User Specified Priors

rstanarm के साथ Bayesian Regression Modeling

Jake Thompson

Psychometrician, ATLAS, University of Kansas

डिफॉल्ट prior क्यों बदलें?

मजबूत कारण कि पैरामीटर एक निश्चित मान लेगा
पैरामीटर पर बाधाएँ

Prior सेट करें

stan_model <- stan_glm(kid_score ~ mom_iq, data = kidiq,
  prior_intercept = normal(location = 0, scale = 10),
  prior = normal(location = 0, scale = 2.5),
  prior_aux = exponential(rate = 1)
)

Prior सेट करें

stan_model <- stan_glm(kid_score ~ mom_iq, data = kidiq,
  prior_intercept = normal(location = 0, scale = 10, autoscale = FALSE),
  prior = normal(location = 0, scale = 2.5, autoscale = FALSE),
  prior_aux = exponential(rate = 1, autoscale = FALSE)
)

Prior सेट करें

stan_model <- stan_glm(kid_score ~ mom_iq, data = kidiq,
  prior_intercept = normal(location = 3, scale = 2),
  prior = cauchy(location = 0, scale = 1))

कई तरह के priors
- normal()
- exponential()
- student_t()
- cauchy()
?priors

Flat priors

stan_model <- stan_glm(kid_score ~ mom_iq, data = kidiq,
  prior_intercept = NULL,
  prior = NULL,
  prior_aux = NULL)
prior_summary(stan_model)

Priors for model 'stan_model' 
 ------
Intercept (after predictors centered)
 ~ flat

Coefficients
 ~ flat

Auxiliary (sigma)
 ~ flat
 ------
See help('prior_summary.stanreg') for more details

अभ्यास करते हैं!

rstanarm के साथ Bayesian Regression Modeling