Feature Terms Analyzing Strategy for Recruiting Websites

As we know text information on web page has grown exponentially. It is a hot research area in data processing by reasonably extracting and analysis for unstructured information, so as to mine novel, latent useful pattern. Focusing on imprecise classified text set about job hunting web site, discovering topic relevant feature terms is an effective way to find new tendency for work ability demanding. In this paper, we propose a job relevant feature extracting method better than methods of TF-IDF, maximum entropy and lexical chain to reflect the demanding of tendency, and prove that it is effective by contrast testing.