Machine Learning topic 7.(Random Questions) Answer each o f the following questions in at most two briefsentences.(a) Why do we expect predictive models with less features to be more generalizable?(b) In training a classifier, after we divide the samples into a training set and a test set, we oftenfurther divide the training set into a training set used for model building and a validation set.What purpose does the validation set serve?(c) Assume that we have 100 samples and we are using 5-fold cross validation. How manydifferent models will be built? How many training samples will be used to train each model?

Your question has three multiple question. Let's solve all questions with explanation. Question-(1) why we experts predict models with less features with more generalized. Answer:- without using more features we get same output is more generalized than using more features. Less features means less dimensions, with less dimensions required less space. So that we get accurate output. Question-(2) in training classifiers after we divide the sample into training set and testing set, we often future divide the training set into training set and validation set. What the purpose does validation set serve? Answer:- validation set use training set to goal of finding best model and tune the parameters. Validation set are used to select and tune the final model. Question-(3) Assume that we have 100 samples and we are using 5-fold cross validation. How many different models will be built? How many training samples will be used to train each model? Answer:- there are total 100 sample with 5-fold cross validation. So that there are 20 sample has each fold. So that there will be 5 different model are built. There 4 different training samples wil be used to training each model. Other one is used for testing set. These are all that you want. Thank you.