Supervised learning models need to find out optimal hyperpar…
Supervised learning models need to find out optimal hyperparameters during the split-test to minimize overfitting and evaluating the predictive performance. Overfitting means a machine learning model is trained with the training dataset and performs well with the training dataset. However, the trained model does not perform well for new data (out-of-sample prediction; generalization). During the training step in the split test, we find out the optimal hyperparameters to avoid overfitting, If we select the optimal hyperparameter during the training step, the optimal hyperparameter should be selected at Area B (dash line). In addition, in this case (training step), the generalization loss is validation set loss. During the training step in the hold out split test for a supervised learning model, • 1. SL model trained with the (1) ___________ dataset • 2. Trained SL model predicts labels in (2)__________ dataset using features in (2)__________ dataset • 3. Repeat the process steps 1 and 2 with different hyperparameters • 4. Finding out the best hyperparameter for the SL model to avoid overfitting During the test step in the hold out split test for a supervised learning model,, • 1. Training SL model with (3)___________ dataset with the optimal hyperparameter • 2. Predict labels in (4)_________ dataset with the trained SL model in Step 1 • 3. Evaluate the model performance
Read Details