Machine learning is a subfield of artificial intelligence that uses mathematics to allow a system to automatically learn and improve from data.

It has become wildly popular over the last decade, and it powers many of the technologies you interact with on a daily basis.

Machine learning works by building a model based on patterns identified in a collection of data.

Machine learning algorithms are grouped into three categories: supervised, unsupervised, and reinforcement. 

The result of applying a machine learning algorithm on training data is called a “model”.  Building a model takes multiple steps: data preprocessing, model selection, training, and evaluation.

The first step is to obtain and clean our training data.

Next, we select the right algorithm based on our data’s attributes and the task we are trying to complete.

For each algorithm, there are some settings we must define to deal with our particular task.

Next we train the model on the data, which it uses to automatically learn and improve itself.

Next, we evaluate the model and make refinements to our model’s settings.

This is repeated until we are satisfied with the model performance.  Then it is applied to make predictions on new, unseen examples.