library(hmcdm)
= dim(Design_array)[1]
N = nrow(Q_matrix)
J = ncol(Q_matrix)
K = dim(Design_array)[3] L
<- sample(1:2^K, N, replace = L)
class_0 <- matrix(0,N,K)
Alphas_0 for(i in 1:N){
<- inv_bijectionvector(K,(class_0[i]-1))
Alphas_0[i,]
}= rnorm(N)
thetas_true = c(-1, 1.8, .277, .055)
lambdas_true <- sim_alphas(model="HO_sep",
Alphas lambdas=lambdas_true,
thetas=thetas_true,
Q_matrix=Q_matrix,
Design_array=Design_array)
table(rowSums(Alphas[,,5]) - rowSums(Alphas[,,1])) # used to see how much transition has taken place
#>
#> 0 1 2 3 4
#> 36 28 91 161 34
<- matrix(runif(J*2,.1,.2), ncol=2)
itempars_true
<- sim_hmcdm(model="DINA",Alphas,Q_matrix,Design_array,
Y_sim itempars=itempars_true)
= hmcdm(Y_sim,Q_matrix,"DINA_HO",Test_order = Test_order, Test_versions = Test_versions,
output_HMDCM chain_length=100,burn_in=30,
theta_propose = 2,deltas_propose = c(.45,.35,.25,.06))
#> 0
= hmcdm(Y_sim,Q_matrix,"DINA_HO",Design_array,
output_HMDCM chain_length=100,burn_in=30,
theta_propose = 2,deltas_propose = c(.45,.35,.25,.06))
#> 0
output_HMDCM#>
#> Model: DINA_HO
#>
#> Sample Size: 350
#> Number of Items:
#> Number of Time Points:
#>
#> Chain Length: 100, burn-in: 30
summary(output_HMDCM)
#>
#> Model: DINA_HO
#>
#> Item Parameters:
#> ss_EAP gs_EAP
#> 0.1698 0.1651
#> 0.1680 0.1600
#> 0.1493 0.1269
#> 0.1329 0.1430
#> 0.1139 0.1124
#> ... 45 more items
#>
#> Transition Parameters:
#> lambdas_EAP
#> λ0 -1.4473
#> λ1 1.9931
#> λ2 0.1998
#> λ3 0.1391
#>
#> Class Probabilities:
#> pis_EAP
#> 0000 0.1429
#> 0001 0.1706
#> 0010 0.1868
#> 0011 0.2530
#> 0100 0.1556
#> ... 11 more classes
#>
#> Deviance Information Criterion (DIC): 19275.41
#>
#> Posterior Predictive P-value (PPP):
#> M1: 0.4937
#> M2: 0.49
#> total scores: 0.623
<- summary(output_HMDCM)
a $ss_EAP
a#> [,1]
#> [1,] 0.1697906
#> [2,] 0.1680309
#> [3,] 0.1492832
#> [4,] 0.1329068
#> [5,] 0.1138727
#> [6,] 0.1661326
#> [7,] 0.1705610
#> [8,] 0.1246988
#> [9,] 0.2354335
#> [10,] 0.1794307
#> [11,] 0.1431058
#> [12,] 0.1529385
#> [13,] 0.1331102
#> [14,] 0.1267657
#> [15,] 0.1497950
#> [16,] 0.1769501
#> [17,] 0.1744780
#> [18,] 0.1824385
#> [19,] 0.2135706
#> [20,] 0.1727740
#> [21,] 0.1837538
#> [22,] 0.1474389
#> [23,] 0.2108857
#> [24,] 0.1130164
#> [25,] 0.1082945
#> [26,] 0.1829056
#> [27,] 0.2060486
#> [28,] 0.1929764
#> [29,] 0.1876752
#> [30,] 0.1507918
#> [31,] 0.1301102
#> [32,] 0.1945162
#> [33,] 0.2057871
#> [34,] 0.1923413
#> [35,] 0.1234793
#> [36,] 0.2201181
#> [37,] 0.2065649
#> [38,] 0.1702183
#> [39,] 0.1853408
#> [40,] 0.1845450
#> [41,] 0.1679590
#> [42,] 0.1603618
#> [43,] 0.1931209
#> [44,] 0.1568479
#> [45,] 0.2033812
#> [46,] 0.1286860
#> [47,] 0.1235941
#> [48,] 0.2308971
#> [49,] 0.1723194
#> [50,] 0.1458139
$lambdas_EAP
a#> [,1]
#> λ0 -1.4473139
#> λ1 1.9930927
#> λ2 0.1997842
#> λ3 0.1390890
mean(a$PPP_total_scores)
#> [1] 0.6246612
mean(upper.tri(a$PPP_item_ORs))
