Research On Association Analysis Of The Text Based On AKD

With the enhancement of the level of information in the government departments, a lot ofinformation is stored in form of electronic data, and is unified managed by informationsystem. However, a large number of text messages only to be used in the form of documents,most of them in the document could not be further used.In this thesis, we studied the knowledge discovery approach of text information in thesocial insurance field, following the purpose of the audit of the state social insurance. We firststudied the research of association rules mining, and summed up the general process of textassociated analysis methods. Second, we analyzed the difficulties and problems of dealingtext data with the characteristics of unstructured text data, and discussed methods of the textpre-processing. Again, on the basis of the text pre-processing, we discussed the associatedanalysis methods which on the structural of the text features, and combinated thecharacteristics of the semantic relationship in the audit knowledge database, and proposed theassociated analysis methods of the text based on the text of the audit knowledge database.Finally, we analyzed the application of text association rules in the audit field.In this thesis, we verified the practical effect of the text of this association analysisstrategy with the text data in social pension insurance, the system of text association Analysisbased on audit-field knowledge database is useful in the field of social pension insurance.