New Approach for Spatiotemporal Data Mining: Combining Spatiotemporal and Visual Data Mining for Traffic Analysis

Abstract

Most of the large databases, which are currently available, have a strong spatiotemporal component, particularly for, transport databases. Spatiotemporal data mining is a complex process, since it takes into account the specificities of spatial and temporal information. Unlike the spatiotemporal data mining process, the visual data mining one uses visualization for exploratory purposes at all levels of the knowledge discovery process. The automatic and visual extractions of data have their respective advantages. Computers can process large amounts of data much faster than humans, while humans are able to recognize objects and visually explore data much more efficiently than computers. This article proposes a new process for the spatiotemporal data mining which integrates visualization as a confirmatory and exploratory tool in the drilling process of large spatiotemporal databases. This consists in combining the spatiotemporal data mining and the visual data mining process in one and the same global approach. On the other hand, it proposes a methodological study that allows judging the utility of combining the spatiotemporal data mining methods with several visualization techniques. For each applied spatiotemporal data mining method, several visualization techniques have been used. In order to validate the presented approach, this study proposes a case study that applies the proposed spatiotemporal data mining process to traffic analysis in the city of Oran.
Copyright © 2019 Praise Worthy Prize - All rights reserved.