Abstract: Semantic hashing for web documents is essential for effective information dissemination. This paper is a sincere effort towards application of a novel method which outputs a semantic hash for an input web document. The need of such method arises as a result of research search where user may be so naïve that they are unaware of domain specific keywords or any labels for satisfying their search goals. The proposed technique in this paper assigns rank from 1 to n based on highly relevant modals of a web document. We have used six of such modals and duly considered their impact in finding semantics of a web document when hashing with a user input document. The algorithm is used in many information retrieval systems and employs a distance metric learning mechanism in practice.

Keywords: Semantic, Retrieval, Multi-Modal, Context, LOMDML.