What is model training?
Model training is the process of feeding the data to a machine learning algorithm, thus creating a model capable of performing a specific task with the data. Model training must necessarily include a particular algorithm (for example, a decision tree) and some data. The more data is fed to the algorithm; the better are the results (but it can also lead to overfitting). This is the most critical step in the machine learning workflow. However, it is not feasible without all the other steps.
Why is it important?
Model training produces a machine learning model, which is the purpose of any ML project. The model can later be used for practical tasks (depending on the data and algorithms used).