| Popis: |
Aiming at the shortcomings of the current classification models of distribution lines, such as low classification accuracy and poor robustness, a rating classification model of distribution network based on dynamic classification is proposed. Based on K-PAM, the model dynamically combines the decision tree classification method C4.5 to efficiently use the prediction information of base classifier to achieve higher classification accuracy. Finally, the real data of the three distribution networks were used to conduct the experimental analysis, and the classification efficiency, classification accuracy and robustness of the three methods were compared. |
