News-Oriented Automatic Chinese Keyword Indexing

摘要

In our information era, keywords are veryrnuseful to information retrieval, text clusteringrnand so on. News is always a domainrnattracting a large amount ofrnattention. However, the majority of newsrnarticles come without keywords, and indexingrnthem manually costs highly. Aimingrnat news articles' characteristics andrnthe resources available, this paper introducesrna simple procedure to index keywordsrnbased on the scoring system. In thernprocess of indexing, we make use of somernrelatively mature linguistic techniques andrntools to filter those meaningless candidaternitems. Furthermore, according to the hierarchicalrnrelations of content words,rnkeywords are not restricted to extractingrnfrom text. These methods have improvedrnour system a lot. At last experimental resultsrnare given and analyzed, showing thatrnthe quality of extracted keywords are satisfying.