MCQs

Disclaimer:

Educational Use Only:

The website links provided are to be used strictly for educational and knowledge-gaining purposes only.

No Misuse:

Do not misuse the links or the information contained therein for any commercial, illegal, or unethical activities.

Respect for Websites and Businesses:

Respect the intellectual property and business interests of the websites and their owners. Do not attempt to alter, copy, or redistribute content without proper authorization.

Acknowledge Value of Knowledge:

Recognize and value the knowledge and resources shared by these websites. Use the information responsibly and ethically.

No Endorsement:

The provision of these links does not imply endorsement of the linked sites, their content, or their products and services. Users are responsible for evaluating the suitability and accuracy of the information.

User Responsibility:

By accessing these links, you agree to use the information at your own risk and to comply with all applicable laws and website terms of use.

Overall MCQs for Reference

What is the primary purpose of cross-validation in machine learning?
a. To eliminate training data
b. To increase test accuracy through memorization
c. To assess model performance on unseen data
d. To reduce computational cost
Answer: cgeeksforgeeks+2
In k-fold cross-validation, how many models are trained per process?
a. 1
b. k
c. 2k
d. 1/k
Answer: bplutuseducation+2
Which of the following best describes “leave-one-out” cross-validation (LOOCV)?
a. Each fold contains 50% of data
b. Each fold contains a single data point for validation
c. Each fold is stratified by output category
d. All folds are used for training
Answer: bgeeksforgeeks+1
What does a larger value of k in k-fold cross-validation generally result in?
a. Greater bias
b. Greater variance
c. Less computational cost
d. Smaller training sets
Answer: bsanfoundry+2
What is the major drawback of the holdout method compared to k-fold cross-validation?
a. Higher bias and variance due to a single split
b. Requires more storage
c. Faster training
d. Only works for classification
Answer: astudocu+2
What is the benefit of stratified k-fold cross-validation?
a. Speed
b. Class imbalance is resolved in each fold
c. Model always overfits
d. Dataset size is reduced
Answer: bplutuseducation+1
Which of the following is NOT a purpose of cross-validation?
a. Hyperparameter tuning
b. Detecting overfitting
c. Data collection
d. Performance evaluation
Answer: cstudocu+1
In repeated k-fold cross-validation:
a. The data is split and evaluated once
b. The process is repeated multiple times with different random splits
c. Only used in regression
d. Used for data augmentation
Answer: bsanfoundry+1
Why is LOOCV computationally expensive?
a. Data preprocessing overhead
b. Model is trained n times for n data points
c. Fewer results are aggregated
d. Small dataset requirement
Answer: bgeeksforgeeks+1
When would you prefer holdout validation over k-fold cross-validation?
a. When the dataset is large and computation is restricted
b. When preventing underfitting is the goal
c. When there’s no test set
d. When model performance is constant
Answer: aplutuseducation+1
What is a common value of k for k-fold cross-validation?
a. 2
b. 5 or 10
c. 100
d. 1
Answer: bwayground+2
In k-fold cross-validation, if k=10, what fraction of data is used for validation in any one run?
a. 0%
b. 10%
c. 50%
d. 90%
Answer: bgeeksforgeeks+1
What is the main objective of cross-validation when developing a model?
a. Increase training error
b. Reduce test error indication bias
c. Increase model complexity
d. Select input features
Answer: bexploredatabase+1
Which metric is NOT typically evaluated with cross-validation?
a. Accuracy
b. Precision
c. Recall
d. Model training speed
Answer: dexploredatabase+1
What happens if you use k=1 in k-fold cross-validation?
a. No model is trained
b. All the data is used for validation
c. Equivalent to LOOCV
d. Equivalent to holdout validation
Answer: bplutuseducation+1
Cross-validation helps to:
a. Reduce overfitting
b. Increase feature count
c. Lower training time
d. Increase label noise
Answer: asanfoundry+2
Stratified k-fold should always be used when:
a. Data is homogeneous
b. Classes are imbalanced
c. Features are continuous
d. Model is non-linear
Answer: bgeeksforgeeks+1
After cross-validation, the overall performance metric is usually:
a. Median over all folds
b. Average over all folds
c. Lowest score of all folds
d. Highest score of all folds
Answer: bplutuseducation+1
Cross-validation is NOT useful when:
a. Data is time series
b. Data is tabular
c. Neural networks are applied
d. Hyperparameters are tuned
Answer: ageeksforgeeks+1
In a typical train-test split, what percent of data is commonly allocated for testing?
a. 5%
b. 10%
c. 20-30%
d. 70-80%
Answer: cplutuseducation+1
What is the effect of insufficient cross-validation on model selection?
a. Accurate generalization estimates
b. Overoptimistic results
c. Underestimated performance
d. Increased training speed
Answer: bexploredatabase+1
Nested cross-validation is particularly helpful for:
a. Imbalanced datasets
b. Outlier detection
c. Hyperparameter tuning within validation
d. Data shuffling
Answer: cscikit-learn+1
Cross-validation performance can vary if:
a. Dataset is static
b. Each split is identical
c. Dataset is small or not representative
d. Model is complex
Answer: cplacementpreparation+1
Which of the following is NOT a variant of cross-validation?
a. Stratified cross-validation
