Title Named Entity Recognition using Wikipedia and Abbreviation Generation

In this paper, we propose a title named entity recognition model using Wikipedia and abbreviation generation. The proposed title named entity recognition model automatically extracts title named entities from Wikipedia so constant renewal is possible without additional costs. Also, in order to establish a dictionary of title named entity abbreviations, generation rules are used to generate abbreviation candidates and abbreviations are selected through web search methods. In this paper, we propose a statistical model that recognizes title named entities using CRFs (Conditional Random Fields). The proposed model uses lexical information, a named entity dictionary, and an abbreviation dictionary, and provides title named entity recognition performance of 82.1% according to experimental results.