Modern Scala Projects

Features and their importance

Each feature in a multidimensional dataset is a contributing factor in arriving at a prediction:

• A prediction is to be made on a new sample; for example, a new breast cancer mass sample belonging to a certain individual 
• Each contributing factor has a certain feature importance number or feature weight

Some features are more important than others in contributing to the final prediction. In other words, a (final) prediction is made on what category a new sample belongs to. For example, in the breast cancer dataset in Chapter 2, Build a Breast Cancer Prognosis Pipeline with Spark and Scala, the Random Forest algorithm can be used to estimate feature importance. In the following list the top-most features have ...