Colin M.

Recent Posts

Understanding A Confusion Matrix

Posted by Colin M. on 8/24/20 9:24 AM

One of the major tasks in machine learning and statistical testing is classification. In classification problems, we use a training set of labeled data to train our model to classify an unlabeled observation into one category or another. At the simplest level, this method uses observable data to make a related yes-or-no classification (like: will it rain today or not rain today). Classification problems can also have more than two classifications (like: will it be cloudy, sunny, rainy, snowy, etc.), but the principles for analyzing the results are largely the same. Many popular techniques exist for classification problems, such as logistic regression, trees (including boosted trees and random forests), and neural networks. To see how well a given method works, we can use a confusion matrix to understand the results of the model.

Read More

Tags: statistics & probability