Low RAM Footprint Algorithm for Small Scale Plagiarism Detection Projects

The paper addresses the specific problem of lowering the RAM required to perform one-to-many document comparison. The lower RAM environment is specific problem to the web applications placed on shared hosting servers. Using simple hash based comparison as the basic model for similarity detection, the paper shows how all pre-processing and comparison operations can be moved into the database. Provided experimental evaluations support the applicability of the proposed approach.