tests/testthat/test-examples.R

In changepoint: Methods for Changepoint Detection

context("man file example tests")

# From changepoint-package.Rd
# change in variance
set.seed(1)
x=c(rnorm(100,0,1),rnorm(100,0,10))
ansvar=cpt.var(x)
test_that('var1',expect_equal(cpts(ansvar),100))

# change in mean
set.seed(1)
y=c(rnorm(100,0,1),rnorm(100,5,1))
ansmean=cpt.mean(y)
test_that('mean1',expect_equal(cpts(ansmean),100))

# change in mean and variance
set.seed(1)
z=c(rnorm(100,0,1),rnorm(100,2,10))
ansmeanvar=cpt.meanvar(z)
test_that('meanvar1',expect_equal(cpts(ansmeanvar),100))







# From cpt.mean.Rd
# Example of a change in mean at 100 in simulated normal data
set.seed(1)
x=c(rnorm(100,0,1),rnorm(100,10,1))
test_that('mean2',expect_equivalent(cpt.mean(x,penalty="SIC",method="AMOC",class=FALSE),c(100,1)))
ans=cpt.mean(x,penalty="Asymptotic",pen.value=0.01,method="AMOC")
test_that('mean3',expect_equal(cpts(ans),100))
ans=cpt.mean(x,penalty="Manual",pen.value=0.8,method="AMOC",test.stat="CUSUM")
test_that('mean4',expect_equivalent(cpts(ans),101))

# Example of multiple changes in mean at 50,100,150 in simulated normal data
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,1),rnorm(50,10,1),rnorm(50,3,1))
test_that('mean5',expect_equal(cpt.mean(x,penalty="Manual",pen.value="2*log(n)",method="BinSeg",Q=5,class=FALSE),c(50,100,150,200)))

# Example of using the CROPS penalty in data set above
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,1),rnorm(50,10,1),rnorm(50,3,1))
out=cpt.mean(x, pen.value = c(4,1500),penalty = "CROPS",method = "PELT")
truth=matrix(NA,ncol=7,nrow=7); truth[1:6,1]=50;truth[1:5,2]=c(96,96,100,100,150)
truth[1:4,3]=c(100,100,133,150);truth[1:3,4]=c(133,133,150);truth[1:2,5]=c(150,150)
truth[1,6]=159;truth[1,7]=180
test_that('crops1',expect_equivalent(cpts.full(out),truth))
truth=c(4.000000,   4.332496,   4.385247,   4.684254 ,559.366988, 646.962719,1311.335695,1500)
test_that('crops2',expect_equal(pen.value.full(out),truth,tolerance=1e-6))

# Example multiple datasets where the first row has multiple changes in mean and the second row has
#no change in mean
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,1),rnorm(50,10,1),rnorm(50,3,1))
y=rnorm(200,0,1)
z=rbind(x,y)
ans=cpt.mean(z,penalty="Asymptotic",pen.value=0.01,method="SegNeigh",Q=5,class=FALSE)
test_that('mean6',expect_equal(ans,list(c(50,100,150,200),200)))
ans=cpt.mean(z,penalty="Asymptotic",pen.value=0.01,method="PELT")
test_that('mean7',expect_equal(cpts(ans[[1]]),c(50,100,150)))
test_that('mean8',expect_equal(cpts(ans[[2]]),numeric()))







# From cpt.meanvar.Rd
# Example of a change in scale parameter (mean and variance) at 100 in simulated gamma data
set.seed(1)
x=c(rgamma(100,shape=1,rate=1),rgamma(100,shape=1,rate=5))
test_that('meanvar2',expect_equivalent(cpt.meanvar(x,penalty="SIC",method="AMOC",test.stat="Gamma",class=FALSE,shape=1),98))
ans=cpt.meanvar(x,penalty="AIC",method="AMOC",test.stat="Gamma",shape=1)
test_that('meanvar3',expect_equivalent(cpts(ans),98))

# Example of multiple changes in mean and variance at 50,100,150 in simulated normal data
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,3),rnorm(50,10,1),rnorm(50,3,10))
test_that('meanvar4',expect_equal(cpt.meanvar(x,penalty="Manual",pen.value="4*log(n)",method="BinSeg",Q=5,class=FALSE),c(50,100,150,152,200)))

