model_exponential_gamma: Run bayesian exponential-gamma model for estimating non-zero...
In eliobartos/bayeselio: Conjugate Bayesians Models

Description Usage Arguments Details Value Author(s) Examples

View source: R/model_exponential_gamma.R

Runs a bayesian estimation of non-zero positive variable using exponential distribution as likelihood and gamma distribution as conjugate prior. Posterior distribution is gamma distribution. Prior used is Gamma(shape, rate)

1	model_exponential_gamma(shape, rate, n_sample, sum_sample, n_post = 1e+05)

`shape`	Parameter for prior distribution representing the number of samples from before - 1
`rate`	Parameter for prior distribution representing the sum of samples from before
`n_sample`	Total number of cases in your data
`sum_sample`	Sum of variable for each successful case.
`n_post`	Size of sample from posterior distribution

Mean of prior distribution is shape/rate. Remember Exp(lambda) has mean of 1/lambda.

Vector of samples from posterior distribution representing mean of exponential distribution (1/lambda).

Posterior distribution is Gamma(shape + n_sample, rate + sum_sample)

Elio Bartoš

post = model_exponential_gamma(0, 0, 20, 100) # No prior information, pror is uniform
post2 = model_exponential_gamma(5, 95, 3, 50) # Prior succes rate is around 5% with estimation strenght as it was estimated on a sample of 100

mean(post)
quantile(post, probs = c(0.05, 0.95)) # 90% highest density posterior interval

mean(post2)
quantile(post2, probs = c(0.05, 0.95))