Model selection with AICs

[@farhanaakhter5532]

Thank you for your explanation. Can we use this equation for calculation AIC when the residuals are not normally distributed?

[@hansvanlennep6376]

Thank you for your clear explanation! Can I ask you a question about your example? I was wondering how the AICc of your quadratic model is lower then you linear model? Because if you look at the CI's of your coefficients in the quadratic model they seem to be bigger then the coefficients of the linear model. Now if i'm not mistaking, this means that the residual sum of squares (used to calculate the CI's for these coefficients) in the quadratic model must be larger then in the linear model? Or does the formula in the AICc utilize a total residual sum of squares? That is still not fully clear to me.... I hope that you can elaborate on this question.

[@aliciahamer154]

The graph at 6:23, where did you get it from? I would like to mention it in my report.

[@bubuchocolate]

Thanks for the video! How did you calculate Delta?

[@xsnowolflyx]

Thank you, your presentation is very clear and easy to follow :)

[@__dekana__]

Thank you so much for this!

[@da2015]

Amazing!

[@martinsahmed9107]

Nice example. But I was expecting to see the data before the analysis. How did you come about those values? Some of us have research papers we are working on. Having an example like this will surely help. But it seems you forgot to explain the nature of the data you used and how you came about the values. we just saw them. How did you come about them? Pls explain.

[@shilpalal9668]

Thank you so much...very clear and very informative

[@gonzalomartinezherrera4451]

Thank you! The explanation is very clear.

[@kahuilim2122]

Thank you! This was so thorough!

[@ellenpasternack9750]

This is so helpful, thank you!

[@kennethburnham9247]

This video is very nice, I like it. Interested persons might look at material on the web sitesites.warnercnr.colostate.edu/kenburnham/model-selection-issues/One point - AIC is widely accepted and used in many contexts, not just ecology/wildlife.

[@methodsinexperimentalecolo5000]

Thank you, sir - as one of the leading proponents and explainers of this approach, your feedback means much. I note that the course I teach is called Methods in Experimental Ecology (thus the ecology emphasis), but yes! This approach applies generally. To anyone reading this, try to refer to the model with the lowest AICc as "most plausible" - I say "best" in the video too often, despite Dr. Burnham's advice.

[@DanielEinor]

thank you for your video!!!

[@brazilfootball]

I bet those peaks every 4-5 years in web searches correlate with statistics students trying to graduate lol

[@monellian]

Thank you!

[@alanhhernandezcastillo6639]

chingon

[@zandere1]

Sorry, but I don't see any of AIC or AICc explained here. The formulas just pop out of thin air without ever being explained, why the terms are in there and what they really mean and do in the equation. Then there is talk about AICc correcting AIC for sample size "because there is some n in the denominator", but, sorry, that is no explanation at all. What does this term correct and why exactly does it take on this form? I could write down any old formulas with n's all over the place, and contend that they "correct" AIC...

[@A-human-like-you]

Did you find a better source that talked about this? If you did can you share it please. Thanks

[@zandere1]

@@A-human-like-you Sorry, no. Not on youtube...