Building Of Classification Method And Classifier About Text Complaints Information Based On Association Rules

A few years, in order to better understand the public opinion of the people’s livelihood,the government department is gradually implemented public telephone system. Citizensthrough the mayor’s office, there are a number of12345hotline to reflect the situation andrequirements, content of each complaint by operator records into text information on thecomputer, each complaint information corresponding to a designated by the unit, this paperstudies the content is to use automatic sorting system to realize the complaint information unitclassification.At this stage, the government’s automatic sorting system mainly USES text categorizationmethod based on statistics, based on statistical classification method for the large-scalesituation has very high classification accuracy, but if the corpus is not comprehensive,representative is not strong, will directly affect the accuracy of automatic classification. Thecontent of this study is based on association rule mining to realize automatic sorting ofcomplaint information. Because in the complaint information, uneven distribution of sample,short text, the feature dimension is not high, in addition to this, keyword stability, correlationdegree is high, so the use of key phrase set to implement will improve overall accuracy. In theentire research process, first of all, it is necessary to design system, and then preprocessingoriginal data. Then using the improved Apriori algorithm for mining association rules in thecategory of, finally build up the need to use all the association rules text informationautomatic classifier.Research is to use the method for text classification based on association rule, actuallysolving government12345automatic classification of complaint information processingsystem kernel module, promote work efficiency. For the realization of the text informationclassification provides a new direction, to further expand and improve the automatic textclassification technology and method system.