Papers by Author: Zhi Feng Hao

Abstract: A general depth-first backtracking algorithm for graph isomorphism with the vertex partition and refinement technique is presented in this paper. The time complexity of this nondeterministic polynomial algorithm is O(nα+3) where nα is the number of backtracking points and (h-1)/2α (h+1)/2 for h=logn in the worst cases. The tests on many types of graphs validated the efficiency of this algorithm for graph isomorphism.

Abstract: This paper proposes an improved web crawler algorithm to climb more useful information since the basic web crawler algorithm is low-efficiency and easy to climb useless repeated information. By the proposed algorithm, the website urls are hierarchical saved to store websites overall topology, which will make crisscross complex web URL system from a graphic structure into a tree structure. The actual website BBS experiments show that the algorithm is much better than the basic web crawler algorithm in crawling speed and download information such as the usefulness of baking. Furthermore, it provides a performing structure mode for the increment crawler algorithm.

Abstract: Association rules provide a concise statement of potentially useful information, and have been widely used in real applications. However, the usefulness of association rules highly depends on the interestingness measure which is used to select interesting rules from millions of candidates. In this study, a probability analysis of association rules is conducted, and a discrete kernel density estimation based interestingness measure is proposed accordingly. The new proposed interestingness measure makes the most of the information contained in the data set and obtains much lower falsely discovery rate than the existing interestingness measures. Experimental results show the effectiveness of the proposed interestingness measure.