#> [1] 0.49
mean(a$PPP_item_means)
#> [1] 0.506
<- numeric(L)
AAR_vec for(t in 1:L){
<- mean(Alphas[,,t]==a$Alphas_est[,,t])
AAR_vec[t]
}
AAR_vec#> [1] 0.9171429 0.9385714 0.9592857 0.9742857 0.9785714
<- numeric(L)
PAR_vec for(t in 1:L){
<- mean(rowSums((Alphas[,,t]-a$Alphas_est[,,t])^2)==0)
PAR_vec[t]
}
PAR_vec#> [1] 0.7200000 0.7800000 0.8514286 0.9142857 0.9200000
$DIC
a#> Transition Response_Time Response Joint Total
#> D_bar 2082.864 NA 15064.31 1289.583 18436.75
#> D(theta_bar) 1811.729 NA 14528.49 1257.873 17598.09
#> DIC 2353.999 NA 15600.12 1321.292 19275.41
head(a$PPP_total_scores)
#> [,1] [,2] [,3] [,4] [,5]
#> [1,] 0.2714286 0.8714286 1.0000000 0.5000000 1.0000000
#> [2,] 0.5428571 0.4000000 1.0000000 1.0000000 0.5285714
#> [3,] 0.2714286 0.5857143 0.4571429 0.5714286 0.8714286
#> [4,] 0.3285714 0.5285714 0.5571429 0.5857143 1.0000000
#> [5,] 0.5428571 0.5285714 0.9285714 0.4571429 0.7714286
#> [6,] 0.7571429 0.8142857 1.0000000 1.0000000 0.2428571
head(a$PPP_item_means)
#> [1] 0.5000000 0.4571429 0.5857143 0.5285714 0.5000000 0.4285714
head(a$PPP_item_ORs)
#> [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8]
#> [1,] NA 0.7714286 0.7285714 0.5285714 0.5571429 0.7142857 0.6000000 0.6714286
#> [2,] NA NA 0.8714286 0.6571429 0.8571429 0.5428571 0.8857143 0.9571429
#> [3,] NA NA NA 0.9714286 0.4428571 0.8142857 0.3142857 0.9857143
#> [4,] NA NA NA NA 0.8285714 0.7714286 0.9142857 0.6428571
#> [5,] NA NA NA NA NA 0.7571429 0.8285714 0.5857143
#> [6,] NA NA NA NA NA NA 0.2857143 0.6285714
#> [,9] [,10] [,11] [,12] [,13] [,14] [,15]
#> [1,] 0.9142857 0.6142857 0.71428571 0.9428571 0.9857143 0.6857143 0.35714286
#> [2,] 0.5428571 0.2714286 0.05714286 0.7571429 0.8714286 0.4571429 0.54285714
#> [3,] 0.2285714 0.3857143 0.27142857 0.1714286 0.7857143 0.1857143 0.35714286
#> [4,] 0.6714286 0.6428571 0.91428571 0.7571429 0.5571429 0.5857143 0.48571429
#> [5,] 0.2571429 0.4000000 0.11428571 0.4714286 0.8142857 0.2857143 0.04285714
#> [6,] 0.8714286 0.4714286 0.42857143 0.1571429 0.4142857 0.2714286 0.17142857
#> [,16] [,17] [,18] [,19] [,20] [,21] [,22]
#> [1,] 0.9714286 0.37142857 0.2285714 0.41428571 0.7000000 0.70000000 0.57142857
#> [2,] 0.6142857 0.42857143 0.6857143 0.15714286 0.4571429 0.72857143 0.11428571
#> [3,] 0.1857143 0.58571429 0.9714286 0.02857143 0.4285714 0.05714286 0.10000000
#> [4,] 0.7857143 0.47142857 0.9285714 0.35714286 0.5714286 0.31428571 0.32857143
#> [5,] 0.3285714 0.04285714 0.0000000 0.12857143 0.2142857 0.50000000 0.55714286
#> [6,] 0.8285714 0.48571429 0.3571429 0.38571429 0.8428571 0.58571429 0.07142857
#> [,23] [,24] [,25] [,26] [,27] [,28] [,29]
#> [1,] 0.1000000 0.8714286 0.7571429 0.17142857 0.1285714 0.6142857 0.6714286
#> [2,] 0.6000000 0.6857143 0.5142857 0.48571429 0.1000000 0.6571429 0.6571429
#> [3,] 0.4571429 0.8000000 0.7142857 0.04285714 0.5000000 0.4714286 0.2571429
