Feature Selection Based on Gain Ratio in Hybrid Incomplete Information Systems
Jiazheng Chen, Jingjing Hu, Gangqiang Zhang · 2021 16th International Conference on Intelligent Systems and Knowledge Engineering (ISKE) · 2021
Fuzzy rough theory is widely used in feature selection because it can deal with fuzzy and uncertain data effectively. Inspired by the decision tree algorithm of machine learning and the kernel fusion concept of multi-kernel learning, the article uses a feature selection method based on gain ratio and a T-norm method to compute the similarity between attributes. We use data sets of mix types of attributes to detect the applicability of the feature selection method. The results manifest that this feature selection method is effective in the hybrid incomplete information systems.