PULM: Prediction of User's Location using Machine Learning with Markov Model

In today's date, the amount of data generated daily is huge. Similarly, a large amount of user mobility data is generated from mobile GPS, Wi-fi, IMU, and other sensors and devices. Mobility analysis and prediction using this mobility data has attracted a lot of attention recently. This prediction includes the users location and mobility tracking. In this paper, a trajectory prediction system is proposed that will predict a user's future trajectory depending upon his/her previous mobility data, also taking into consideration the data of users having similar mobility data. This system usesan improved DBSCAN algorithm for extracting Stay Points from a user's mobility data and an ExtendedMobility Markov Chain (MMC) model for getting a single user's prediction model. Location Prediction Accuracy is used as the performance measure for the proposed work.