E-viri
Recenzirano
-
Zhou, HongFang; Zhang, JiaWei; Zhou, YueQing; Guo, XiaoJie; Ma, YiMing
Expert systems with applications, February 2021, 2021-02-00, 20210201, Letnik: 164Journal Article
•Use the median as the criterion for the feature weight threshold.•Use the improved ReliefF as a feature pre-filtering method.•Take the feature weight as the criterion for evaluating features in the decision tree. In order to improve the classification accuracy, a preprocessing step is used to pre-filter some redundant or irrelevant features before decision tree construction. And a new feature selection algorithm FWDT is proposed based on this. Experimental results show that FWDT our proposed method performs better for the measures of accuracy, recall and F1-score. Furthermore, it can reduce the required time in constructing the decision tree.
