Consider a binary classification problem where the target variable is highly imbalanced. You may imagine problems like detecting fraudulent transactions, predicting attrition, cancer detection, etc. where the number of positive examples is relatively fewer as compared to the number of negative examples. In such cases, training a classification algorithm to detect the positive classes accurately becomes difficult as the model becomes biased towards predicting the negatives.
Date : March 25, 2020 at 08:31AM
Tag(s) : #AI ENG