# Example of using the CROPS penalty in the above example
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,3),rnorm(50,10,1),rnorm(50,3,10))
out=cpt.meanvar(x,pen.value=c(2*log(length(x)),100*log(length(x))),penalty="CROPS",method="PELT")
truth=matrix(NA,ncol=9,nrow=6);truth[1:5,1]=c(rep(15,2),rep(50,3));truth[1:4,2]=c(17,17,100,100)
truth[1:4,3]=c(22,22,133,150);truth[1:3,4]=c(44,50,151);truth[1:2,5]=c(46,100)
truth[1:2,6]=c(50,133);truth[1:2,7]=c(100,151);truth[1,8]=133;truth[1,9]=151
test_that('crops3',expect_equal(cpts.full(out),truth))
truth=c(10.59663,  10.68431,  11.31088, 11.38307,  119.78669, 191.42622,529.83174)
test_that('crops4',expect_equal(pen.value.full(out),truth,tolerance=1e-6))

# Example multiple datasets where the first row has multiple changes in mean and variance and the
#second row has no change in mean or variance
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,5,3),rnorm(50,10,1),rnorm(50,3,10))
y=rnorm(200,0,1)
z=rbind(x,y)
ans=cpt.meanvar(z,penalty="Asymptotic",pen.value=0.01,method="SegNeigh",Q=5,class=FALSE)
test_that('meanvar5',expect_equivalent(ans,list(c(50,100,150,200),200)))
ans=cpt.meanvar(z,penalty="Asymptotic",pen.value=0.01,method="PELT")
test_that('meanvar6',expect_equivalent(cpts(ans[[1]]),c(50,100,150)))
test_that('meanvar7',expect_equivalent(cpts(ans[[2]]),numeric()))








# From cpt.range-class.Rd
x=new("cpt.range")
test_that('class1',expect_is(x,"cpt.range"))
cpts(x)<-c(10,50,100)
test_that('class2',expect_equivalent(cpts(x),c(10,50,100)))

# Example of multiple changes in variance at 50,100,150 in simulated normal data
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
out=cpt.var(x,pen.value=c(log(length(x)),10*log(length(x))),penalty="CROPS",method="PELT")
ll=logLik(out,ncpts=3)
test_that('class3',expect_equivalent(ll,c(925.8085, 947.0578))) # raw likelihood of the data with changepoints, second value is likelihood + penalty




#



# From cpt.reg-class.Rd
x=new("cpt.reg") # creates a new object with the cpt.reg class defaults
test_that('class4',expect_is(x,"cpt.reg"))
test_that('class5',expect_is(data.set(x),"matrix"))
data.set(x)<-matrix(1:10,nrow=5,ncol=2) # replaces the data.set slot from x with a matrix
test_that('class6',expect_equivalent(data.set(x),matrix(1:10,nrow=5,ncol=2)))







# From cpt.var.Rd
# Example of a change in variance at 100 in simulated normal data
set.seed(1)
x=c(rnorm(100,0,1),rnorm(100,0,10))
test_that('var2',expect_equivalent(cpt.var(x,penalty="SIC",method="AMOC",class=FALSE),c(100,1)))
ans=cpt.var(x,penalty="Asymptotic",pen.value=0.01,method="AMOC")
test_that('var3',expect_equivalent(cpts(ans),100))

# Example of multiple changes in variance at 50,100,150 in simulated data
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
out=cpt.var(x,penalty="Manual",pen.value="log(2*log(n))",method="BinSeg",test.stat="CSS",Q=5,
            class=FALSE)
truth=list();truth$cps=matrix(c(99,53,150,50,140,3.156304,3.156304,3.156304,3.074743,1.254542),byrow=T,nrow=2)
truth$cpts=c(50,53,99,150,200);truth$op.cpts=4;truth$pen=2.360536
test_that('var4',expect_equal(out,truth,tolerance=0.00001))

