Consider that you are training a classifier for a 10-class classification problem. Each I/P is represented as a 512 dimensional vector. There are 1000 samples out of which first 100 will be used for testing. Let Leave-one-out-Cross-Validation (LOOCV) be used for selection of the classifier model before testing. Which of the following option is the correct option for no. of validation split that will be generated?
Correct Answer :
900
Solution :
The correct option is 900.
To understand why this is the correct answer, let's break down the process step-by-step:
1. Identify the total number of samples and the split between training/validation and testing:
We are given a dataset with a total of 1000 samples.
The problem states that the first 100 samples are reserved and used exclusively for testing the final model.
Therefore, the remaining samples available for training and model selection (validation) are:
Total training/validation samples = 1000 - 100 = 900 samples.
2. Understand Leave-One-Out Cross-Validation (LOOCV):
LOOCV is a specific type of K-fold cross-validation where the number of folds, K, is equal to the number of data points available for training/validation.
In each split of LOOCV:
- Exactly 1 sample is held out as the validation set.
- All the remaining samples (N - 1) are used to train the model.
This process is repeated such that every single sample in the training/validation pool is used as the validation set exactly once.
3. Calculate the number of validation splits:
Since there are 900 samples available for model selection (after excluding the 100 test samples), and LOOCV selects exactly one sample at a time to act as validation, there will be exactly 900 unique ways to choose the validation sample.
Thus, the number of validation splits generated is equal to the number of training/validation samples:
Access expert-curated educational resources and study materials—completely free.
Create, conduct, and manage professional online assessments with Mindyard. Perfect for teachers and institutes.
Copyright © 2026 Mindyard. All Rights Reserved.