Ridge regression explained: Regression robust to multicollinearity (Excel)

What can one do when independent variables in a regression model are non-orthogonal or even near-multicollinear? How to address the issue of multicollinearity and make your regression model robust to it? A common technique is the ridge regression as proposed by Hoerl and Kennard (1970). Today we are learning how to apply ridge regression in Excel, comparing the results with OLS, and discussing the bias-variance tradeoff inherent to the problem.
Don't forget to subscribe to NEDL and give this video a thumbs up for more videos in Econometrics!
Please consider supporting NEDL on Patreon: / nedleducation

[@NEDLeducation]

You can find the spreadsheets for this video and some additional materials here: drive.google.com/drive/folders/1sP40IW0p0w5IETCgo464uhDFfdyR6rh7
Please consider supporting NEDL on Patreon: www.patreon.com/NEDLeducation

[@jacopotonziello2879]

Hi, thanks for the interesting video. if you have the chance to check i think there is a slightly mistake in the spreadsheet in excel page Ridge_additional, when you evaluate the standard error of (constant, population, capital stock) within the very long formula i think is missing the (lambda x I) of the final part of Z matrix.

[@MrMahankumar]

Great!!!!

[@gannikim-jd6zx]

Hi, can you do the same practical but in SPSS? its gonna really help me

[@azyzmeddeb9157]

For k optimum please, have a question, you said number of variables which is 2, multiplied by error-type squared. You used the number of parameters. So should I use the number of parameters or the number of explanatory variables. Thanks.

[@NEDLeducation]

Hi Azyz, and thanks for the question! It is the number of parameters.

[@EliteBestGamers]

Hello NEDL, thanks for the video!! Can you do some videos related to cryptocurrencies? I mean, some forms of predicting or risk metrics...

[@NEDLeducation]

Hi, and thanks for the question! I have made a tutorial on the past with regards to cryptocurrency mining profitability: ua-cam.com/video/lrQzbFlBSmk/v-deo.html In general, many techniques applicable to conventional assets can be generalised to cryptocurrencies. For example, you can model their returns using fat-tailed distributions such as Cauchy, Johnson's SU, or error (please check out my modelling stock returns series in the Mathematical Finance playlist). Alternatively, if you are interested in something in particular, let me know!

[@EliteBestGamers]

@@NEDLeducation Thank you very much for the answer NEDL. Perfect... I will check it out.
Personally speaking I would like to see if hurst coefficient is a good indicator to check for tendencies. I tried to create in python but it seems that it's wrong. Maybe combining hurst coefficient and another indicator together that is related forecast could be interesting. Do you see interesting combinations of indicatores?
Thanks

[@NEDLeducation]

@@EliteBestGamers Hi, and thanks for the follow-up question! I have loads of videos on Hurst exponent and its applications actually, here is a video on the Hurst exponent in Excel (more conceptual): ua-cam.com/video/l08LICz8Ink/v-deo.html and in Python (more applied): ua-cam.com/video/v0sivj2wGcA/v-deo.html.