A Study on Mining and Analysis of Trajectory Databases with Multi-dimensional Feature Sets

Abstract

AbstractThe new means in sensing technology, high end mobile devices with wireless communication embedded with GPS, enable large collections of moving objects with ordered information called trajectory data. Trajectory data sources are currently being piled and managed in scores of application domains. There is a need to put efforts on analyzing large scale trajectory data from a variety of aspects. With the availability of advanced communication devices, applications and machinery that capture data the society adopted the electronic search process for their daily needs. The data of moving objects, changing values, sequencing events with underlying trajectory nature can provide large information potential if is processed intelligently. The research in trajectory data mining is limited the trajectory datasets with spatiotemporal information only. But today there is the availability of trajectory data with other attributes like color and type (for vehicles) monitory value and purchase value (for transactional data), blood pressure reading and blood sugar reading (for medical data). A few attempts to mine such datasets are lagging to get scalable and portable methods to analyze and mine trajectory data. The present study tried to propose better means for multi-attribute trajectory data analysis and mining in terms of reduced effort. The proposed framework can provide interesting and useful knowledge that can solve some of the problems of real life through data analytics. KeywordsTrajectoryData miningData analyticsClusteringClassificationMulti attribute data