# Example of using CROPS in the above example
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
out=cpt.var(x,pen.value=c(log(length(x)),100*log(length(x))),penalty="CROPS",method="PELT")
truth=matrix(NA,ncol=7,nrow=7);truth[1:6,1]=50;truth[1:5,2]=c(77,rep(99,3),150);truth[1:4,3]=c(79,114,140,150)
truth[1:3,4]=c(99,133,150);truth[1:2,5]=c(114,150);truth[1,6]=133;truth[1,7]=150
test_that('var5',expect_equivalent(cpts.full(out),truth))
truth=c(5.298317,5.548538,6.149305,7.083099,26.592259,142.417161,145.146279,529.831737)
test_that('var6',expect_equivalent(pen.value.full(out),truth))

# Example multiple datasets where the first row has multiple changes in variance and the second row
#has no change in variance
set.seed(10)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
y=rnorm(200,0,1)
z=rbind(x,y)
truth=list();truth[[1]]=c(50,100,149,200);truth[[2]]=200
ans=cpt.var(z,penalty="Asymptotic",pen.value=0.01,method="SegNeigh",Q=5,class=FALSE)
test_that('var7',expect_equivalent(ans,truth))
ans=cpt.var(z,penalty="Asymptotic",pen.value=0.01,method="PELT")
test_that('var8',expect_equivalent(cpts(ans[[1]]),c(50,100,149)))
test_that('var9',expect_equivalent(cpts(ans[[2]]),numeric()))






# From cpt-class.Rd
x=new("cpt") # creates a new object with the cpt class defaults
test_that('class7',expect_is(x,"cpt"))
test_that('class8',expect_equivalent(cpts(x),numeric()))
cpts(x)<-c(10,50,100) # replaces the cpts slot from x with c(10,50,100)
test_that('class9',expect_equivalent(cpts(x),c(10,50,100)))

# Example of a change in variance at 100 in simulated normal data
set.seed(1)
x=c(rnorm(100,0,1),rnorm(100,0,10))
ans=cpt.var(x)
test_that('class10',expect_equivalent(logLik(ans),c(1003.2283241358,1012.438665)))




# From cpts.full.Rd
x=new("cpt.range") # new cpt.range object
test_that('class11',expect_is(x,"cpt.range"))
test_that('class12',expect_is(cpts.full(x),"matrix")) # retrieves the cpts.full slot from x




# From cpts.full-.Rd
x=new("cpt.range") # new cpt.range object
cpts.full(x)<-matrix(c(10,20,10,NA),nrow=2,byrow=TRUE)
test_that('class13',expect_equivalent(cpts.full(x),matrix(c(10,20,10,NA),nrow=2,byrow=TRUE) ))




# From cpts.Rd
x=new("cpt") # new cpt object
test_that('class14',expect_equivalent(cpts(x),numeric())) # retrieves the cpts slot from x





# From cpts.ts.Rd
x=new("cpt") # new cpt object
test_that('class15',expect_equivalent(cpts.ts(x),numeric()))




# From cpts-.Rd
x=new("cpt") # new cpt object
cpts(x)<-10 # replaces the vector of changepoint in object x with 10
test_that('class16',expect_equivalent(cpts(x),10))





# From cpttype.Rd
x=new("cpt") # new cpt object
test_that('class17',expect_equivalent(cpttype(x),"Not Set")) # retrieves the cpttype slot from x





# From cpttype-.Rd
x=new("cpt") # new cpt object
cpttype(x)<-"mean" # replaces the existing cpttype in object x with "mean"
test_that('class18',expect_equivalent(cpttype(x),'mean'))




# From data.set.Rd
x=new("cpt") # new cpt object
test_that('class19',expect_equivalent(data.set(x),ts()))




# From data.set.ts.Rd
x=new("cpt") # new cpt object
test_that('class20',expect_equivalent(data.set.ts(x),ts()))





# From data.set-.Rd
x=new("cpt") # new cpt object
data.set(x)<-c(1,2,3,4,5) # replaces the existing data.set slot in x with c(1,2,3,4,5)
test_that('class21',expect_equivalent(data.set(x),1:5))




# From distribution.Rd
x=new("cpt") # new cpt object
test_that('class22',expect_equivalent(distribution(x),character()))





# From distribution-.Rd
x=new("cpt") # new cpt object
distribution(x)<-"normal" # replaces the current distribution slot of x with "normal"
test_that('class23',expect_equivalent(distribution(x),"normal"))






