Machine Learning - Testing and Error Metrics

Overview

Embark on a friendly journey into the process of evaluating and improving machine learning models in this 45-minute video tutorial. Explore essential concepts such as training and testing, evaluation metrics including accuracy, precision, recall, and F1 score, and types of errors like overfitting and underfitting. Delve into cross-validation techniques, including K-fold cross-validation, and learn to interpret model evaluation graphs. Discover the power of grid search for optimizing model performance. Gain practical insights through examples of medical models, spam classifiers, and credit card fraud detection. Master the art of diagnosing model performance using confusion matrices and understand the tradeoffs between different types of errors. By the end, acquire valuable skills in problem-solving and effectively applying machine learning techniques to real-world scenarios.

Syllabus

Introduction
Which model is better
Why Testing?
Golden Rule # 1
How do we not 'lose' the training data?
K-Fold Cross Validation
Randomizing in Cross Validation
Evaluation Metrics
Medical Model
Spam Classifier Model
Confusion Matrix Diagnosis
Accuracy
Precision and Recall
Credit Card Fraud
Harmonic mean
F1 Score
Types of Errors
Classification
Error due to variance overfitting
Error due to bias underfitting
Tradeoff
Solution: Cross Validation Testing
Training a Logistic Regression Model
Training a Decision Tree
Training a Support Vector Machine
Grid Search Cross Validation
Parameters and Hyperparameters
How to solve a problem
How to use machine learning
Thank you!