I partioned the data into 2 viz. Test and train.
But is it imorptant to do validation of data.
I am using WBDC dataset (of 532 tuples).
Usual practice is to partition the dataset into three such as 1. Training, 2.Validation, 3. Testing
Validation set is really important to prevent the model you train from overfitting to the training data. If you have a validation set you can validate the model during training wheter it is overfitting / not. With this you can have an idea where you need to stop training in order to get a generalized model
Usual practice is to partition the dataset into three such as 1. Training, 2.Validation, 3. Testing
Validation set is really important to prevent the model you train from overfitting to the training data. If you have a validation set you can validate the model during training wheter it is overfitting / not. With this you can have an idea where you need to stop training in order to get a generalized model
- Badges
- Science topic
- Similar topics
- Organizational Psychology
- Training