Research On Named Entity Recognition In Automotive Filed

With the rapid popularization of the computer,and the rapid development of the Internet,the amount of information is becoming more and more.So,how to obtain necessary information from the massive data becomes a topic of concern.The main purpose of information extraction is to convert unstructured natural language text into semistructured or structured data so that people can get information quickly and accurately.Named entity recognition,as the sub-task of information extraction,has gradually become the focus in professional field.In this paper,the named entity recognition in automotive filed will be discussed as follow:(1)A method of ontology construction for automobile domain based on thesauri is proposed.In this paper,an ontology construction method based on the thesauri is proposed,and a new ontology construction method based on the thesauri is proposed.Good ontology as a feature of the automotive field named entity recognition.The follow-up experiments show that the ontology of automobile field constructed in this paper is reasonable and feasible.(2)This paper proposes an entity recognition method based on ontology.Firstly,based on the car-like corpus provided by the COAE,this paper analyzes the characteristics of naming entities in the corpus,uses the n-gram model to transform the corpus,and extracts the words and words from the converted corpus as the basic features Then,the context structure of the corpus is analyzed,and a window is set up to extract the words with the instructive meanings.At the same time,according to the features of the evaluation phrases,the affective tendencies of the sentences are extracted and the effect of the emotion characteristics on the named entity recognition effect is explored.Finally,the auto-domain ontology is constructed as another feature,and several other features are combined into different feature templates.The conditional random field model is used to label the vehicles to identify the naming entities.The contrast experiment of feature combination shows that the recognition effect is improved greatly after adding the ontology feature.(3)A method for identifying naming entities of hidden car is proposed.In this paper,we define the naming entities composed of unlisted words in the corpus as hidden entities,analyze the syntactic dependencies of implicit entities in corpus,and propose an implicit entity based on syntactic dependencies Recognition method.Experiments show that this method can further improve the accuracy of naming entity recognition in automobile field.