source("c:\\documents and settings\\garreel.oncology_nt.000\\desktop\\cancerclinicaltrials\\LikelihoodSurvival\\Rfunctions.R")

library(survival)



###############################################################################
# h1 = 0.02, h2 = 0.02
M<-50 # number of datasets
time<-24 # analysis time
atime<-24 # accrual time
n<-30  # number of indivs in each of two subpops
h1<-0.02 # hazard rate for subpop 1
h2<-0.02  # hazard rate for subpop 2
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation

# generate M datasets

data<-mixture.exp(n,h1,n,h2,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=time)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue(100,h1,h2,time, study.time)
plot.both(lik,freq.mat,true)

################################################################################
# h1 = 0.02, h2 = 0.01
M<-50 # number of datasets
time<-24 # analysis time
atime<-24 # accrual time
n<-30  # number of indivs in each of two subpops
h1<-0.01 # hazard rate for subpop 1
h2<-0.02  # hazard rate for subpop 2
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation

# generate M datasets

data<-mixture.exp(n,h1,n,h2,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=time)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue(100,h1,h2,time, study.time)
plot.both(lik,freq.mat,true)

################################################################################
# h1 = 0.03, h2 = 0.005
M<-50 # number of datasets
time<-24 # analysis time
n<-30  # number of indivs in each of two subpops
atime<-24 # accrual time
h1<-0.03 # hazard rate for subpop 1
h2<-0.005  # hazard rate for subpop 2
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation

# generate M datasets

data<-mixture.exp(n,h1,n,h2,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=24)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue(100,h1,h2,time, study.time)
plot.both(lik,freq.mat,true)

################################################################################
# h1 = 0.05, h2 = 0.005
M<-50 # number of datasets
time<-24 # analysis time
n<-30  # number of indivs in each of two subpops
atime<-24 # accrual time
h1<-0.05 # hazard rate for subpop 1
h2<-0.005  # hazard rate for subpop 2
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation

# generate M datasets

data<-mixture.exp(n,h1,n,h2,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=24)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue(100,h1,h2,time, study.time)
plot.both(lik,freq.mat,true)


#####################################################################################
#####################################################################################
# % cured:  h1=0.03, n=60, pct=0.5

M<-50 # number of datasets
time<-24 # analysis time
atime<-24 # accrual time
n<-30  # number of indivs in each of two subpops
h1<-0.03 # hazard rate for subpop 1
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation
pct<-0.50

# generate M datasets

data<-mixture.cured(pct,h1,2*n,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=24)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue.cured(pct,100, h1, study.time)
plot.both(lik,freq.mat,true) 

#####################################################################################
# % cured:  pct=0.3, h1=0.01, n=60

M<-50 # number of datasets
time<-24 # analysis time
atime<-24 # accrual time
n<-30  # number of indivs in each of two subpops
h1<-0.01 # hazard rate for subpop 1
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation
pct<-0.30

# generate M datasets

data<-mixture.cured(pct,h1,2*n,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=24)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue.cured(pct,100, h1, study.time)
plot.both(lik,freq.mat,true) 


#####################################################################################
# % cured:  pct=0.7, h1=0.05, n=60

M<-50 # number of datasets
time<-24 # analysis time
atime<-24 # accrual time
n<-30  # number of indivs in each of two subpops
h1<-0.05 # hazard rate for subpop 1
study.time<- 60 # total study time (i.e. accrual time + min follow-up time)
lmba<-seq(0.001,0.05,0.0001) # range of lambda's for estimation
pct<-0.70

# generate M datasets

data<-mixture.cured(pct,h1,2*n,atime,M,study.time)

# likelihood estimates of data
lik <- lik.result(lmba, data, time=time) 

# analyze data by frequentist approach
freq.mat<-matrix(0,M,3)
for(i in 1:M) {
	freq.mat[i,]<-freq.int(data[,c(2,5),i],time=24)
}

# estimate true survival time and plot both estimates for M datasets
true<-simtrue.cured(pct,100, h1, study.time)
plot.both(lik,freq.mat,true)