Abstract: In information retrieval,minwise hashing algorithm is often used to estimate similarities among documents,and frbit minwise hashing is capable of gaining substantial advantages in terms of computational efficiency and storage space by only storing the lowest h bits of each(minwise) hashed value(e. g. ,b=1 or 2). Fractional bit minwise hashing has a wider range of selectivity for accuracy and storage space requirements. For the fixed fraction f,there are so many combinations of f. We theoretically analyzed limited combinations of fractional bit hhe optimal fractional bit was found. Experimental results demonstrate the effectiveness of this method.

Key words: Similarity estimation, Hasing, Optimal fractional bit