Cross-validation

Cross-validation is a way to evaluate the accuracy of a model on a dataset that was not used for training, that is, a sample of data that is unknown to trained models. This ensures generalization of a model on independent datasets when deployed in a production environment. One of the methods is dividing the dataset into two sets—train and test sets. We demonstrated this method in our previous examples.

Another popular and more robust method is a k-fold cross-validation approach, where a dataset is partitioned into k subsamples of equal sizes. Where k is a non-zero positive integer. During the training phase, k-1 samples are used to train the model and the remaining one sample is used to test the model. This process is repeated for k times with one of the k samples used exactly once to test the model. The evaluation results are then averaged or combined in some way, such as majority voting to provide a single estimate.

We will generate a 5 and 10 fold cross-validation on the...