The Two Important Elements: Model Training and Validation – Machine Learning Fundamental Concepts

The Two Important Elements: Model Training and Validation

Training is where the machine learns something. As much as possible, the goal is to record the connections between features and their respective labels. Algorithms can be “trained” by making small changes over and over until they fit the data better or capture the relationships that are important.

When the training phase of the model is complete, it is time to put the model through the testing phase.

It is not recommended to use the same dataset for model training and model validation at any point. To build a reliable machine learning model, you need to split your dataset into a training set, a validation set, and a test set.

As was covered earlier in this chapter, there are a wide variety of evaluation metrics that can be applied to the model so that it can be evaluated. The thing that we need to keep in mind is that the purpose of machine learning is to locate a model that comes as close as it can to accurately predicting the label that will be applied.

NoteAlways keep in mind that even the most accurate model can still have a certain deviation.

We can conclude the training and validation process in the following steps:

1.	Develop a training set and a validation set out of the available dataset.
2.	Use a fitting algorithm to create a model from the training data.
3.	Relationships in the data are captured by the trained model.
4.	Predict results focusing on the validation data using the model.
5.	Measure the degree to which clusters are distinguishable from one another or contrast predicted labels with actual ones using evaluation metrics.
6.	Start another cycle from step 1.