There are different kinds of problems that require a machine learning solution. For instance, our target can be forecasting future outcomes or identifying patterns from the data. The starting point is a set of objects (for example, items) or people (for example, customers of a supermarket). In most situations, a machine learning technique identifies the solution, starting from some features that describe objects/people. The features are numeric and/or categorical attributes, and they are the base of the machine learning model. Having the right features will improve the performance and accuracy of the model, so it is extremely important to define some features that are relevant to the problem.
In this chapter, you will:
Build machine learning solutions
Build a feature data
Clean the data
Explore the defined features
Modify the features
Rank the features using a filter