Skip to content

Latest commit

 

History

History
102 lines (71 loc) · 2.35 KB

Notes.md

File metadata and controls

102 lines (71 loc) · 2.35 KB

Key Concepts

  • Sample
  • Population
  • Distributions
  • Null & Alternative
  • Central limit theorm

https://openintro-ims.netlify.app/foundations-mathematical.html#foundations-mathematical

Outline

week 1

Idea behind monte carlo Intro / Background / Prerequisites Resources Goals, concepts

Motivations

What questions do we ask when we use statistics? Null models

Main applications & Example uses

Cautionary tales ACF/CCF hummingbirds

Problem set: Give a data set that was generated from a null & ask them to visualize and summarize. Some simple programming - pulling samples, assigning and tracking ids, loops and mapping

Week 2

Sampling from data 1- better names

Bootstrap

The General Procedure

Randomization General workflow Simple designs & examples Show models slide Randomization use cases

Other methods Jackknife Bootstrap Problem set: skills needed - loops, mapping, functions, seeds, tracking ids. Pulling out test statistics, Perform test on a dataset violating assumptions - standard, nonparametric, randomization CI jackknife same as CI on mean hypoth test with jackknife

Week 3

Sampling from data sets: complex designs

Sampling from data 2 Conservative vs. anti-conservative tests As a method to answer - is this ok? Ratios - as an example

Complex designs 2 way anova, multiple regression non Nesting, time series, levels, multivariate Different ways to parallelize Base R furrr Scripts that accept parameters, calling R scripts from bash scripts

Problem set - parallel stuff, seeds, lewis

Sampling from data sets: decision errors and predicting new data

Generality - cross validation False positive rate - empirical Power simulations Multiple testing and how many permutations, empirical fdr

Simulations 1

simulations to verify methods they have learned simulating a biological null

eddible package: https://emitanaka.org/edibble-book/index.html Distributions simulation: https://mgoodman.shinyapps.io/distributions-lab/ Simstudy:https://cran.r-project.org/web/packages/simstudy/ Faux https://debruine.github.io/faux/ Problem set

Simulations 2 When the null needs to be simulated Simulating data Power Problem set: assigned at start - use your own data to use one method from class - consult ahead of time with instructors

Problem set: assigned at start - use your own data to use one method from class - consult ahead of time with instructors