Designs Of Mobile Crowd Sensing System For User Behavior Recognition

In the process of user behavior recognition research,user behavior data is an important cornerstone for the research of the project.The traditional manual data collection methods have problems such as large errors and low efficiency.With the rapid development of Internet of Things technology,communications technology and smart mobile devices,the emergence of Mobile Crowd Sensing System as a new data collection method has been promoted.The Mobile Crowd Sensing System is based on the user's smart mobile device as the basic sensing unit,the use of communication technology to form a Mobile Crowd Sensing network,the server to achieve the distribution of sensing tasks,and the data collection of large-scale and wide-coverage is completed through a large number of users' mutual cooperation.This thesis designs a Mobile Crowd Sensing Prototype System for user behavior recognition.It elaborates on the construction of Mobile Crowd Sensing Systems for user behavior recognition and the process of analyzing and identifying user behavior data using Naive Bayes algorithm.The system client is based on the Android platform and the server uses the Bmob backend cloud.The client contains seven functional modules,including registration,login,task reception,map,positioning,application keep-alive,task participation,feedback,sharing,and data display.The server side mainly includes data maintenance and task push.In order to ensure that the system can be used stably and normally,the system has been rigorously tested.The test results show that the system has good robustness,reliability and practicality.Based on Naive Bayes algorithm,the thesis trained the user behavior recognition model by using the data which collected by the Mobile Crowd Sensing System,and realized the classification of learning state by using the user's current time and place information,and the classification accuracy rate on the test set is as high as 80.5%.The system has efficiently completed the task of collecting and identifying user behavior data.The data collection function can also serve other follow-up scientific research tasks and contribute to large-scale data collection.