Department of Information Science and Intelligent Systems, Faculty Engineering, University of Tokushima, Tokushima, Japan

---

Abstract—A trie is an ordered tree data structure to store keywords. It is used in natural language processing and so on. The trie is represented by the double array. The double array can retrieve fast at time complexity of O(1). The double array using linear functions (DALF) is proposed as a compression method of the double array. DALF reduces space usage of the double array to 60%. DALF is built by using parameters, and its space usage and its construction time depend on these parameters. However, appropriate values of them are not determined. This paper observes these parameters and evaluates parameters by experiments. From experiments, appropriate parameters are found, and it turns out that DALF can be built more efficiently by keyword sets including multibyte characters.

Index Terms—Trie, double array, construction method, keyword search.

[PDF]

Cite: Shunsuke Kanda, Kazuhiro Morita, Masao Fuketa, Jun-Ichi Aoe, "Experimental Observations of Construction Methods for Double Array Structures Using Linear Functions," Journal of Software vol. 10, no. 6, pp. 739-747, 2015.