b. Cluster cross-validation
c. Monte Carlo cross-validation
d. Sequential cross-validation
Answer: dgeeksforgeeks+1
Which method best reduces data leakage risk during validation?
a. Using the test set for tuning
b. Cross-validation without shuffling
c. Stratified k-fold cross-validation
d. Validating with training data only
Answer: cplutuseducation+1
Cross-validation can be used to compare:
a. Model types
b. Feature extraction algorithms
c. Regularization strategies
d. All of the above
Answer: dsanfoundry+1
Which of the following is FALSE?
a. Cross-validation can identify overfitting
b. Cross-validation uses unseen data in every fold
c. Cross-validation is inefficient on large datasets
d. Testing error is estimated by averaging over all folds
Answer: csanfoundry+2
Leave-P-Out cross-validation differs from LOOCV, because:
a. It leaves out p>1 examples in each fold for testing
b. It randomizes the target variable
c. It is identical to LOOCV
d. It is used only for classification
Answer: ascikit-learn+1
What should be avoided when splitting data for cross-validation on temporal (time series) data?
a. Random splitting
b. Forecasting future from past
c. Training on newest data
d. Sequential validation
Answer: ageeksforgeeks
A major advantage of k-fold over LOOCV:
a. Requires fewer model trainings
b. Always less accurate
c. Eliminates randomness
d. Needs more data points
Answer: aplutuseducation+1
In repeated k-fold cross-validation:
a. Bias is higher than single k-fold
b. Variance is reduced
c. Number of models is decreased
d. Only feature selection is performed
Answer: bgeeksforgeeks+1
Cross-validation helps to assess:
a. The best possible score on the unseen data
b. Variance of the prediction performance
c. Amount of feature leakage
d. Frequency of model retraining
Answer: bstudocu+2
k-fold cross-validation with k=n is called:
a. Holdout validation
b. Leave-one-out validation
c. Bootstrap validation
d. Nested validation
Answer: bplutuseducation+1
For imbalanced binary classification, the best k-fold variant is:
a. Random k-fold
b. Stratified k-fold
c. Standard holdout
d. Leave-P-Out
Answer: bgeeksforgeeks+1
Cross-validation process involves:
a. Splitting data, training, testing, repeating, averaging
b. Only splitting data once
c. Averaging hyperparameters
d. Removing rows
Answer: aplutuseducation+1
Which is NOT a cross-validation benefit?
a. More robust model evaluation
b. Reliable hyperparameter selection
c. Zero computational cost
d. Avoids overfitting
Answer: cgeeksforgeeks+1
What is the first step in k-fold cross-validation?
a. Calculate test error
b. Shuffle and partition data into k groups
c. Set model hyperparameters
d. Train on all data
Answer: bplutuseducation+1
As dataset size increases, LOOCV tends to be:
a. Faster
b. More computationally expensive
c. Less reliable
d. Preferred over other methods
Answer: bgeeksforgeeks+1
K-fold cross-validation repeats the following for each fold except:
a. Training on k-1 subsets
b. Validating on the remaining subset
c. Randomizing output values
d. Averaging validation results
Answer: cplutuseducation+1
Which method gives the most pessimistic estimate of model error?
a. Training data evaluation
b. Holdout
c. k-fold cross-validation
d. LOOCV
Answer: bgeeksforgeeks+1
Cross-validation is inappropriate for which data type?
a. Time series with autocorrelation
b. Images
c. Tabular data
d. Text data
Answer: aplutuseducation+1
If the model performs well on cross-validation but poorly on real-world data, the problem could be:
a. Overfitting to train/test splits
b. Underfitting
c. Model complexity too low
d. Shuffling features
Answer: aplacementpreparation+1
Which of the following is a key indicator of data leakage?
a. High performance only on validation sets
b. Similar results on test data and real-world data
c. Cross-validation results not matching deployed results
d. Low bias
Answer: cplacementpreparation+1
Cross-validation is most informative when:
a. The dataset is small and diverse
b. There are lots of missing values
c. Dataset is already perfectly balanced
d. The model is very simple
Answer: ageeksforgeeks+1
In cross-validation, the variance of the model’s score across folds can indicate:
a. Data quality
b. Model instability
c. Training size
d. Algorithm type
Answer: bplutuseducation+1
Monte Carlo cross-validation primarily involves:
a. Shuffling data and splitting randomly multiple times
b. Sequential validation
c. Using adjacent values
d. No repetition
Answer: ageeksforgeeks+1
Which of the following can cause a misleading cross-validation score?
a. Data leakage
b. Feature engineering on full dataset
c. Not shuffling before splitting
d. All above
Answer: dplacementpreparation+1
Why is cross-validation not recommended for all types of time series?
a. Predictors are not independent
b. Sequential dependence may leak future information into training
c. Decay is not handled
d. K is always too small
Answer: bplutuseducation+1
Which step should always be included in a cross-validation process?
a. Shuffling dataset (except for temporal data)
b. Preprocessing after all splits
c. Data duplication for bigger training set
d. Feature leakage between folds
Answer: ageeksforgeeks+1
If the mean and variance of validation scores are both low, it typically indicates:
a. Model is unstable
b. Model generalizes well and is robust
c. Training data is insufficient
d. Model is overfitted
Answer: bplutuseducation+1

Blog

Systemethic Technologies

Leave a Reply Cancel reply

Useful links

Other Terms & Policies

Contact Info