From the course: Predictive Analytics Essential Training: Data Mining
Join today to access over 25,500 courses taught by industry experts.
Establishing proof that the model works
From the course: Predictive Analytics Essential Training: Data Mining
Establishing proof that the model works
- [Instructor] We've seen that you want to conduct your search open to unanticipated patterns and solutions, so you don't begin with a hypothesis. But nonetheless, you need evidence that your a model is a good one before you deploy it. In some ways it's a bit more like engineering than it is like science. You need to take your model out on a test drive. So our next element is proof that we're right. We've already discussed that data mining by its very nature does not have a priori hypothesis. A priori is just a fancy Latin phrase, meaning hypothesis from theory and not from experience. It's commonly used in statistics. Even better, you can think of a priori as having the hypothesis before you observe. But we are basing our model on observation. We're examining and exploring the data. So we are doing something quite different. Yet we still need proof. Is this some kind of contradiction? No theory based hypothesis, yet…
Download courses and learn on the go
Watch courses on your mobile device without an internet connection. Download courses using your iOS or Android LinkedIn Learning app.