Study On Recognition Of Chinese Agricultural Named Entity With CRF

Named entity (NE), as the basic information unit of text, is essential to the correctunderstanding of a text. Named entity recognition (NER) is to identify the words in adocument belonging to NE and further classify them into some predefined categories.Named entity has been widely used in machine translation, text classification, informationretrieval, automatic summarization or other Natural Language Processing applications.Accordingly, its solution will promote the research of the relevant fields. In this thesis,attention is concentrated on the agriculture named entity, studied and discussed from thefollowing aspects.1. This article summarizes the methods of previous studies on Named EntityRecognition and research status of in-depth understanding of the named entity recognitionevaluation activities at home and abroad in recent years. In addition, it has deeply studiedHidden Markov (HMM) model, the maximum entropy (ME) model and conditionalrandom (CRF) model, also analyzes the advantages and disadvantages of the three models.Model that is used to agriculture named entity recognition research in this paper isCRF.In this paper, we has chosen CRF model.2. This paper uses CRF models and enters the word as granularity. Combininginternal and external features of the agriculture named entity, we select the appropriatefeatures, and train model to get a better recognition efficiency model file. The experimentresults show that, F values can reach92.2%based on the CRF model. In order toimprove the F value again, we add a rule to correct the CRF model recognition results, theresults show that the value of F can finally reach95%.3. This paper implements a CRF-based named entity recognition system of agricultureand discusses several important modules of the framework in detail. This paper hasconducted four experiments through the system, and proved that the F value of agriculturalNER eventually can be able to reach95%with stable performance.