From the course: Machine Learning and AI Foundations: Classification Modeling

The importance of binary classification

From the course: Machine Learning and AI Foundations: Classification Modeling

Start my 1-month free trial

The importance of binary classification

“

- [Instructor] Let's start by dispelling two myths. For many, there's an implicit assumption that there's one best algorithmic approach for each use case, and if you simply define the problem carefully, you'll know which algorithm to use. As we'll see, there's a one-to-many relationship between problems and potential solutions. Just in this short discussion of classification techniques, we'll explore a dozen different approaches to a single-use case, binary classification. Another myth is that the goal of predictive analytics projects are to produce a single predictive model. Real-world projects always have multiple models performing different functions and often sending information back and forth to each other, working together as a decision management system. In virtually every instance, at least one of these models is a binary classifier. In fact, in my 20 years of doing this kind of work, it's always been present on the projects I've been a part of. Why? Well, the ultimate goal of a predictive analytics project is to drive decisions, and the easiest way to do that is to have at least one of the models tell you when to take action and when not to. These two themes and many others are explored in my course Essential Elements of Predictive Analytics and Data Mining. So for now, we'll just accept the following premise. Binary classifiers play an important role in virtually every project, so understanding them constitutes a critical part in anyone's professional development in predictive analytics, data science, and data mining.

Contents