Skip to content

Instantly share code, notes, and snippets.

@fusaroli

fusaroli/trial.r

Last active December 20, 2019 10:24
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save fusaroli/64d4b4d85221da026cda274647acfb8d to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save fusaroli/64d4b4d85221da026cda274647acfb8d to your computer and use it in GitHub Desktop.
Download ZIP

Revisions

  1. fusaroli revised this gist Dec 20, 2019. 1 changed file with 23 additions and 1 deletion.
    24 changes: 23 additions & 1 deletion trial.r
    View file
    Original file line number Diff line number Diff line change
    @@ -49,4 +49,26 @@ hypothesis(m,"languagedk:moSymptom > languagedk:moSymptom:data$AgeS", class = "b

    # ERROR
    #Error: Some parameters cannot be found in the model:
    # 'bsp_languagedk:moSymptom:data', 'bsp_AgeS'
    # 'bsp_languagedk:moSymptom:data', 'bsp_AgeS'

    ## Without the mo() all runs fine

    # Define model

    SymptomVoice_Age_f2 <- bf(Outcome ~ 0 + language + language : Symptom + language : Symptom : AgeS + (1 | gr(ID, by = language)))

    # Define priors

    SkepticPrior1 <- c(
    prior(normal(0, 1), class=b),
    prior(normal(0, .1), class=sd)
    )

    # Run model

    m2 <- brm(SymptomVoice_Age_f2,
    data = data,
    family = gaussian,
    prior = SkepticPrior1,
    sample_prior = T
    )
  2. fusaroli created this gist Dec 20, 2019.
    52 changes: 52 additions & 0 deletions trial.r
    View file
    Original file line number Diff line number Diff line change
    @@ -0,0 +1,52 @@
    # Load libraries
    library(brms)
    library(tidyverse)

    # Load data

    data <- read_csv("temp_data.csv")

    # Define model

    SymptomVoice_Age_f <- bf(Outcome ~ 0 + language + language : mo(Symptom) + language : mo(Symptom) : AgeS + (1 | gr(ID, by = language)))

    # Define priors

    SkepticPrior1 <- c(
    prior(normal(0, 1), class=b),
    prior(normal(0, .1), class=b, coef = languagedk:moSymptom),
    prior(normal(0, .1), class=b, coef = languageus:moSymptom),
    prior(normal(0, .1), class=sd)
    )

    # Run model

    m <- brm(SymptomVoice_Age_f,
    data = data,
    family = gaussian,
    prior = SkepticPrior1,
    sample_prior = T
    )

    # Error:
    # Error in eval(predvars, data, env) : object 'AgeS' not found

    ## Pseudo-Fixing it

    SymptomVoice_Age_f <- bf(Outcome ~ 0 + language + language : mo(Symptom) + language : mo(Symptom) : data$AgeS + (1 | gr(ID, by = language)))


    m <- brm(SymptomVoice_Age_f,
    data = data,
    family = gaussian,
    prior = SkepticPrior1,
    sample_prior = T
    )

    # Hypothesis testing

    hypothesis(m,"languagedk:moSymptom > languagedk:moSymptom:data$AgeS", class = "bsp")

    # ERROR
    #Error: Some parameters cannot be found in the model:
    # 'bsp_languagedk:moSymptom:data', 'bsp_AgeS'