00:10:43 jonathan.bratt: :-D
00:11:02 Laura Rose: chapter 6 is really long!
00:11:07 jonathan.bratt: Yes, it is!
00:11:38 Laura Rose: I can probably not do 12/21 either, due to another meeting
00:11:50 jonathan.bratt: second
00:11:52 Laura Rose: 2
00:12:52 SriRam: except Ryan :D
00:13:17 Ryan Metcalf: People run away if I were to grow a beard! Patchy!
00:13:24 SriRam: :D
00:35:32 jonathan.bratt: > morphemepiece::morphemepiece_tokenize("unbiasedly")
[[1]]
un## bias ##ed ##ly
11995 4903 4030 4057
00:36:06 Jon Harmon (jonthegeek): lol I did the same thing Jonathan
00:36:32 jonathan.bratt: 😆
00:36:48 Jon Harmon (jonthegeek): (that's a package Jonathan and I are working on for NLP)
01:01:05 jonathan.bratt: That was really good.
Meeting chat log
00:34:42 Federica Gazzelloni: tidy models: https://www.tmwr.org/dimensionality.html
00:35:05 Ryan Metcalf: https://www.tmwr.org/dimensionality.html#partial-least-squares
00:35:09 Federica Gazzelloni: pca for unsupervised
00:35:18 Federica Gazzelloni: pls for supervised
6.6.2 Cohort 2
Meeting chat log
00:08:25 Jim Gruman: Hello
00:08:36 Federica Gazzelloni: Hi
00:17:45 Michael Haugen: Hi, Sorry I am late.
00:35:35 Ricardo: Check the broom package
00:36:25 Ricardo: https://broom.tidymodels.org/reference/augment.lm.html
00:36:56 Jim Gruman: +1 yeah, to uncover the method help details, you can run class() on the object to see the S3 class, and then ask for help on autoplot.class to see the specific help on the autoplot in the dependent package
00:50:40 Ricardo: PCA is unsupervised, PLS is supervised (check the outcome argument)
6.6.3 Cohort 3
Meeting chat log
00:07:07 Nilay Yönet: https://emilhvitfeldt.github.io/ISLR-tidymodels-labs/resampling-methods.html
00:33:41 Mei Ling Soh: Maybe we can wrap up the lab soon?
00:37:48 Fariborz Soroush: 👍
01:04:24 Mei Ling Soh: Two more minutes to go
01:05:19 Nilay Yönet: https://onmee.github.io/assets/docs/ISLR/Resampling-Methods.pdf
01:05:22 Nilay Yönet: https://waxworksmath.com/Authors/G_M/James/WWW/chapter_5.html
Meeting chat log
00:23:20 Mei Ling Soh: Each resample is created within the stratification variable. Numeric strata are binned into quartiles.
00:34:05 Jeremy Selva: https://dionysus.psych.wisc.edu/iaml/unit-06.html#elastic-net-regression
00:54:24 Jeremy Selva: Here are the link to the slides and the lab
https://jauntyjjs.github.io/islr2-bookclub-cohort3-chapter6
https://jauntyjjs.github.io/islr2-bookclub-cohort3-chapter6-lab/
Source codes for the slides and the lab
https://github.com/JauntyJJS/islr2-bookclub-cohort3-chapter6
https://github.com/JauntyJJS/islr2-bookclub-cohort3-chapter6-lab
00:55:57 Rose Hartman: Thanks, Jeremy!
00:56:01 Nilay Yönet: thank you Jeremy!
6.6.4 Cohort 4
Meeting chat log
00:38:01 Ronald Legere: https://stats.stackexchange.com/
Meeting chat log
00:06:57 Kevin Kent: https://noamross.github.io/gams-in-r-course/
00:27:23 Ronald Legere: https://stats.stackexchange.com/questions/410849/when-does-partial-least-squares-provide-1-component-solutions
6.6.5 Cohort 5
Meeting chat log
00:05:47 Lucio Cornejo: Hello Angel
00:07:22 Lucio Cornejo: Arnab, can you hear us?
00:07:48 Arnab: No :/
00:08:14 Arnab: Can anyone hear me?
00:08:43 Arnab: I am getting a lot of echo
00:44:58 Derek Sollberger (he/his): Fig. 6.7: Lasso (L1 norm, left picture) coefficients will meet RSS on the axes … while ridge (L2 norm, right picture) RSS will intersect at a combination of coefficients (i.e. reinforcing why lasso forces some coefficients to be zero, while ridge does not)
00:45:39 Lucio Cornejo: https://stats.stackexchange.com/questions/318322/lp-regularization-with-p-1
01:02:45 Lucio Cornejo: no comments from me. Great presentation, thanks