Study On The Classification Mapping Between IPC And CLC

Massive growth of information resources has led to an "Information Explosion" situation that S & T information users are suffering from "Knowledge Hunger" and "Failure in finding the real needs and the latest information and data". Patent literature, as a special kind of document, includes plenty of advanced technology, meanwhile some problems exists such as management difficulties, isolation, low utilization etc.International Patent Classification (IPC) is the most popular tool to manage and make use of patent documentations at the moment. Establishing the relationships between IPC and CLC through mapping can provide great convenience for cross-browsing and cross-search in patents and periodicals data.After doing mass of researches and verifications on the Classification Mapping topic, two different methods are proposed to achieve mapping between IPC and CLC, which are based on category similarity and cross-concordance respectively. A conceptual model for a category is defined in the process of calculating the category similarity. To get better results, two methods of lexical similarity are introduced. There will be more manual intervention in the first method to build mapping so that another method based cross-concordance is designed, which is more dependent on mathematical statistic rather than human knowledge. At last, experiments show that both methods can effectively realize classification mapping.