Research And Application Of Qualitative Data Analysis Method Based On Granular Computing

ABSTRACT: Along with the development of computer technology and thepopularization, the data obtained by people are increasing sharply with unprecedentedspeed, therefore, the method of the large qualitative data statistical analysis is burningdevelopment and improvement. The granular computing theory of the rough settheory covers all the particle size of the theory, technology and method, and itprovides a new view for qualitative data analysis method innovation, and is a newmethod of the current application of computational intelligence solving complexqualitative data analysis.Based on granular computing theory knowledge and traditional qualitative datastatistical analysis method, the granular computing of the field of computer into thequalitative data statistical analysis is introduced, and a thoroughly study on the largequalitative data statistical analysis method is carried out, and the qualitative datavariable reduction model is established based on variable reduction qualitative dataclustering analysis model. The main contents of this paper are listed in the followingsection:1. On the basis of the basic principle of granular computing theory and thepresent research situation of qualitative data, according to the characteristics of largequalitative data, the qualitative data of granular computing description is completedby combining with the information system description method, which provides atheoretical basis for large qualitative data statistical analysis method research.2. In rough set theory, on the basis of variable reduction by recognition matrix,according to the insufficient of processing mass qualitative data operation efficiencywith a lot of limitations in the above method, a mathematical model of variablereduction with granular computing is given by applying the definition of variableimportance degree, which provides a theoretical evidence for the elinimination ofredundancy variables in qualitative data, and the simulation results prove that thismodel is feasibility and effectiveness in processing large qualitative data.3.In order to do the further statistical analysis for the large qualitative data, onthe basis of variable reduction model, a mathematical model of qualitative dataclustering analysis is given with variable importance. Because the clustering result istoo simple with a larger error, the optimal variable cluster subset is selected to solvethis problem, and a mathematical model of qualitative data clustering analysis withgranular computing variable reduction is establised. Finally, the simulation resultsshow that the model has a higher accuracy and effectiveness.