Now, a feature in machine learning and data science represents an individual measurable property or characteristic of the phenomena being observed.

Now there are a few big words in there, so essentially think of a feature as the historical data that you're using in order to help generate a prediction.

If you remember back to when we worked through the decision tree case study we had all of this historical data. So we had the year, make, mileage, fuel, repairs, and service items and so all of this data here is a set of features.

These are the data points that we're looking at when we're building our training model.

Now you may notice that I left off status and that's because status is a label status is something that's different. We're going to actually have a guide dedicated just to label.

So for right now know that in a dataset like this where you have all of this hit historical training data could be in tabular form it could be from a Web site scraper. There's all kinds of different ways you can store it this is a way that is a little bit more straightforward to understand what your set of features are.

But you're going to see features in all kinds of different shapes and sizes as you build out programs but you can think of your features as the attributes and as the data points that you're looking at when you're building out your training model.