PEkit (version 1.0.0.1000)

tSimLab: Simultaneously predicted labels of the test data given the training data classification.

Description

Classifies the test data x based on the training data object. All of the test data is used simultaneously to make the classification.

Usage

tSimLab(training, x)

Arguments

training

A training data object from the function classifier.fit().

x

Test data vector or matrix with rows as data points and columns as features.

Value

A vector of predicted labels for test data x.

Details

The test data are first labeled with the marginal classifier. The simultaneous classifier then iterates over all test data, assigning each a label by finding the maximum predictive probability given the current classification structure of the test data as a whole. This is repeated until the classification structure converges after iterating over all data.

References

Amiryousefi A. Asymptotic supervised predictive classifiers under partition exchangeability. . 2021. https://arxiv.org/abs/2101.10950.

Corander, J., Cui, Y., Koski, T., and Siren, J.: Have I seen you before? Principles of Bayesian predictive classification revisited. Springer, Stat. Comput. 23, (2011), 59<U+2013>73, (<10.1007/s11222-011-9291-7>).

Examples

# NOT RUN {
## Create random samples x from Poisson-Dirichlet distributions with different
## psis, treating each sample as coming from a class of its own:
set.seed(111)
x1<-rPD(1050,10)
x2<-rPD(1050,1000)
test.ind1<-sample.int(1050,50) # Sample test datasets from the
test.ind2<-sample.int(1050,50) # original samples
x<-c(x1[-test.ind1],x2[-test.ind2])
## create training data labels:
y1<-rep("1", 1000)
y2<-rep("2", 1000)
y<-c(y1,y2)

## Test data t, with first half belonging to class "1", second have in "2":
t1<-x1[test.ind1]
t2<-x2[test.ind2]
t<-c(t1,t2)

fit<-classifier.fit(x,y)

## Run the classifier, which returns
tS<-tSimLab(fit, t)

##With multidimensional x:
set.seed(111)
x1<-cbind(rPD(500,1),rPD(500,5))
x2<-cbind(rPD(500,10),rPD(500,50))
test.ind1<-sample.int(500,50)
test.ind2<-sample.int(500,50)
x<-rbind(x1[-test.ind1,],x2[-test.ind2,])
y1<-rep("1", 450)
y2<-rep("2", 450)
y<-c(y1,y2)
fit<-classifier.fit(x,y)
t1<-x1[test.ind1,]
t2<-x2[test.ind2,]
t<-rbind(t1,t2)

tS<-tSimLab(fit, t)
# }