Finding longest frequent trajectory of dynamic objects using association approaches

Article type: Research Article

Authors: Shaw, Arthur A.^{; *} | Gopalan, N.P.

Affiliations: Department of Computer Applications, National Institute of Technology, Thiruchirappalli, Tamil Nadu, India

Correspondence: [*] Corresponding author: Arthur A. Shaw, Department of Computer Applications, National Institute of Technology, Thiruchirappalli 620015, Tamil Nadu, India. Tel.: +91 948 718 7668; Fax: +91 431 250 0133; E-mail: 405107053@nitt.edu.

Abstract: Data mining is mainly concerned with analyzing large volumes of unstructured data and automatically discovering interesting relationships among them. This information leads to better knowledge and power. Finding Frequent Trajectory patterns are the recently emerging area in data mining. Using the principles of data mining frequent trajectory patterns are derived and knowledge can be obtained out of it. Maximum length of the frequent path will give information about traveling time taken to cross the stretch, alternate path taken to avoid congestion etc. Finding the longest trajectory from the frequent trajectory pattern is a more challenging task and is done efficiently in this paper. Currently existing methods use spatial and spatial-temporal data and follows histogram method to find the frequency of occurrence and clustering method to group the data to find frequent patterns. All these cases spatial or spatiotemporal data may not have any standard approach to find the longest frequent path. This issue is addressed specifically by applying the association based mining concepts in spatiotemporal data. The path derived by applying the modified Apriori and frequent pattern tree methods are compared with a standard graph based method currently available and their performance is analyzed. This approach may be applied to interesting game domains to find the longest frequent trajectory of balls.

Keywords: Data mining, frequent pattern mining, association mining, apriori algorithm, requent trajectory pattern

DOI: 10.3233/IDA-140661

Journal: Intelligent Data Analysis, vol. 18, no. 4, pp. 637-651, 2014