R: NIMBLEをためす

ちょっとNIMBLEをためしてみました。

Rコードです。統計研修の例題でためしてみました。

結果です。

> ex2.out$summary
$chain1
             Mean     Median    St.Dev.  95%CI_low  95%CI_upp
beta   -2.1403865 -2.1251510 0.50126977 -3.1455574 -1.1732269
beta.x  0.5518773  0.5444307 0.09992961  0.3698632  0.7614172

$chain2
             Mean     Median   St.Dev.  95%CI_low 95%CI_upp
beta   -2.1478326 -2.1250499 0.5490874 -3.3233251 -1.184588
beta.x  0.5550663  0.5536948 0.1119390  0.3526774  0.800066

$chain3
             Mean     Median   St.Dev. 95%CI_low  95%CI_upp
beta   -2.1648437 -2.1021086 0.5091905 -3.232323 -1.3204283
beta.x  0.5590768  0.5489405 0.1037532  0.376029  0.7790607

$all.chains
             Mean     Median   St.Dev.  95%CI_low  95%CI_upp
beta   -2.1510209 -2.1174662 0.5201975 -3.2645064 -1.2197713
beta.x  0.5553401  0.5487436 0.1053325  0.3665114  0.7760678

> summary(ex2.out$samples)

Iterations = 1:1000
Thinning interval = 1 
Number of chains = 3 
Sample size per chain = 1000 

1. Empirical mean and standard deviation for each variable,
   plus standard error of the mean:

          Mean     SD Naive SE Time-series SE
beta   -2.1510 0.5202 0.009497       0.024311
beta.x  0.5553 0.1053 0.001923       0.004891

2. Quantiles for each variable:

          2.5%     25%     50%     75%   97.5%
beta   -3.2645 -2.4805 -2.1175 -1.7906 -1.2198
beta.x  0.3665  0.4801  0.5487  0.6224  0.7761

> gelman.diag(ex2.out$samples)
Potential scale reduction factors:

       Point est. Upper C.I.
beta         1.01       1.01
beta.x       1.01       1.01

Multivariate psrf

1
> 
> plot(ex2.out$samples)

タグ：R Nimble