Linear Regression Model Residuals Quadratic Regression Model...

Linear Regression Model Residuals Quadratic Regression Model Residuals Exponential Regression Model Residuals A set of data was used to create a linear, a quadratic, and an exponential regression model. The residual plots for the three models are shown above. Based on the three residual plots, which of the following could be an appropriate model for the data? y = -0.3 + 2.4 log x y = 5.7 - 2.9x y = -0.3(1.72)^x y = -1.3x^2 + 2.1x + 0.3

Transcript

00:01 We want to model a set of data, and we've got some residuals shown to us.

00:07 Linear, quadratic, and exponential.

00:16 Based on this, what is a plausible model for the data? so we have four options here.

00:23 We've got one using logarithms.

00:25 One is linear, so if i get the rhythm, i'll label what they are, a, b, c, and d.

00:31 So a is a logarithmic one.

00:34 B is linear.

00:36 We have no exponential on x, it's just x as it is.

00:41 Y is an exponential one, something to the power of x, which is quite rare, really.

00:51 It's found in things like radioactive decay or population growth.

00:55 And finally, d is quadratic.

00:58 Notice we have that x squared term in there.

01:01 That's the highest exponential that x gets to, so that's a quadratic.

01:07 Okay, so we've got these residual models.

01:11 If i look at these, what do i see? well, linear, i don't really see a pattern.

01:15 They're just somewhere above zero, somewhere below.

01:18 So no obvious pattern.

01:20 If i look at the quadratic, i see a line, like so.

01:26 If i look at the exponential, i see a curve.

01:29 Really, it looks kind of like a quadratic, like that.

01:32 Based on that, which one is appropriate to use here? well, what is a residual? a residual is the predicted value, y hat, minus the actual value.

01:46 This represents the error for each data point that we're using to make our model...

Question

Please give Ace some feedback