Bayesian Linear Regression

For the majority of standard prediction algorithms, after training the model we provide some features as input and get a prediction of the label in turn. This label tends to take the form of a single number, but often that's not enough to be of any use as we don't have a sense of how confident the model is in its prediction or what its margins of error are.

When we fit a Bayesian Linear Regression model, we can easily compute the uncertainty associated with a prediction, allowing us to make more informed decisions!

Online resources

• This introduction to Bayesian Statistics for beginners;
• Section 3.3 of Bishop's Pattern Recognition and Machine Learning for a mathematical introduction to Bayesian Linear Regression

Click the links below to access the Jupyter Notebooks for Bayesian Linear Regression

• Bayesian Linear Regression - Empty [Online notebook | .ipynb file]
• Bayesian Linear Regression - Redacted [Online notebook | .ipynb file]
• Bayesian Linear Regression - Complete [Online notebook | .ipynb file | HTML file]