviors,and constructs prediction model for mobile phone users' individual travel behaviors,which can provide users with more accurate location-based services,help understand human mobility patterns,and provide decision-making basis for urban transportation planning,location selection of commercial facilities,etc.The prediction of individual travel behavior is based on the in-depth understanding of the characteristics of human activities.In the mobile Internet era,the online behavior of cyberspace is inseparable from the travel behavior of physical space.However,the relationship between individual online behavior and travel behavior is less considered in current individual travel behavior prediction models.This paper analyzes the relationship between different travel status and APP usage,divides the travel behavior into two parts,stay and move.And the paper integrates the characteristics of mobile users' online behavior to construct an individual travel behavior prediction model.The main research contents of this paper are listed below.1)A data-driven method is used to analyze the relationship between mobile phone user travel behavior and online behavior characteristics.The paper uses mobile phone tracking data to analyze the spatio-temporal characteristics of mobile phone users and the movement pattern of users between base stations and uses mobile Internet traffic data to analyze the frequency mobile phone users use APP,the size of data traffic,and the difference in usage patterns of different types of APP;A data-driven method is used to analyze the association between the use of different time periods and different types of APPs and user travel behavior.It is found that mobile phone users tend to use mobile phones to access the Internet more frequently when they travel,and the average number of online records is about 5 times higher.After using 14 combinations of navigational travel,weather,and life services,the growth probability of individual travel distance is higher.2)A mobile phone user stay behavior prediction model based on the characteristics of online app usage behavior is proposed.Firstly,the time-space constraint is used to define the mobile phone user's stay behavior.Then,from multi-source data,the paper extracts the individual travel behavior's space-time preference,the app usage characteristics such as the APP combination,Internet traffic,Internet access times and other Internet behavior characteristics and weather information,etc.Feature engineering is done in a time and space crossing way,and the mobile phone user stay behavior prediction model with high interpretability from feature to model is constructed.The prediction accuracy of the model is 80.31%.After the integration of online behavior characteristics,weather and other external factors,the prediction accuracy is improved by 12.08%,compared with the model using only individual travel characteristics.3)A multi-model fusion model of individual travel location prediction based on Markov and machine learning methods is proposed.Firstly,the traditional Markov travel location prediction model is constructed.And the CART,GBDT and RF models are constructed by using the travel behavior feature set,the online behavior feature set and the external factor feature set.Considering the classification probability of prediction results,an adaptive fusion strategy based on frequency distribution graph is proposed.The prediction results of traditional Markov model and machine learning multi-classification model are merged together to obtain the final mobile phone user travel location prediction result.The experiments show that the top1 accuracy and top3 accuracy of the multi-model fusion location prediction model based on histogram are respectively 73.58% and 94.15%,higher than the prediction accuracy of the basic model with the highest accuracy and the vote strategy,and under the time granularity of 30 minutes,the individual travel location prediction is better.