Incremental Patent Semantic Annotation Based On Keyword Extraction And List Extraction

At present, there is a lack of in-depth processing and indexing of Chinese patents in China, which makes the patent data retrieval inaccurate and incomplete, leading to duplication of applications and waste of resources. Aiming at the problem of lacking annotated patent data in Chinese patent indexing, this paper studies an incremental patent annotation method. By using co-training method, keyword extraction and list extraction can cooperate with each other and iteratively annotate the functional clauses, which achieves the effect of obtaining much more annotated data through a small quantity of training data. Experiment results indicate this method can gradually improve the recall without sacrificing much precision.