From the course: Applied Machine Learning: Supervised Learning
Join today to access over 24,900 courses taught by industry experts.
Metrics for classification and regression
From the course: Applied Machine Learning: Supervised Learning
Metrics for classification and regression
- [Instructor] We've talked about what supervised learning is, we've talked about regression in classification, we've talked about how we have to have our data in this tabular format of X and Y. Let's talk about metrics for our model. Metrics are a way to evaluate the performance of our model. Oftentimes, folks will ask me, is a model good enough, given a given metric? That's often a hard question to answer with a basic metric like something like accuracy. Oftentimes, I'll tell my clients that they want to create a metric that ties business value to a model. Although we will use metrics to compare models, and if a model has a better metric, we can often say that model is a better model. Because regression and classification are different, we will use different metrics for them. Some common metrics for regression are the mean absolute error. What this represents is when we are making predictions on data that we haven't seen, what is the delta from the true value to the predicted value?…
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.