Train, score, and evaluate
Before making the prediction, you need to train an algorithm with the example data or training dataset where the target value or the label is known. After training the model, you can make a prediction with the trained model.
Continuing with the preceding illustration, the trained model may be considered as the mathematical function f to make a prediction.
Usually, when you have to build a model from a given dataset, you split the dataset into two sets and use one as a training dataset and the other as a test dataset. After the model is trained with the training data, you use the test dataset to see how the model is performing, that is, how many errors it has.
After the model is trained, you can use the test data to make a prediction or to score. In scoring, the feature values are used and then the target value is predicted. At this point, you are not sure how your model is performing. You need to evaluate it to find out its performance. During evaluation, you take...
