วีดีโอ

Yes it helps student think about nonlinearity and shows how nonlinearity can be introduced even in a linear regression model. It's also pretty common in quick and dirty forecasts to use squared (and higher terms.) However, over fitting quickly becomes a risk.

@@statswithgavin6297 you've explained the concepts really well. Thank you 🙏

We use this book in the statistics class that I teach and you can find them there: www.amazon.com/Statistics-Business-Economics-David-Anderson/dp/1337901067?ref=d6k_applink_bb_dls&dplnkId=65b8466f-2a03-4fa3-acd9-5bee417b0c2c

Small edit required: The #1 rule of forecasting (when forecast is performed by a Supply Chain Manager) is it is NEVER right :) as brought to us by Supply Chain Management 3050

You should be able to download the data here: www.zillow.com/research/data/

No one's getting younger. Except for Musk. He appears to be getting younger...

18:40 should be "if you estimate the standard deviation with the sample standard deviation, use a t-stat." You could have a problem where a sample mean is given but a population standard deviation is provided/assumed rather than calculating a sample standard deviation.

If it's a log-linear regression with just one independent variable and you are building the intervals around logged predicted values then you don't need to make any adjustments to the method in the video.

@@statswithgavin6297 Thanks!

You're very welcome. I'm glad it was helpful.

I'm glad it was helpful!

Glad to hear it was helpful!

You're welcome! I'm glad you found it helpful.

No problem!

You can use ordinary multiple regression but it might lead to biased estimates, so most people would use something else. For example an ordered probit or ordered logit model. This would be very complicated to set up in Excel so I would suggest using different software such as Stata or R if you decided to do that. en.wikipedia.org/wiki/Ordered_probit en.wikipedia.org/wiki/Ordered_logit

@@statswithgavin6297 I have changed my variables entirely. My independent variable is the respondents' monthly income and my dependent variable is their total expenditure (for the last 12 months). I ran the tests using simple linear regression and pearson correlation and found that they have no correlation or significant relationship at all.. Is there something wrong with my data, my tests, or is that an uncommon result? I dont really know to move forward 😕

@@MangoAndMucho Standard OLS regression, like used in video, will be a better fit for that data.

You're welcome, Raja. I'm glad it was helpful.

They probably see this behavior as pretty standard.