英语论文网

6.1 Introduction of attribute features
The research of attribute features describes the inclusion and exclusion relationsamong attributes. Compared with the previous researches, the Reference material aboutattribute features is rare. Up to present, the only available one is The Complete Definitionsof Attributes and Abstract Description of Attribute Features of the Formal Context byprofessor Hong (2012). In this paper, Hong (2012) describes the complete attributedefinitions and the abstract descriptions of condition formal context and decision formalcontext. Specific examples are also given during the description. Professor Hong dividesthe attribute features into 11 kinds: the largest common attribute, common attribute,opposite attributes, mutually repulsive attributes, mutually non-inclusive attributes,concomitant attribute, the nearest concomitant attribute, the common concomitant attributeof several associated attributes, the unique attribute, the unique attribute of the congenericobjects and the unique composite attributes of the congeneric objects (Hong, 2012). It isnoted that the unique attribute, the unique attribute of the congeneric objects and theunique composite attributes of the congeneric objects play a part in the classification ofthe three senses of preposition on.

…………

Conclusion

On the basis of Formal Concept Analysis tool and corpus-based method, this paperconstructs the word sense disambiguation model of preposition on, extracts three groupsof rules used for semantic recognition and introduces a new rule extraction method forfurther researches. Compared with the previous researches, this paper contributes to thestudy of preposition on and natural language processing. Certain achievements andfindings are reached.The problem of data sparsity is likely to appear in the previous word sensedisambiguation researches for the involvement of one training set and one testing set. Datasparsity may influence the accuracy of word sense disambiguation and rule extraction.Based on the previous situation, this research extracts one training set and two testing sets.The addition of more data improves the problems caused by data sparsity to some extent.On the basis of one training set and two testing sets, the formal contexts are built andthe attribute positive sequence diagrams are generated, which are the foundation of wordsense disambiguation with the final accuracy reaches 95%. On the combination of onetraining set and two testing sets, three groups of rules are extracted with a re-examinationaccuracy reaches 91.56%. These research results, on one hand, prove the effectiveness ofword sense disambiguation; on the other hand, prove the correctness of the extracted rules.
…………
Reference (omitted)

本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。