Linear Regression: One-Line Test

In this video, I'm going to go over the one-line linear regression command that essentially does everything that that linear regression solution just did, but in a single line. And so, I'm going to call this variable, output. And the command here is stats. Then regress with the scipy stats linked here. And so, we just say stats.line regress. We give it our first variable we give it our second variable and then we can just print the output.

And so, the output includes several parts. So we've got slope intercept r value p-value standard error on the slope, standard error on the intercept, and so there's lots of variables here. So we rarely want all of them in particular. We're interested here in the slope, which is just output dot slope and the intercept which is just output dot intercept, so we can run this. And so we can see here our slope and our intercept values. And if we look up here, they are exactly the same. So if I hide some cells and delete, so we can sort of just barely there we can see that there, so here is our manual output, here is our automatic one-line output. And we can see that they're the exact same number. So this is something that you know uses that linear algebra solution comes the exact same answer, but is a lot quicker to implement for us. And then we can also use those outputs. Since they've been stored very nicely in this output function we can use those to make a basic prediction. So let's say we want to predict the nuclear power for a summer in which we had 475 drowning deaths. So we call that xnew.

Our prediction we always call yhat, so this is not the observed data y, but is the predicted version of that data yhat. We can say output.slope times xnew plus output dot intercept. So this is just our equation of a line y equals mx plus b. And then we can print the predicted nuclear power generation. Officially, it is in megawatts. We can run that, and we can see that the predicted megawatt generation is 762 megawatts. If there were 475 drowning deaths, and this is something we could have done with the manual linear regression as well, just change out this output slope to one of to this first beta value and intercept to the second beta value. But for the remainder of this lesson we'll be focusing on these one-line implementations of linear regressions, which as I said are a little bit faster to run.