From the course: Problem Identification and Solution Design for Data Scientists
Join today to access over 24,800 courses taught by industry experts.
Estimates of accuracy
From the course: Problem Identification and Solution Design for Data Scientists
Estimates of accuracy
- [Instructor] I'm going to suggest avoiding obsessing over estimates of accuracy. Many think that the key to getting a project approved is to determine that an accurate model is possible. It seems to make sense, but you'll manage your time poorly if you worry about this too soon. The key to model accuracy is assembling and preparing the data, and you don't have time to do that properly at this point. So I wouldn't share any modeling outcomes at all during this stage. I know how tempting it is, especially if you get some preliminary good news, you'll want to share it. But how are you going to explain if the accuracy actually goes down later? And that might happen, the accuracy might be artificially inflated. This phenomenon is sometimes called data leakage, and it's easy to make that mistake at this stage because you haven't done data prep yet. Or it might be that you're looking at a small data sample, but on a more comprehensive data sample, the model is less impressive. Your good…
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.