Finding Frequently Occurring Paths of Flying Objects Using FP-Tree Approach

During the last two decades frequent pattern mining has been an emerging and active field in data mining research. In this work the process of applying algorithms to extract specific patterns from data sets from a particular domain plays a significant role. Ear-lier methods to mine frequently occurring paths of objects in spatial–temporal databases adopt complex algorithms and require many data-base scans. In this paper a modified frequent pattern tree (FP_tree) algorithm is used to find the frequently occurring paths of flying objects by two phase process. Initially, the flying object paths data-base (2D) is scanned to generate the condensed smaller data structure called FP_tree so as to compress the data and make traversal fast. Finally, the database is scanned using a partition based divide and conquer method to decompose the mining task into a set of smaller tasks for FP_tree based mining frequent paths of flying objects. The proposed algorithm outperforms the mining frequent trajectory pat-terns in spatial–temporal databases approach in terms of number of scans and candidate set generation. This may be applied to interesting game domains and find the frequently occurring paths of a ball shot by a player in some games which follows a trajectory path.