search

QCBSRworkshops / workshop04

Workshop 4 - Linear models
https://r.qcbs.ca/workshops/r-workshop-04/
Other
9 stars 57 forks source link

Model explanation 11 #16

Closed dschoenig closed 4 years ago

dschoenig commented 4 years ago

Fixes issue #11 for the French and English versions. For details see discussion of the issue.

pedrohbraga commented 4 years ago

Hi, Daniel (@dschoenig)! Thank you so much for this! I am going to wait for Travis CI's pass on the workshop, which should take a few hours. Once it is done, we will approve, merge the pull request and get back to you about the changes you have added!

LaurieMaynard commented 4 years ago

I found some typos and here and there and some slide missing translation. I can fix it on this branch before you merge it.

LaurieMaynard commented 4 years ago

My changes has been implemented into this branch and it should be good to be reviewed and merge with the main branch

dschoenig commented 4 years ago

Great, thanks for having another look at this, Laurie!

dschoenig commented 4 years ago

Going through the slides there was only one potential inaccuracy I noticed: The condition for the residuals is that they should be independent and identically distributed -- I think the equivalent French term would be residus idépendants et identiquement distribués (see https://fr.wikipedia.org/wiki/Variables_ind%C3%A9pendantes_et_identiquement_distribu%C3%A9es), as uniformément distribués (current version line 586) would imply that they come from a uniform distribution when they come in fact all from the same normal distribution N(0, σ²).

However my knowledge on the correct statistical terms in French is definitely spotty, so it would be great if someone of you could shed some light on this!

LaurieMaynard commented 4 years ago

Sorry, I know you wrote identiquement, but I changed it because I never heard that word and thougth it didn't exist (I just check and apparently it does). When I wrote uniformément distribué, to me it meant that the points are spread evenly across the plot. I never heard identiquement in that context, but we can put it back. If not, maybe another word would be better, such as "également" not to confuse anyone.