#> [4,] 0.3285714 0.9571429 0.9571429 0.30000000 0.9142857 0.6571429 0.1000000
#> [5,] 0.2000000 1.0000000 0.6000000 0.64285714 0.4000000 0.6285714 0.7714286
#> [6,] 0.2571429 0.9000000 0.4428571 0.65714286 0.1285714 0.3000000 0.4142857
#> [,30] [,31] [,32] [,33] [,34] [,35] [,36]
#> [1,] 0.6428571 0.3285714 0.6714286 0.8714286 0.6428571 0.81428571 0.80000000
#> [2,] 0.6428571 0.8714286 0.2857143 0.6142857 0.4428571 0.01428571 0.05714286
#> [3,] 0.2571429 0.9142857 0.3000000 0.4714286 0.8857143 0.91428571 0.91428571
#> [4,] 0.3285714 0.4571429 0.2571429 0.8571429 0.2571429 0.27142857 0.21428571
#> [5,] 1.0000000 0.4142857 0.3857143 0.4571429 0.7285714 0.42857143 0.42857143
#> [6,] 0.1000000 0.4571429 0.2285714 0.6714286 0.5571429 0.12857143 0.20000000
#> [,37] [,38] [,39] [,40] [,41] [,42] [,43]
#> [1,] 0.8857143 0.9428571 0.9857143 0.9285714 0.5571429 0.1714286 0.11428571
#> [2,] 0.6000000 0.3428571 0.1285714 0.7000000 0.1714286 0.7285714 0.38571429
#> [3,] 0.8285714 0.6285714 0.8142857 0.4428571 0.7571429 0.9142857 0.22857143
#> [4,] 0.7142857 0.2714286 0.9714286 0.1428571 0.8857143 0.9428571 0.07142857
#> [5,] 0.5428571 0.3857143 0.9714286 0.5857143 0.6000000 0.5857143 0.18571429
#> [6,] 0.8142857 0.3571429 0.7857143 0.9571429 0.1857143 0.9714286 0.10000000
#> [,44] [,45] [,46] [,47] [,48] [,49] [,50]
#> [1,] 0.01428571 0.00000000 0.2714286 0.04285714 0.44285714 0.1571429 0.1142857
#> [2,] 0.37142857 0.02857143 0.6000000 0.40000000 0.35714286 0.3285714 0.2857143
#> [3,] 0.82857143 0.51428571 0.9857143 0.91428571 0.52857143 0.2571429 0.9428571
#> [4,] 0.20000000 0.67142857 0.5714286 0.54285714 0.70000000 0.6285714 0.1000000
#> [5,] 0.32857143 0.01428571 0.9000000 0.65714286 0.04285714 0.6714286 0.2571429
#> [6,] 0.08571429 0.18571429 0.4571429 0.57142857 0.12857143 0.2571429 0.3000000
library(bayesplot)
pp_check(output_HMDCM)
pp_check(output_HMDCM, plotfun="dens_overlay", type="item_mean")
pp_check(output_HMDCM, plotfun="hist", type="item_OR")
#> `stat_bin()` using `bins = 30`. Pick better value with `binwidth`.
pp_check(output_HMDCM, plotfun="stat_2d", type="item_mean")
pp_check(output_HMDCM, plotfun="scatter_avg", type="total_score")
pp_check(output_HMDCM, plotfun="error_scatter_avg", type="total_score")
Checking convergence of the two independent MCMC chains with
different initial values using coda
package.
# output_HMDCM1 = hmcdm(Y_sim, Q_matrix, "DINA_HO", Design_array,
# chain_length=100, burn_in=30,
# theta_propose = 2, deltas_propose = c(.45,.35,.25,.06))
# output_HMDCM2 = hmcdm(Y_sim, Q_matrix, "DINA_HO", Design_array,
# chain_length=100, burn_in=30,
# theta_propose = 2, deltas_propose = c(.45,.35,.25,.06))
#
# library(coda)
#
# x <- mcmc.list(mcmc(t(rbind(output_HMDCM1$ss, output_HMDCM1$gs, output_HMDCM1$lambdas))),
# mcmc(t(rbind(output_HMDCM2$ss, output_HMDCM2$gs, output_HMDCM2$lambdas))))
#
# gelman.diag(x, autoburnin=F)