KNN Ensemble Learning Integration Algorithm Based on Three-Way Decision

Abstract

AbstractThe KNN algorithm is affected by overlapping classification, unbalanced data, and K-value selection. It is difficult to apply to some environments with uncertain phenomena. At the same time, the three-way decision is a decision theory that conforms to the human cognitive model with subjective characteristics, so the idea of a three-way decision is introduced into the KNN ensemble learning algorithm and the KNN ensemble learning algorithm based on the three-way decision is proposed. Based on the KNN ensemble learning algorithm, the conditional Probability of each class is calculated and combined with the cost function, which is used to determine the positive domain, negative domain, and boundary domain in the three-way decision theory. This paper performs the three-way decision KNN ensemble learning classification on seven real UCI datasets. The experimental results show that it can effectively improve the classification accuracy and F1-score of the data.KeywordsKNNThree-way decisionEnsemble learning algorithm