HYBRID APPROACH FOR GENERATING NON OVERLAPPED SUBSTRING USING GENETIC ALGORITHM

Akila Rani.M, Shanthi.D, Farzhana.I

Abstract: Approximate Membership Localization (AML) is the process that provides user with most relevant matched substrings. In a document, one word position belongs to only one reference matched substring. There should not be any overlap in a true matched substring. In the Approximate Membership Extraction (AME) technique when searching a document in a web it displays all coordinated substring .So redundancy occurs and it causes less efficiency. To overcome this problem, AML is used. It provides non overlapped substring during searching process and avoids redundancy by using optimized algorithm called P-prune algorithm. The pruning algorithm eliminates unwanted data that is overlapped data and increases the efficiency of searching process. The high comparison load and time taken for generating the result is minimized. This enhancement can be achieved by Genetic algorithm which helps in identifying true matched substring with the help of fitness function. The equivalent key term are compared instead of comparing all the terms and hence reduces the time taken.

Keywords: Approximate Membership Localization (AML), Approximate Membership Extraction (AME), Pruning algorithm, Genetic algorithm.

DOI: https://doi.org/10.15623/ijret.2014.0305045