Construction And Application Of Three-way Bayesian Classification Models

In view of the high error rate of the two-way decision Bayesian classifier in processing uncertain data in the classification process,the three-way decision idea is integrated into the Bayesian classification and a three-way Bayesian classification model is constructed.Considering the dynamic update of the data,research and construct a dynamic three-way incremental Bayesian classification model,and apply it to the recommendation system.Firstly,using the incremental learning feature of the NB classifier to update the model parameters and calculate the conditional probability.Formulating three-way classification rules to construct three-way incremental naive Bayesian classification model.And verify the effectiveness of the method on the selected UCI dataset.Secondly,proposing a new three-way extended TAN Bayesian classifier,which builds a TAN Bayesian classification model and uses prior probabilities and class conditional probabilities to estimate the conditional probabilities.Formulating three-way classification rules in the 3WD-TAN classifier to construct a 3WD-TAN classifier.And conduct comparative experiments on the UCI dataset.Thirdly,based on the construction of the 3WD-TAN classifier,considering the dynamic update of the data,integrating the idea of incremental learning,constructing a dynamic three-way incremental tree augmented naive Bayesian classification model,and verify the effectiveness of the method through experiments.Finally,three-way Bayesian classification application frameworks are constructed and applied to the recommendation system.Considering the influence of item attributes on users,using NB classifier to effectively fill the matrix,the three-way decision ideas are introduced into the recommendation system,and the three-way Bayesian recommendation model is constructed.And the effectiveness of the algorithm is verified through examples.Figure 21;Table 26;Reference 58...