Summary

Here’s a brief summary of what you learnt in this session.

• When one variable might not be enough
  • A lot of variance values are not explained by just one feature.
  • Predictions are inaccurate.
• Formulation of MLR
• New considerations to be made when moving from SLR to MLR
  • Overfitting
  • Multicollinearity
  • Feature selection
• Dealing with categorical variables
  • Dummy variables for fewer levels
• Feature scaling
  • Standardisation
  • MinMax scaling
  • Scaling for categorical variables
  • Model assessment and comparison
    • Adjusted R-squared
    • AIC, BIC
  • Feature selection
    • Manual feature selection
    • Automated feature selection
    • Finding a balance between the two

Coming up

Now that you have gained an understanding of the theoretical considerations in building a multiple linear regression model, you will create one such model in Python in the next session.

Report an error