Design And Implementation Of User Location Prediction System Based On Spark

At present,intelligent terminals play an increasingly important role in people's lives.Geographic location information can serve the intelligent transportation field,such as prediction of traffic flow and congestion and the design of solution,navigation path planning of vehicles running,etc.It also serves social networks and e-commerce,and analyzes user behavior and location-based recommendations based on the user's location.In geo-based services,it is important to extract the user's location of interest from the user's trajectory.Prediction of the user's next location based on the extracted location is also a hot research topic today.The subject fully investigated the current location prediction research at home and abroad,focusing on studying and researching the Markov model,random forest model and deep belief network model.Aiming at the problem that the traditional prediction model considers the single dimension,this paper applies the belief network model to the position prediction,and from a set of solutions,from the preprocessing of the user's trajectory to extracting the user's interested position,then to the prediction based on the extracted location.Features introduced by the model include the user's access to the previous location,the user's visit time,the user's residence time and the geographic attributes of the geographical location.Compared with the traditional location prediction model,the model has a 10%improvement in accuracy.Based on the open source distributed computing framework Spark,this topic applies the location prediction model based on deep belief network to practice,and completes the design and implementation of the prototype system.The system is divided into data analysis module,position extraction module,position prediction module,data visualization module and so on,which can realize the model training and data prediction function,and provide a user-friendly interface.Based on the distributed test environment,this paper verifies the indicators of the system requirements,and the system has been tested to meet the requirements.In addition,from the performance point of view,to meet the system stability requirements and operational efficiency.