
So this is actually a surprisingly challenging question if you've never done

this before.

So let me dissect it for you first.

I'm going to only look at the very first feed over here and

ask how much does Y grow as you vary X.

And you can see as we add 20 to X, from 20 to 40, or another 20 to X.

Every time, x1 goes by 20, y goes by 100.

So the factor that goes in over here is y divided by x.

It is a five.

I hope you got that right.

As we do the same thing for the x2 variable, we find when x2 goes by 20.

Y only goes up by 50.

So we divide 50 by 20 and divide, at 2.5.

Now if we left it at this and believe this is the formula and

this number of here doesn't exist then what we will predict for

this value over here is 20 times 5 plus 80 times 2.5 and that is 300.

So we get 200 too large.

So we have to subtract minus 200.

And it turns out, the way I made this table,

this is actually the correct equation.

So you can check any value in this table over here and you'll find that

this specific linear equation exactly explains the value over here.

Now that's an artifact.

In reality we don't have such nice situation, where a,

what we call a linear hyperplane explains data so well.

But it gives you a sense that the predictor variable, in this case,

net worth, could be a factor of more than one input variable and

that is called multivariate regression.

And all you have to know about this,

there's a piece of software that just does it for you.

So you can actually plug data like this into a table and

out comes the correct result.