DEVELOPMENT OF LIBRARY FOR ANALYSIS MOVING POINT DATA IN STREAM ENVIRONMENT

• Main Author: Halasan, Rizki
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43424
• Institution: Institut Teknologi Bandung

Moving point data in stream environment is point data that collected in real time and its position likely change over time. The quantity of data stream is increasing so that more information will be obtained. To obtain that information, acquisition, processing, and analysis of that data is needed. In this final task already developed a library that have functionality for application to do acquisition, processing, and analysis of moving point data in stream environment. Main focus of this library is in the analysis part that consist of trajectory clustering and trajectory prediction. Trajectory clustering algorithm that used is CuTiS. Trajectory prediction algorithm that used in LinearRegression. The analysis part of this library has extensibility characteristic, new algorithm can be added for trajectory clustering and trajectory prediction. Data acquisition purpose is to collect moving point data in real time. After the data successfully collected, it is processed to change its model data to defined one because data streams has each data model. The processed model then analysed. The analysis part in library consists of trajectory clustering and trajectory prediction. Trajectory clustering is grouping trajectories that have similarities. Trajectory prediction is predicting the future coordinate moving points based on previous coordinate. The result of trajectory clustering and trajectory prediction represented on map visualization. Extensible characteristic on this library proved by the new algorithm that successfully added for trajectory clustering and trajectory prediction.