PowerPriorVari/modifiedMDICprogram.R at master · lxt3/PowerPriorVari · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
# One-arm trial (OPC) example

# reset vectos
res1<-res2<-bias1<-bias2<-alpha<-sd.1<-sd.2<-phat<-sigmaD<-sigmaD1<-vector(length=nsim)


# Generate nsim current data sets and compute posterior inference using prior D0
for(i in 1:nsim){

  if((i/1000)%in%c(1:20)) print(i)

  # generate ith D from this mu
  D<-rnorm(mean=mu, n=n, sd=1)

  # set D1 = x% of D
  D1<- D[1:(percent*n)]


  # estimate alpha0 using D0 and D1

  if(external){
    D1<-rnorm(mu, n=percent*n, sd=1)
    fit<-mu_posterior(mu      = mean(D1),
                      sigma2   = var(D1),
                      N       = length(D1),
                      mu0     = mean(D0),
                      sigma02  = var(D0),
                      N0      = length(D0),
                      N0_max=length(D0)*max_alpha,
                      number_mcmc=nmcmc,
                      D0=D0, D=D1
    )
  }
  else{
  fit<-mu_posterior(mu      = mean(D1),
               sigma2   = var(D1),
               N       = length(D1),
               mu0     = mean(D0),
               sigma02  = var(D0),
               N0      = length(D0),
               N0_max=length(D0)*max_alpha,
               number_mcmc=nmcmc,
               D0=D0, D=D1
  )
  }

 alpha_use<- alpha[i]<-ifelse(fit$alpha_loss>max_alpha, max_alpha, fit$alpha_loss)


  # use alpha0 from above to estimate posterior

  # first no drop
  D2<-D

  fit<-mu_posterior2(mu      = mean(D2),
                    sigma2   = var(D2),
                    N       = length(D2),
                    mu0     = mean(D0),
                    sigma02  = var(D0),
                    N0      = length(D0),
                    N0_max=length(D0)*max_alpha,
                    number_mcmc=nmcmc,
                    alpha_loss=alpha_use
  )

  res1[i]<-mean(fit$mu_posterior > null)>=prob.H1  # prob H1 (or prob reject H0)

  sd.1[i]<-sd(fit$mu_posterior)
  sigmaD[i]<-mean(fit$sigma2_posterior)

  bias1[i]<-mean(fit$mu_posterior)-mu


  # drop D1 (if percent<1)
  if(percent<1) {

    D2 <- D[-(1:(percent*n))]

    # use alpha0 from above to estimate posterior
    fit<-mu_posterior2(mu      = mean(D2),
                     sigma2   = var(D2),
                     N       = length(D2),
                     mu0     = mean(D0),
                     sigma02  = var(D0),
                     N0      = length(D0),
                     N0_max=length(D0)*max_alpha,
                     number_mcmc=nmcmc,
                     alpha_loss=alpha_use
    )

    res2[i]<-mean(fit$mu_posterior > null)>=prob.H1
    sigmaD1[i]<-mean(fit$sigma2_posterior)


    sd.2[i]<-sd(fit$mu_posterior)

    bias2[i]<-mean(fit$mu_posterior)-mu
  } else {

    res2[i]<-res1[i]
    sigmaD1[i]<-sigmaD[i]
    sd.2[i]<-sd.1[i]
    bias2[i]<-bias1[i]

  }

}

# print to sink output
sink(sinkfname,split=T, append=T) # put this in modifiedMDICprogram and open/close for output only

cat("percent = \n")
print(percent)
cat("mu = \n")
print(mu)
cat("n = \n")
print(n)
cat("null = \n")
print(null)
cat("external = \n")
print(external)

cat("No dropping\n")

print(mean(res1))
print(mean(sd.1))

cat("bias: No dropping\n")
print(mean(bias1))


cat("alpha\n")
print(mean(alpha))
#print(mean(sd.1))

cat("With dropping\n")

print(mean(res2))
print(mean(sd.2))

cat("bias: Dropping\n")
print(mean(bias2))

sink()