A Classification Method based on Multi-decision Framework Using Regression Model
Yi‐Chong Zeng · 2020
Most classification approaches utilize a single classifier and result in the classification performance depends on the classifier. In this paper, we propose a classification method based on a multi-decision framework instead of a single classification, which performs like consensus decision-making. The multi-decision framework consists of several decision operators in two or more stages. The core of the operator is a regression model, such as Support Vector Regression, Gaussian Process Regression, and random forest. Combining threshold discriminators and a voting system, the final operator fuses the decision results generated by the previous ones. The proposed scheme compares to the existing approaches, and the experiment results demonstrate our method achieves better accuracies than the single classifier does.