Phase 4 Analytical Report

Phase 4 – Regression modeling (200 points)

In this phase, you will create a regression model using the modeling process. The details of the process are covered in Chapters 23, 24, and 25, but this phase entails the following steps.

1. Identify dependent (response) and explanatory (independent) variables. You may have more independent variables than you need at this stage.
2. Run a multiple regression model (MRM) and identify any issues with multicollinearity.
3. Address multicollinearity, if found.
4. Choose independent variables. Select the model. You may also use the partial F test, if appropriate.
5. For the chosen model, run the regression diagnostics. Check (i) the linearity requirement by creating the scatterplot matrix between the response and all explanatory variables, (ii) heteroskedasticity, (iii) autocorrelation.
6. Validate the goodness-of-fit of the model.

Run the Overall F test.

Run the statistical significance test of individual X variables.

Examine R-squared and adjusted R-squared. Interpret the value of R-squared.

Examine RMSE of the model and interpret its value.

If the model is valid, then use it for prediction.