home.social

#linearmodels — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #linearmodels, aggregated by home.social.

  1. Look at what I found at thriftbooks... Probably from a library , original 1971 edition (published two years before I was born!) and one of the very good (?best) and clear books written on the topic of #linearmodels @[email protected] @[email protected] @[email protected]

  2. ‼ Announcement: Online Unfold.jl workshop ‼

    📅 09.05.2025
    💶 Free!
    👉🏼 github.com/s-ccs/workshop_unfo
    ❓ rERPs, mass univariate models & deconvolution!

    If you are interested in combined #EEG / #EyeTracking, in natural experiments, sequential sampling models + EEG (e.g. DriftDiffusion), #VR+EEG, - this could be a useful workshop for you!

    #EEG #linearmodels #statistics
    #julialang

    Organized with Romy Frömer (CHBH)
    and the S-CCS lab (@uni_stuttgart)

  3. 📈 Models simplify complex observations by filtering out details that might not generalize to new instances, but… simplification requires assumptions.

    Take #LinearModels: they assume data is fundamentally linear, dismissing deviations as mere noise.

    The art lies in knowing what to keep and what to discard.

    #DataScience #MachineLearning #ml #ai

  4. "The Robust Beauty of Improper Linear Models in Decision Making" lives rent free in my mind. I think about this paper from 1979 ALL. THE. TIME!

    TL;DR: experts can make robust linear models by just picking a few salient features from their experience. See cmu.edu/dietrich/sds/docs/dawe

    In today's parlance the TL;DR would read "feature selection is really important."

    #DataScience #MachineLearning #LinearModels

  5. In today's lecture on #StatisticalModeling, I explained how to define meaningful non-orthogonal hypotheses/contrasts in (generalized) #LinearModels.

    I only learned about the difference between specifying a contrast matrix vs. a hypothesis matrix in this paper:

    How to capitalize on a priori contrasts in linear (mixed) models
    (by Daniel Schad et al., 2020)
    doi.org/10.1016/j.jml.2019.104

    Preprint: arxiv.org/abs/1807.10451