**statistics:bayesian**45

Inferential Statistics is not Inferential – sci five | University of Basel – Medium

3 days ago by hallucigenia

"The earth is flat (p > 0.05).

I confess. Throughout my scientific life, I have used a method that I knew or felt was deeply flawed. What’s more, I admit to have taught — and I do still teach — this method to my students. I have a number of questionable excuses for that. For example, because the method has shaped a big part of science in the last century, I think students ought to know about it.

But I have increasingly come to believe that science was and is largely a story of success in spite of, and not because of, the use of this method. The method is called inferential statistics. Or more precisely, hypothesis testing.

The method I consider flawed and deleterious involves taking sample data, then applying some mathematical procedure, and taking the result of that procedure as showing whether or not a hypothesis about a larger population is correct."

statistics:bayesian
statistics:frequentist
confidence_intervals
to_teach
regression
statistics
philosophy_of_science
philosophy_of_statistics
statistics:error_based
I confess. Throughout my scientific life, I have used a method that I knew or felt was deeply flawed. What’s more, I admit to have taught — and I do still teach — this method to my students. I have a number of questionable excuses for that. For example, because the method has shaped a big part of science in the last century, I think students ought to know about it.

But I have increasingly come to believe that science was and is largely a story of success in spite of, and not because of, the use of this method. The method is called inferential statistics. Or more precisely, hypothesis testing.

The method I consider flawed and deleterious involves taking sample data, then applying some mathematical procedure, and taking the result of that procedure as showing whether or not a hypothesis about a larger population is correct."

3 days ago by hallucigenia

Bayesian Basics

march 2018 by mozzarella

nice bookdown, lets see how it reads

statistics:bayesian
nice-thinking
An introduction to Bayesian data analysis.

march 2018 by mozzarella

mjskay/tidybayes: Bayesian analysis + tidy data + geoms (R package)

statistics:bayesian
tidyverse
R
r_packages
visualization
to_try

february 2018 by hallucigenia

tidybayes is an R package that aims to make it easy to integrate popular Bayesian modelling methods into a tidy data + ggplot workflow.

Tidy data frames (one observation per row) are particularly convenient for use in a variety of R data manipulation and visualization packages. However, when using MCMC / Bayesian samplers like JAGS or Stan in R, we often have to translate this data into a form the sampler understands, and then after running the model, translate the resulting sample ...

february 2018 by hallucigenia

How often does the best team win?A unified approach to understanding randomness in North American sport

nice-thinking
statistics
statistics:bayesian
basketball-reference

january 2018 by mozzarella

In this manuscript, we develop Bayesian state-space models using betting

market data that can be uniformly applied across sporting organizations

to better understand the role of randomness in game outcomes.

These models can be used to extract estimates of team strength,

the between-season, within-season, and game-to-game variability of

team strengths, as well each team’s home advantage. We implement

our approach across a decade of play in each of the National Football

League (NFL), National Hockey League (NHL), National Basketball

Association (NBA), and Major League Baseball (MLB), finding that

the NBA demonstrates both the largest dispersion in talent and the

largest home advantage, while the NHL and MLB stand out for their

relative randomness in game outcomes.

january 2018 by mozzarella

VAST: Spatio-temporal analysis of univariate or multivariate data, e.g., standardizing data for multiple species or stage

september 2017 by hallucigenia

VAST

Is an R package for implementing a spatial delta-generalized linear mixed model (delta-GLMM) for multiple categories (species, size, or age classes) when standardizing survey or fishery-dependent data.

Builds upon a previous R package SpatialDeltaGLMM (public available here), and has unit-testing to automatically confirm that VAST and SpatialDeltaGLMM give identical results (to the 3rd decimal place for parameter estimates) for several varied real-world case-study examples

Has built in diagnostic functions and model-comparison tools

Is intended to improve analysis speed, replicability, peer-review, and interpretation of index standardization methods

Background

This tool is designed to estimate spatial variation in density using spatially referenced data, with the goal of habitat associations (correlations among species and with habitat) and estimating total abundance for a target species in one or more years.

The model builds upon spatio-temporal delta-generalized linear mixed modelling techniques (Thorson Shelton Ward Skaug 2015 ICESJMS), which separately models the proportion of tows that catch at least one individual ("encounter probability") and catch rates for tows with at least one individual ("positive catch rates").

Submodels for encounter probability and positive catch rates by default incorporate variation in density among years (as a fixed effect), and can incorporate variation among sampling vessels (as a random effect, Thorson and Ward 2014) which may be correlated among categories (Thorson Fonner Haltuch Ono Winker In press).

