Introduction to general linear models

[@carolinacampos7988]

The explanations in your videos are clear and helpful for those of us that do not find stats intuitive. I am amazed your videos do not have at least 100,000 views.

[@iriswang8401]

can't agree more!

[@simplicityandchaos976]

This is the best video I have ever encountered on the internet on Generalized linear models

[@riddhimatanwar3533]

THANK YOU SO MUCH!!! THIS HELPED ME A LOT!!!!!!!!! I CAN'T THANK YOU ENOUGH!

[@dagemayele1727]

Very helpful videos.... it is surprising the less number of views???? I expect more than 80000 views

[@funmilayosaporu7853]

Can we have a hard copy. Good lecture

[@buildlackey]

great explanation ! now i understand relationship between linear algebra and stats stuff i just (re)learned from brandon foltz / sal khan / statquest. You are in esteemed company. thnx !

[@yousefmohammadi5982]

Very well explained. Thank you 😊

[@copernicus99]

Thanks for this very helpful lecture!

[@Tascioni49]

Super simple and clear. Thank you.

[@rezat.ashtiani1338]

Thank you so much! You make it so easy

[@funmilayosaporu7853]

I need a hard copy

[@rmttbj]

Can non-linear regression also fall into General Linear Models? In other words, if linear regression falls into GLM, are there any different categorization for non-linear models? Thank you

[@statstree9205]

A general linear model needs to represent a linear combination of predictors. There are some tricks for dealing with non-linear relationships using linear predictors - for example, including x and x^2 terms, or log(x), which are linear representations of nonlinear relationships. There is a limit to the kind of nonlinearities that these tricks can deal with. Some non-linearities cannot be dealt with in a linear model.

[@carolinacampos7988]

@@statstree9205 What would be the best way to deal with non-linearities once linear models have been discarded?

[@statstree9205]

@@carolinacampos7988 This gets trickier. If you just want to visualize the shape of the relationship then you could use a generalized additive model (GAM) such as a spline or loess fit. In other cases, the nonlinearity follows some other known relationship and a custom model could be created for something like that (e.g. a custom package in R perhaps).

[@carolinacampos7988]

@@statstree9205 Thank you very much! I am now studying GAMs so I can apply them.