Monstermash, an online game app development company, has built a predictive model to identify gamers who are likely to make in-app purchases. The model classifies gamers who are likely to make in-app purchases in Class 1 and gamers who are unlikely to make in-app purchases in Class 0. Applying the model on the validation data set generated a table that lists the actual class and Class 1 probability of the gamers in the validation data set. A portion of the table is shown below.

a. Specify the predicted class membership for the validation data set using the cutoff values of 0.25, 0.5, and 0.75. Produce a confusion matrix based on the classification results from each cutoff value.
b. Compute the misclassification rate, accuracy rate, sensitivity, precision, and specificity of the classification model from each of the three cutoff values specified in part a.
c. Create the cumulative lift chart, decile-wise lift chart and ROC curve for the classification model.
d. What is the lift that the classification model provides if 20% of the observations are selected by the model compared to randomly selecting 20% of the observations?
e. What is the lift that the classification model provides if 50% of the observations are selected by the model compared to randomly selecting 50% of the observations?

Transcribed Image Text:

Gamer 1 2 ⠀ 920 Actual Class 0 0 ⠀ 0 Predicted Class 1 Probability 0.022666 0.032561 0.040652