An efficient method for extracting subtrees against forest query

This is the authors' abstract. We don't add key points for this paper.

In this paper, we present an algorithm to search and rank top-k approximately matched subtrees from a tree database, where the query is a collection of trees i.e. a forest. Even though existing algorithms can handle a single tree query, we argue that forest query would be significantly useful in some real life applications including biological domain. To address the issue we have proposed a method to find relevant subtrees and rank those given a tree database and a forest query. Tree edit distance is used to find and rank a set of subtrees with a pruning technique to improve the performance of the algorithm. We have tested our algorithm on different data sets and the efficiency of the searching and ranking process show promising results. Experimental results suggest that our algorithm improve run time at this stage and in future we would like to make it more useful for practical large data set.