Research And Application Of Na(?)ve Bayesian Classification Model Based On Clustering Algorithms

The classification is an important research branch in the data mining domain. There are many amazing achievements have been obtained. Owing to its highly efficiencies and highly precise calculation, as well as its strict theoretical foundation, Naive Bayesian Classifier has obtained widespread application. However, its condition independence assumption and perfection data requisition limit its real application. The local searching ability of k-means algorithm is exerted, and the precise of Naive Bayesian Classifier improved. It can solve classification problem effectively. The main work of the dissertation is as follows:1. Introducing and analyzing k-means algorithms of clustering algorithms and Naive Bayesian Classifier algorithm. The basic theory of Naive Bayesian is studied, and some common models of Naive Bayesian Classifier are discussed.2. A Naive Bayesian classification based on clustering principle(CNBC) by introducing clustering algorithm into Naive Bayesian classification. The similarity between every recorder in absent data subsets and the centers K cluster is calculated by clustering complete data subsets of initial data by k-means algorithm, then the recorder is set to the nearest cluster and the absent value of the record is filled by the corresponding attribute of the cluster, finally, the handled data set is clustered by Naive Bayesian classification algorithm. The experiments show that Naive Bayesian classification based on clustering algorithms has the higher precise of clustering comparing with Naive Bayesian classification.3. The model of Naive Bayesian Classifier based on clustering algorithms is designed to help the students to select employment or continue studying in university. By constructing the model and using the experience gained by the students in the past in their selection of specialties, students can base their conditions to select their way.