Machine Learning

General Concepts In Machine Learning

Interview questions on General Concepts In Machine Learning.

14 questions

All subtopics Basics Cross Validation Similarity Measures Sampling Techniques and Creating Training Data Feature Engineering Bias and Variance

Cross Validation

Q1. Below Fig depicts two different cross-validation approaches. Name them.

Cross Validation

Q2. 1. What is the purpose of following Python code snippet?

        skf = StratifiedKFold(y, n_folds=5, random_state=989, shuffle=True)
        ```
2. Explain the benefits of using the K-fold cross validation approach.
3. Explain the benefits of using the Stratified K-fold cross validation approach.
4. State the difference between K-fold cross validation and stratified cross validation.
5. Explain in your own words what is meant by “We adopted a 5-fold cross-validation approach to estimate the testing error of the model”.

Cross Validation

Q3. True or False: In a K-fold CV approach, the testing set is completely excluded from the process and only the training and validation sets are involved in this approach.

Cross Validation

Q4. True or False: In a K-fold CV approach, the final test error is:

$CV_k = \frac{1}{k}\sum_{i=1}^{k}MSE_i$

Cross Validation

Q5. Mark all the correct choices regarding a cross-validation approach:

A 5-fold cross-validation approach results in 5-different model instances being fitted.
A 5-fold cross-validation approach results in 1 model instance being fitted over and over again 5 times.
A 5-fold cross-validation approach results in 5-different model instances being fitted over and over again 5 times.
Uses K-different data-folds.

Cross Validation

Q6. Mark all the correct choices regarding the approach that should be taken to compute the performance of K-fold cross-validation:

We compute the cross-validation performance as the arithmetic mean over the K performance estimates from the validation sets.
We compute the cross-validation performance as the best one over the K performance estimates from the validation sets.

Cross Validation

Q7. A data-scientist who is interested in classifying cross sections of histopathology image slices decides to adopt a cross-validation approach he once read about in a book.

Cross Validation

Q8. 1. True or false: The leave-one-out cross-validation (LOOCV) approach is a sub-case of k-fold cross-validation wherein K equals N , the sample size.

True or false: It is always possible to find an optimal value $n$ , $K = n$ in K-fold cross-validation.

Cross Validation

Q9. What is the main difference between RandomizedSearchCV and GridSearchCV?

Cross Validation

Q10. When would you prefer to use RandomizedSearchCV over GridSearchCV, and vice versa?

Cross Validation

Q11. What are the advantages of RandomizedSearchCV?

Cross Validation

Q12. What are the advantages of GridSearchCV?

Cross Validation

Q13. What is cross-validation in the context of hyperparameter tuning?

Cross Validation

General Concepts In Machine Learning

Q1. Below Fig depicts two different cross-validation approaches. Name them.

Q2. 1. What is the purpose of following Python code snippet?

Q3. **True or False:** In a K-fold CV approach, the testing set is completely excluded from the process and only the training and validation sets are involved in this approach.

Q4. **True or False:** In a K-fold CV approach, the final test error is:

Q5. Mark all the correct choices regarding a cross-validation approach:

Q6. Mark all the correct choices regarding the approach that should be taken to compute the performance of K-fold cross-validation:

Q7. A data-scientist who is interested in classifying cross sections of histopathology image slices decides to adopt a cross-validation approach he once read about in a book.

Q8. 1. **True or false**: The leave-one-out cross-validation (LOOCV) approach is a sub-case of k-fold cross-validation wherein K equals N , the sample size.

Q9. What is the main difference between RandomizedSearchCV and GridSearchCV?

Q10. When would you prefer to use RandomizedSearchCV over GridSearchCV, and vice versa?

Q11. What are the advantages of RandomizedSearchCV?

Q12. What are the advantages of GridSearchCV?

Q13. What is cross-validation in the context of hyperparameter tuning?

Q14. Can you combine RandomizedSearchCV and GridSearchCV techniques for hyperparameter tuning?

Q3. True or False: In a K-fold CV approach, the testing set is completely excluded from the process and only the training and validation sets are involved in this approach.

Q4. True or False: In a K-fold CV approach, the final test error is:

Q8. 1. True or false: The leave-one-out cross-validation (LOOCV) approach is a sub-case of k-fold cross-validation wherein K equals N , the sample size.