Displaying 4 results from an estimated 4 matches for "betadraw".
2005 May 19
1
R 2.1.0 RH Linux Built from Source Segmentation Fault
...+ # for each physician i draw a beta[i] using MH algorithm
+ #
+ Data=list(m=nalt,X=X[rowx1[i]:rowx2[i],],y=Y[rowy1[i]:rowy2[i],])
+ bbar=t(theta)%*%z[i,]
+ prior=list(A=delta,betabar=bbar)
+ a=rmnlRwMetrop1(Data,prior,Mcmc=parameters,beta[i,])
+ beta[i,]=a$betadraw
+ accept[i]=a$acceptr
+ }
+ mregout=rmultireg(beta,z,thetabar,A1,nu1,V)
+ delta=mregout$Sigma
+ theta=mregout$B
+ if (j%%keep==0){
+ thetas[j/keep,]=as.vector(theta)
+ betastore[j/keep]=list(beta)
+ accepts[j/keep]=mean(accept)
+ ftime=proc.time()[3]
+ cat('Time taken by iteration:...
2009 Mar 21
0
Can not replicate estimates with rScreen function from ROSSI "Bayesian Statistics and Marketing"
...ot;A.3 HIERARCHICAL BAYES MODELING ? AN EXAMPLE" pages 303- 322
Eugen.
> source("C:\\Documents and Settings\\Administrator\\My Documents\\rScreen.R")
> source("C:\\Documents and Settings\\Administrator\\My Documents\\run screening rules.R")
> t(matrix(apply(out$betadraw[,,500:1000],2,mean),ncol=18))
[,1]
[1,] -1.832448426
[2,] -0.355615613
[3,] -1.210309206
[4,] -0.052342073
[5,] 0.182849361
[6,] 0.350787951
[7,] 0.598865166
[8,] -0.186233704
[9,] 0.652014888
[10,] -0.625699004
[11,] 0.469491733
[12,] 0.921956623
[13,] 1.304895684
[14...
2007 Jun 20
0
Multi-variate Probit model using Bayesm
...is replicated
in four columns (diagonal values) to correspond to the first four choices of
y. Every 5th row is 0. This pattern is repeated for all n respondents. Does
this mean for the FIFTH customer choice - there is NO calculation done??
5) Once I run rmvpgibbs with this data structure, my "betadraw" output
summary is of 36 variables (showing the first 11):
mean std dev num se rel eff s size 1 -2.999 3.32 0.216 77 234 2 -20.62 5.3
0.514 170 106 3 -17.973 2.87 0.276 166 108 4 -14.595 5.1 0.505 177 102 5
2.551 2.13 0.165 109 165 6 -0.235 1.11 0.071 74 240 7 -0.772 0.78 0.061
110 162...
2007 Jun 21
0
Multi-variate Probit model using Bayesm in R
...is replicated
in four columns (diagonal values) to correspond to the first four choices of
y. Every 5th row is 0. This pattern is repeated for all n respondents. Does
this mean for the FIFTH customer choice - there is NO calculation done??
5) Once I run rmvpgibbs with this data structure, my "betadraw" output
summary is of 36 variables (showing the first 11):
mean std dev num se rel eff s size 1 -2.999 3.32 0.216 77 234 2 -20.62 5.3
0.514 170 106 3 -17.973 2.87 0.276 166 108 4 -14.595 5.1 0.505 177 102 5
2.551 2.13 0.165 109 165 6 -0.235 1.11 0.071 74 240 7 -0.772 0.78 0.061
110 162...