Chinese Personal Name Disambiguation Based on Vector Space Model

This paper introduces the task of Chinese personal name disambiguation of the Second CIPS-SIGHAN Joint Conference on Chinese Language Processing (CLP) 2012 that Natural Language Processing Laboratory of Zhengzhou University took part in. In this task, we mainly use the Vector Space Model to disambiguate Chinese personal name. We extract different named entity features from diverse names information, and give different weights to various named entity features with the importance. First of all, we classify all the name documents, and then we cluster the documents that cannot be mapped to names that have been defined. Eventually the results of classification and the clustering are combined. In the test corpus experiments, the accuracy rate is 0.6778, the recall rate is 0.7205 and the F value is 0.6985 for all names.