When we use logistic regression for classification, the training risk is an unbiased estimate for the (true) classification error rate.

SECTION B – CONTEXTUAL QUESTIONS QUESTION 5 – CASE STUDY

In logistic regression, if the model is not a good fit, which options below might improve the goodness of fit? Select ALL correct answers.

Both controlling variables and explanatory variables must be considered in the variable selection process for potential discarding from the regression model.

Backward stepwise regression was carried on the above model and the partial result from the first step is shown in the table below. > backwards

What is the number of observations (or sample size)?

QUESTION 6 – REFERENDUM

2.8 Personalist dictatorships is when power lies with a single person. The military or any party cannot control this person. (1)

What are the values for MSE and F-statistic?

Let , be the set of quantitative predictors and y be the binary response in a logistic regression model. The estimated coefficient corresponding to is