Spatial and spatiotemporal variation are approximated as Gaussian Markov random fields (Thorson Skaug Kristensen Shelton Ward Harms Banante 2014 Ecology), which imply that correlations in spatial variation decay as a function of distance.

statistics:gams
statistics:time_series
statistics:fisheries
fisheries
fisheries:methods
statistics:bayesian
statistics:spatial
R_packages
Is an R package for implementing a spatial delta-generalized linear mixed model (delta-GLMM) for multiple categories (species, size, or age classes) when standardizing survey or fishery-dependent data.

Builds upon a previous R package SpatialDeltaGLMM (public available here), and has unit-testing to automatically confirm that VAST and SpatialDeltaGLMM give identical results (to the 3rd decimal place for parameter estimates) for several varied real-world case-study examples

Has built in diagnostic functions and model-comparison tools

Is intended to improve analysis speed, replicability, peer-review, and interpretation of index standardization methods

Background

This tool is designed to estimate spatial variation in density using spatially referenced data, with the goal of habitat associations (correlations among species and with habitat) and estimating total abundance for a target species in one or more years.

The model builds upon spatio-temporal delta-generalized linear mixed modelling techniques (Thorson Shelton Ward Skaug 2015 ICESJMS), which separately models the proportion of tows that catch at least one individual ("encounter probability") and catch rates for tows with at least one individual ("positive catch rates").

Submodels for encounter probability and positive catch rates by default incorporate variation in density among years (as a fixed effect), and can incorporate variation among sampling vessels (as a random effect, Thorson and Ward 2014) which may be correlated among categories (Thorson Fonner Haltuch Ono Winker In press).

Spatial and spatiotemporal variation are approximated as Gaussian Markov random fields (Thorson Skaug Kristensen Shelton Ward Harms Banante 2014 Ecology), which imply that correlations in spatial variation decay as a function of distance.

september 2017 by hallucigenia

bayes: a kinda-sorta masterpost

august 2017 by hallucigenia

"I have written many many words about “Bayesianism” in this space over the years, but the closest thing to a comprehensive “my position on Bayes” post to date is this one from three years ago, which I wrote when I was much newer to this stuff. People sometimes link that post or ask me about it, which almost never happens with my other Bayes posts. So I figure I should write a more up-to-date “position post.”

I will try to make this at least kind of comprehensive, but I will omit many details and sometimes state conclusions without the corresponding arguments. Feel free to ask me if you want to hear more about something."

statistics:bayesian
philosophy_of_science
philosophy_of_statistics
I will try to make this at least kind of comprehensive, but I will omit many details and sometimes state conclusions without the corresponding arguments. Feel free to ask me if you want to hear more about something."

august 2017 by hallucigenia

xcelab.net

june 2017 by sechilds

Statistical Rethinking is an introduction to applied Bayesian data analysis, aimed at PhD students and researchers in the natural and social sciences. This audience has had some calculus and linear algebra, and one or two joyless undergraduate courses in statistics. I've been teaching applied statistics to this audience for about a decade now, and this book has evolved from that experience.

The book teaches generalized linear multilevel modeling (GLMMs) from a Bayesian perspective, relying on a simple logical interpretation of Bayesian probability and maximum entropy. The book covers the basics of regression through multilevel models, as well as touching on measurement error, missing data, and Gaussian process models for spatial and network autocorrelation.

This is not a traditional mathematical statistics book. Instead the approach is computational, using complete R code examples, aimed at developing skilled and skeptical scientists. Theory is explained through simulation exercises, using R code. And modeling examples are fully worked, with R code displayed within the main text. Mathematical depth is given in optional "overthinking" boxes throughout.

statistics
statistics:bayesian
book
R
The book teaches generalized linear multilevel modeling (GLMMs) from a Bayesian perspective, relying on a simple logical interpretation of Bayesian probability and maximum entropy. The book covers the basics of regression through multilevel models, as well as touching on measurement error, missing data, and Gaussian process models for spatial and network autocorrelation.

This is not a traditional mathematical statistics book. Instead the approach is computational, using complete R code examples, aimed at developing skilled and skeptical scientists. Theory is explained through simulation exercises, using R code. And modeling examples are fully worked, with R code displayed within the main text. Mathematical depth is given in optional "overthinking" boxes throughout.

june 2017 by sechilds

How Bayesian inference works

may 2017 by mozzarella

distinction between probabilities: conditional, joint, marginal

statistics
statistics:bayesian
may 2017 by mozzarella

**related tags**

Copy this bookmark: