From the course: Mistakes to Avoid in Machine Learning

Unlock the full course today

Join today to access over 22,600 courses taught by industry experts or purchase this course individually.

Interpreting your coefficients without properly treating for multicollinearity

Interpreting your coefficients without properly treating for multicollinearity

From the course: Mistakes to Avoid in Machine Learning

Start my 1-month free trial

Interpreting your coefficients without properly treating for multicollinearity

“

- [Instructor] Traditional statistics techniques can be useful when you're looking for an explanation of the relationships underlying your data, as opposed to pure predictive capability. But there are many best practices to be mindful of. In particular, the presence of multicollinearity, which can throw off the interpretations of your model input. So if you don't know, multicollinearity is when one predictor variable in your regression model can be accurately predicted from the others. And this can have consequences when you review your results. Now let's see an example of using the iris dataset. We'll import the iris data using seaborne. Then turn this into a binary classification problem by dropping one of the three classes. So I run the cell and there we go. Now we have the iris dataset as a data frame. As we formulate this as a regression problem, we assign species to y, and the remaining variables to x. Recall y is…

Contents