From the course: AWS Certified Machine Learning - Specialty (MLS-C01) Cert Prep: 3 Modeling
Unlock this course with a free trial
Join today to access over 23,400 courses taught by industry experts.
Avoid overfitting and underfitting - Amazon Web Services (AWS) Tutorial
From the course: AWS Certified Machine Learning - Specialty (MLS-C01) Cert Prep: 3 Modeling
Avoid overfitting and underfitting
- [Instructor] Let's take a look at the concept of overfitting versus unfitting. In one way, you can think about a underfitting problem as the line is not directly following the points of a scatter plot. So if this is a scatter plot here and we have a bunch of data, notice there's a slightly superlinear pattern here, where there's some data that's appearing up here. And if you're putting your regression line through this part of the data set, you're missing out on predicting, potentially, some of these other values as they rise up. Likewise, if you wanted to create a more balanced model here, notice that have a slight curve to the data and you can see that the line is appearing directly through the middle of the dataset. So you can call this a balanced prediction. There's a overfitting problem, though, as well, that can occur and that's when you try to too closely match the data. And so the prediction could be…