Implementation of the paper- LRID: A new metric of multi-class imbalance degree based on likelihood-ratio test by Rui Zhu et. al.
A new measure of imbalance degree using likelihood ratio is introduced in order to compute the extent of imbalance in a multi-class data set. Usually Imbalance Ratio (IR) is used as a metric to measure extent of imbalance. However, it takes into consideration only the highest majority class and the lowest minority class, thereby failing to capture any information from the intermediate classes. Imbalance Degree (ID) is a metric was proposed while keeping in mind that it has to consider the intermediate classes as well. But it requires us to choose a distance metric that largely affected the results. It also assumes that data with more minority classes are more imbalanced than data with less minority classes, which is not always true.
A new metric is proposed to measure the class imbalance extent of multi-class data based on the likelihood-ratio test. It can provide effective measurement of class-imbalance extent and can be easily applied in practice. In the experiments, it demonstrates its superior performances over IR and ID.