Fast String Searching Based on Multilevel Hashing

This work aims to design and develop a system for fast searching of variable length of strings taking the advantage of using the enumeration and multilevel hashing to improve the searching time by providing a semi-uniform searching performance, with the use of weighting scheme and some data structure concepts. The designed system takes into consideration various kinds of factors that affect the searching process (e.g. the search space size). Also, in order to provide an efficient system performance capable of handling some spelling mistakes an updated algorithm for string similarity measure was proposed, which is a combined set of well-known algorithms with some improvements by using certain conditions which are introduced after conducting on a number of experimental tests. The proposed algorithm ¿Optimized Dice Coefficient, N-grame and Damerau-Levenshtein Algorithms using Numeration Operations (ODADNN)¿ provides efficient results to assess the similarity between two words without the need to restrict the words length.