From the course: Predictive Analytics Essential Training: Data Mining
Join today to access over 24,900 courses taught by industry experts.
Searching for optimal solutions
From the course: Predictive Analytics Essential Training: Data Mining
Searching for optimal solutions
- [Narrator] We've discussed that we don't need hypotheses when we're doing machine learning but what might still be unclear is that it's potentially counterproductive to burden yourself by constantly guessing or speculating about what the relationships might be in letting all of that speculation drive your strategy. The trick is to be very systematic about trying all possible relationships between your input variables and your target variable. So I phrase this element in a way that might be surprising to you at first. You should have nothing to prove. Let me explain. If you're verifying an outcome, certain that you're right, having carefully chosen predictors in advance and you're just curious how well it fits the data, you aren't doing data mining. Of course, it's possible that you do know or think you know exactly what's going on from prior research. But if so, you're not faced with a data mining problem. There's…
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.