# From likelihood.Rd
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
out=cpt.var(x,penalty="Manual",pen.value="2*log(n)",method="BinSeg",Q=5)
ll=likelihood(out)
test_that('logLik1',expect_equivalent(ll,c(925.8085, 957.5984)))





# From logLik-methods.Rd
set.seed(1)
x=c(rnorm(50,0,1),rnorm(50,0,10),rnorm(50,0,5),rnorm(50,0,1))
out=cpt.var(x,penalty="Manual",pen.value="2*log(n)",method="BinSeg",Q=5)
ll=logLik(out)
test_that('logLik1',expect_equivalent(ll,c(925.8085, 957.5984)))




# From method.Rd
x=new("cpt") # new cpt object
test_that('class24',expect_equivalent(method(x),character()))





# From method-.Rd
x=new("cpt") # new cpt object
method(x)<-"mean" # replaces the existing method slot in x with "mean"
test_that('class25',expect_equivalent(method(x),"mean"))




# From minseglen.Rd
x=new("cpt") # new cpt object
test_that('class26',expect_equivalent(minseglen(x),numeric()))





# From minseglen-.Rd
x=new("cpt") # new cpt object
minseglen(x)<-5 # replaces the existing minseglen slot in x with 5
test_that('class27',expect_equivalent(minseglen(x),5))




# From ncpts.max.Rd
x=new("cpt") # new cpt object
test_that('class28',expect_equivalent(ncpts.max(x),numeric()))





# From ncpts.max-.Rd
x=new("cpt") # new cpt object
ncpts.max(x)<-10 # replaces the vector of changepoint in object x with 10
test_that('class29',expect_equivalent(ncpts.max(x),10))






# From ncpts.Rd
x=new("cpt") # new cpt object
test_that('class30',expect_equivalent(ncpts(x),0)) # returns the number of changepoints (i.e. length of the cpts slot in x minus 1)






# From nseg.Rd
x=new("cpt") # new cpt object
test_that('class31',expect_equivalent(nseg(x),1))





# From param.est.Rd
x=new("cpt") # new cpt object
test_that('class32',expect_equivalent(param.est(x),list()))





# From param.est-.Rd
x=new("cpt") # new cpt object
param.est(x)<-list(mean=0) # replaces the current param.est list in x with list(mean=0)
test_that('class33',expect_equivalent(param.est(x),list(mean=0)))






# From param.Rd
set.seed(1)
x=c(rnorm(100,0,1),rnorm(100,0,10))
ans=cpt.var(x,penalty="Asymptotic",pen.value=0.01,method="AMOC",param.estimates=FALSE)
ans=param(ans) # fills the param.est slot with the parameter estimes.
test_that('class34',expect_equivalent(param.est(ans),list(variance=c(0.7986945, 90.8356989))))






# From pen.type.Rd
x=new("cpt") # new cpt object
test_that('class35',expect_equivalent(pen.type(x),character()))





# From pen.type-.Rd
x=new("cpt") # new cpt object
pen.type(x)<-"SIC" # replaces the existing pen.type slot in x with "SIC"
test_that('class36',expect_equivalent(pen.type(x),"SIC"))






# From pen.value.full.Rd
x=new("cpt.range") # new cpt.range object
test_that('class37',expect_equivalent(pen.value.full(x),numeric()))





# From pen.value.full-.Rd
x=new("cpt.range") # new cpt.range object
pen.value.full(x)<-5 # replaces the existing pen.value.full slot in x with 5
test_that('class38',expect_equivalent(pen.value.full(x),5))





# From pen.value.Rd
x=new("cpt") # new cpt object
test_that('class39',expect_equivalent(pen.value(x),numeric()))




# From pen.value-.Rd
x=new("cpt") # new cpt object
pen.value(x)<-5 # replaces the existing pen.value slot in x with 5
test_that('class40',expect_equivalent(pen.value(x),5))





# From seglen.Rd
x=new("cpt") # new cpt object
test_that('class41',expect_equivalent(seg.len(x),numeric()))





# From test.stat.Rd
x=new("cpt") # new cpt object
test_that('class42',expect_equivalent(test.stat(x),character()))




# From test.stat-.Rd
x=new("cpt") # new cpt object
test.stat(x)<-"normal" # replaces the current test.stat slot of x with "normal"
test_that('class43',expect_equivalent(test.stat(x),"normal"))