100daysbytiffany.txt

Tiffany Fabianac
Tiffany Fabianac
Laboratory Informatics Consultant, Data Science Graduate Student
18h • Edited
Linear Regression for prediction in R:

r = cor(X, y)
slope = r * sd(y) / sd(X)
intercept = mean(y) - slope * mean(X)

r is the correlation coefficient and represents how closely related two variables are. An r of 0 means that the two variables are not related, negative r tells us that as one variable gets big the other gets small, and positive r says that both variables get bigger together.

A regression line allows us to predict y based on the relationship between X and y. This is only really helpful if X and y are linearly related quantitative variables, so we can see that if there is no correlation between X and y then r ~0, slope ~ 0, and the prediction of y can only be assumed to be mean(y). This is why we want to check for independence. 

hashtag#100DaysOfMLCode