The Description And Search Of Chemical Structures In Scientific Documents

There are a large amount of chemical resources which are described through Systematic Nomenclatures, Line Notations etc on the Internet. In face of such massive information, search engines are expected to find out the information that users are intersted. However, the retrieval of chemical structure is hard to be realized by the currently general search engines, such as Google and Baidu which can only search through keywords. In order to use these chemical resources effectively, a chemical structure search engine is requried.Based on the analysis of different chemical structure representation approaches, CML is used as the mark language for the chemical information in the ScienceML(Science Markup Language) for facilitating chemical structure search. Combining modern information retrieval technology and the features of chemical structure, a solution for chemical structure search engine named Chem Search is proposed. The Chem Search provides full structure, substructure and structure similarity search. For identifying the aim structure quickly, a hash function is adopted to locate the chemical structures stored in the database."Keywords matching algorithm", which can reduce the search scope and improved the efficiency of grabbing web pages, is used when Robot grasps the web pages which containing chemical information.. The chemical structure similar searching algorithm based on Feature-Graph Matrix Index is strengthened to further improving the efficiency of the similarity searching. Users can access Chem Search through the browser, input the chemical structure's SMILES/InChI, or draw chemical structures directly, then query the database and get the results.The effectiveness of the proposed Chem Search is demonstrated by experiment. How to identify and get the picture information effectively to enlarge the search scope of Chem Search is our future research keystone.