Space-efficient implementation of a compressed trie

Abstract

Trie is the typical data structure for keyword searching algorithms. The algorithms are classified into two ways. One is the array representation, the other one is the succinct representation. In array representation, one can access the child node using the one dimensional array which stores the pointers from the node to its child. This shows fast lookup time, but takes a lot of space. In succinct representation, one separates a tree structure and index array, and represents a tree structure by succinct representation. This shows slower query time than array representation, but takes less space. In this thesis, we first use the degree sequence and its variants to implement the space-efficient trie on the small fixed alphabet. The experimental result shows that for small alphabet (

Σ

≤10), our implementation gives fast query time and less space usage compared to the exist implementations.