Extracting Noun Phrases from Large-Scale Texts: A Hybrid Approach and Its Automatic Evaluation