Machine Learning with Imbalanced Data - Part 1 (Confusion matrix, precision, and recall)

Imbalanced data substantially compromises the learning process, since most of the standard machine learning algorithms expect balanced class distribution. In this video, we talk about the class imbalance problem that arises in various domains, such as computer vision, medical diagnosis, fraud detection, and bioinformatics, to name just a few. To get you a real feel of the problem, we examine the thyroid data set and employ the logistic regression model. We show that accuracy doesn't provide a reliable and meaningful metric on class-imbalanced data sets. Instead, one should plot the confusion matrix and compute precision and recall scores. We show how to use Scikit-learn from Python to compute several evaluation metrics.

#ImbalancedData #ConfusionMatrix #Precision