Gaussian Linear Regression: Assumptions and Residuals

This lecture by the instructor covers the four basic assumptions of the Gaussian linear regression model: linearity, homoskedasticity, Gaussian distribution of errors, and uncorrelated errors. The lecture explains how to check these assumptions using scientific reasoning and graphical methods, focusing on residuals as a key tool. It discusses the implications of incorrect assumptions and the importance of checking for linearity and homoskedasticity through plots of residuals against explanatory variables and fitted values. The lecture also delves into the concept of normality by comparing the distribution of standardised residuals against a normal distribution using quantile plots.