Question

1. A good predictive model is one that fits the data closely whereas a good explanatory model is one that predicts new cases accurately. A. True B. False

2. The specificity of a classifier is its ability to detect the important class members correctly and sensitivity is its ability to rule out C0 members correctly. A. True B. False

3. This method of finding the best subset of predictors relies on partial, iterative search through the space of all possible regression models. The end product is one best subset of predictors. A. Exhaustive Search B. Subset Selection Algorithms C. Stepwise Regression D. All the above

4. This Model is used to fit a linear relationship between a quantitative dependent variable also called outcome or response variable and a set of predictors also called independent variables. A. Multiple Linear Regression B. Simple Linear Regression C. Stepwise Regression D. All the above

5. The advantage of choosing K > 1 is A. fitting to the noise of data B. That higher values of k provide smoothing that reduces the risk of overfitting due to noise in the training data. C. helps in finding the outliers D. None of the above

6. The number of records required in the training set to qualify as large increases exponentially with the number of predictors p. This is because the expected distance to the nearest neighbor goes up dramatically with p unless the size of the training set increases exponentially with p. This Phenomenon is known as A. Curse of Dimensionality B. Overfitting C. Smoothing D. All the above

7. The naive Bayes classifier's beauty is in its A. Simplicity B. Computational efficiency C. Good Classification Performance D. All the above

8. This is the difficulty with the practical exploitation of the power of the k-NN approach A. Time to find the nearest neighbors in a large training set can be prohibitive B. Number of records required in the training set to qualify as large increases exponentially with the number of predictors. C. The time consuming computation is deferred to the time of prediction D. All the above  

9. Which of the following statement is not true A. Naive Bayes classifier requires a very large number of records to obtain good results. B. With Naive Bayes, Good performance is obtained when the goal is classification or ranking. C. Naive Bayes method is used frequently in Credit Scoring. D. None of the above

10. This is a recursive partitioning method that predates CART (Classification and Regression Tree) procedures by several years and is widely used in database marketing applications to this day. A. Pruning B. CHAID C. Naive Bayes D. All the above

Answer 1

1. A good predictive model is one that fits the data closely whereas a good explanatory model is one that predicts new cases accurately. ANS: A.TRUE

2. The specificity of a classifier is its ability to detect the important class members correctly and sensitivity is its ability to rule out C0 members correctly. ANS:B. FALSE

It is other way round. Sensitivity is defined as the ability of a test to identify as positive, all the patients who actually have the disease. Specificity is defined as the ability of a test to identify as negative all the patients who do not have the disease.

3. This method of finding the best subset of predictors relies on partial, iterative search through the space of all possible regression models. The end product is one best subset of predictors. ANS: D. All the above

4. This Model is used to fit a linear relationship between a quantitative dependent variable also called outcome or response variable and a set of predictors also called independent variables. ANS: A. Multiple Linear Regression

5. The advantage of choosing K > 1 is ANS:B. That higher values of k provide smoothing that reduces the risk of overfitting due to noise in the training data.

6. The number of records required in the training set to qualify as large increases exponentially with the number of predictors p. This is because the expected distance to the nearest neighbor goes up dramatically with p unless the size of the training set increases exponentially with p. This Phenomenon is known as ANS: A. Curse of Dimensionality

7. The naive Bayes classifier's beauty is in its ANS: D. All the above

8. This is the difficulty with the practical exploitation of the power of the k-NN approach ANS: D. All the above