A Study on Small-sized Index Structure and Fast Retrieval Method Using The RCB trio

This paper proposes RCB(Reduced Compact Binary) tie to correct faults of both CB(Compact Binary) tie and HCB(Hierarchical Compact Binary) trie. First, in the case of CB trie, a compact structure was tried for the first time, but as the amount of data was increasing, that of inputted data gained and much difficulty was experienced in insertion due to the dummy nods used in balancing trees. On the other hand, if the HCB trie realized hierarchically, given certain depth to prevent the map from increasing on the right, reached the depth, the method for making new trees and connecting to them was used. Eventually, fast progress could be made in the inputting and searching speed, but this had a disadvantage of the storage space becoming bigger because of the use of dummy nods like CB trie and of many tree links. In the case of RCB trie in this thesis, the tree-map could be reduced by about 35% by completely cutting down dummy nods and the whole size by half, compared with the HCB trie.