What is Feature Selection in Machine Learning and How is it Used?

Real-world application of data analysis is a far cry from what we are taught in classrooms or online courses. Whereas the data we are shown in educational settings is typically homogenous and presented in tabular columns where all rows are filled, when it comes to real-world applications we often get datasets that are not unlike a typical city street (patchy, filled with potholes, and dotted with diversion signs). In these cases, we are generally faced with thousands of rows and columns that require a lot of clean up (for example, deciding what to keep and what to discard). This type of dataset is often referred to as a high dimensional dataset, and it can create a number of headaches. For example, this high dimensionality will significantly increase the training time of your machine learning model; making your model very complicated which in turn may lead to overfitting.