Short Message Orientated Chinese Named Entity Recognition

Named Entity Recognition is a basic task in Natural Language Processing. It plays an important role in information retrieve and extraction, machine translation and question answering system. Our work is focus on the named entity recognition of short message. The main work in this paper includes:1. Construction of knowledge base for Chinese short message orientated Named Entity Recognition.Basing on the analysis of Chinese short message, we abstract large number of morphological characteristics on Person, Location and Organization name in Chinese short message. Using these characters, we make up short message knowledge base respectively for Person, Location and Organization name recognition.2. Combining expert knowledge with Conditional Random Field model to solve the Problem in Chinese short message orientated Named Entity Recognition.For different entities, different methods are used. Person and Location name are recognized with the method of combining expert knowledge with Conditional Random Field model; the organization name recognition is based mainly on the expert rules.As the experiments show, we achieve a good performance on Chinese short message orientated Named Entity Recognition.