| The indexing and retrieval of generic structures has always been among the most problematic aspects of patent information and the most expensive. In this paper, the problem posed by the requirement for storage and manipulation of generic structure definitions in patents is reviewed. Chemists and patents agents have developed an armory of methods of representation over many decades so that a generic structure description can describe large and often unlimited numbers of substances as a result of combinatorial opportunities provided. The nature and theoretical foundations devised during the Sheffield project, CAS, Japan and China research groups for the successful solution of the problem in order to provide the desired and practical retrieval facilities are reviewed. An innovative method of representation for generic structures using scaffold is presented, which comprises the following two steps: first, the method analyses all possibilities of variations described in generic expressions and then divided generic structures into rings and fragments by virtue of the ability of detecting ring and branch, especially aromatic rings, of extended SMILES notations; second, generic structures are represented by reduced graph of scaffold. The whole process of encoding with particular reference to patents and the application to searching system are also described. The program based on OOP is developed, which accomplishes the division and scaffold of query structures automatically. Hie retrieval system designed is built on the platform of Windows 2000+HS5.0 and tested on local machine, which supports searches including queries comprising specific structures, for which inclusion as a member of a generic class should be the criterion for retrieval, generic structures, for which an overlap of one or more structures between the query and a database structure should be determining. The result of searches seems to be satisfying. |
PDF Full Text Request