00:51:58 Jon Harmon (jonthegeek): Because I'm obsessed now with "Why ξ?":
"KS" for "knot/spline" = "ks" = "ξ"?
I'm going to try to use that now, at least 🙃
00:54:45 Laura Rose: 👍
01:01:55 Federica Gazzelloni: h(x,ξ) taylor : https://en.wikipedia.org/wiki/Taylors_theorem
Meeting chat log
00:20:42 Federica Gazzelloni: https://en.wikipedia.org/wiki/Smoothing_spline
00:23:23 Federica Gazzelloni: lambda > = 0 is a smoothing parameter,
7.6.2 Cohort 2
Meeting chat log
00:20:40 Ricardo Serrano: Orthogonal expansion in polynomial regression https://www.dropbox.com/s/j39rd74l51q2vch/Chapter12-Regression-PolynomialRegression.pdf?dl=0
00:21:05 jlsmith3: Thank you, Ricardo!
00:21:36 Federica Gazzelloni: @jenny you can have a look at the picture in this article for getting a sense of orthogonal expansion (in this case it is searching for next values that follow these paths) (https://www.researchgate.net/figure/Orthogonal-and-diagonal-node-expansion-in-the-A-search-algorithm_fig2_222666188)
00:22:26 Ricardo Serrano: 👍
00:48:51 jlsmith3: Perfect, thank you Federica!
00:52:02 Federica Gazzelloni: analysis of wage-education relationship
00:52:16 Federica Gazzelloni: brand choice
00:52:28 Jim Gruman: 👍🏼
00:52:37 Federica Gazzelloni: number of trips to a doctor's office
00:54:52 Federica Gazzelloni: Generalized additive models (GAMs) are a powerful generalization of linear, logistic, and Poisson regression models.
00:56:06 Federica Gazzelloni: should we do the lab?
00:59:16 Ricardo Serrano: Let's split the lab problems https://en.wikipedia.org/wiki/Taylors_theorem
Meeting chat log
00:34:11 Ricardo Serrano: https://www.andreaperlato.com/mlpost/polynomial-regression-smoothing-splines/
00:34:25 Ricardo Serrano: https://stats.stackexchange.com/questions/517375/splines-relationship-of-knots-degree-and-degrees-of-freedom
00:34:50 Ricardo Serrano: https://parsnip.tidymodels.org/reference/mars.html
00:35:00 Ricardo Serrano: https://www.rdocumentation.org/packages/stats/versions/3.6.2/topics/smooth.spline
00:35:10 Ricardo Serrano: https://www.youtube.com/watch?v=bESJ81dyYro
00:35:25 Ricardo Serrano: https://www.youtube.com/watch?v=Vf7oJ6z2LCc
00:36:22 Ricardo Serrano: https://emilhvitfeldt.github.io/ISLR-tidymodels-labs/moving-beyond-linearity.html
00:38:13 Federica Gazzelloni: mgcv::gam()
00:38:44 Jim Gruman: 🐕 looking for attention :)
00:44:06 Ricardo Serrano: I’m back!
00:50:57 Ricardo Serrano: lo() - loess fit in a gam model
7.6.3 Cohort 3
7.6.4 Cohort 4
Meeting chat log
00:11:44 Kevin Kent: https://stackoverflow.com/questions/29710525/symbol-in-r-lm
00:49:09 Ron: https://en.wikipedia.org/wiki/Smoothness
00:50:55 shamsuddeen: This chapter is so mathy
00:51:04 Sandra Muroy: :D
01:14:29 Ron: Are we doing "bring your won questions" next week?
01:14:30 Ron: I hope?
01:14:40 Ron: *own
01:18:07 Kevin Kent: yup!
01:22:05 Ron: I was wrong Loess is briefly looked at in the Lab section .
01:26:30 shamsuddeen: I had to leave. See u next week.
7.6.5 Cohort 5
Meeting chat log
00:23:05 Derek Sollberger: (a) jump discontinuity (models do not intersect)
00:23:17 Derek Sollberger: (b) models intersect, but lower complexity
00:23:36 Derek Sollberger: (c) models intersect, restore complexity, consider bias-variance trade-off
00:23:50 Derek Sollberger: (d) clamped cubic spline: ensure slopes match
00:24:00 Derek Sollberger: but doubles amount of equations and coefficients