Method Of Classification Mapping Between IPC And CLC Based On Machine Learning

With the rapid development of modern science and technology, there is a dramatic increase in the number of literature resources, and how to find the latest and most comprehensive information in which people are interested in the mass of resources has become a problem to be solved. Patents can reflect some information such as the latest scientific research news, and the development status, technology level and legal status of the research, and are of great scientific value and commercial value. Compared with other literature resources, however, the utilization of patent information is lower. With the growing concern in the patent information today, to achieve the interoperability between patent and journal literature is of great significance.International Patent Classification (IPC) is currently the most common management tools for patent documentation. Chinese Library Classification (CLC) is the most widely used classification in China. Completing the mapping between IPC and CLC is an important way to achieve the cross-browse and cross-search in different organization systems. Based on the research of the existing interoperability projects and the algorithms for classification mapping, this paper proposes a method that is based on machine learning to achieve the category mapping between IPC and CLC. By training the corpus which is identified by a category we can get the category classifier, and then use it to classify the corpus that is identified by other categories, and analyze the result to determine with which category or categories there will be a corresponding relationship. Finally, experiments show that the method can complete the mapping between different categories.