BE544 Lecture 10 - Performance Metrics and Transfer Learning

The video explains the concepts of a confusion matrix and how to fill its components. It also covers how to handle n by n confusion matrices. Additionally, it demonstrates how to extract the four components: TP (True Positive), TN (True Negative), FP (False Positive), and FN (False Negative) from the confusion matrix for both binary and multi-class classification. The video further discusses various performance metrics and how to calculate them using weighted, macro, and micro averaging. Moreover, it shows how to build the confusion matrix and extract the four components (i.e., TP, TN, FP, and FN) using the three approaches (i.e., weighted, macro, and micro averaging) in Python code. The video then explains the concepts of transfer learning and pretrained models. Finally, it includes the implementation of transfer learning using Keras pretrained models (i.e., Keras Applications) in Python code.
All code used in the lectures will be available in my GitHub repository (github.com/Hos...) in the "Lectures Scripts" folder.