for IT

no.5 Write short answers for the following questions. a. Is scalingormalization necessary in PCA? If yes, why? What will happen if scalingormalization is not performed? b. You are working on Churn Prediction dataset for a well-known streaming service. While working on a dataset, you analyzed that there are many dimensions/features. Some of the features have constant values, some have missing values, and some have zeros. How do you select important variables? Explain atleast 3 methods with examples. c. Why the measure Lift is preferred over support and confidence