selfTrainingG {ssc}

Self-training generic method

Description

Self-training is a simple and effective semi-supervised learning classification method. The self-training classifier is initially trained with a reduced set of labeled examples. Then it is iteratively retrained with its own most confident predictions over the unlabeled examples. Self-training follows a wrapper methodology using one base supervised classifier to establish the possible class of unlabeled instances.

Usage

selfTrainingG(y, gen.learner, gen.pred, max.iter = 50, perc.full = 0.7,
  thr.conf = 0.5)

Arguments

y

A vector with the labels of training instances. In this vector the unlabeled instances are specified with the value NA.

gen.learner

A function for training a supervised base classifier. This function needs two parameters, indexes and cls, where indexes indicates the instances to use and cls specifies the classes of those instances.

gen.pred

A function for predicting the probabilities per classes. This function must be two parameters, model and indexes, where the model is a classifier trained with gen.learner function and indexes indicates the instances to predict.

max.iter

Maximum number of iterations to execute the self-labeling process. Default is 50.

perc.full

A number between 0 and 1. If the percentage of new labeled examples reaches this value the self-training process is stopped. Default is 0.7.

thr.conf

A number between 0 and 1 that indicates the confidence theshold. At each iteration, only the newly labelled examples with a confidence greater than this value (thr.conf) are added to the training set.

Details

SelfTrainingG can be helpful in those cases where the method selected as base classifier needs learner and pred functions with other specifications. For more information about the general self-training method, please see the selfTraining function. Essentially, the selfTraining function is a wrapper of the selfTrainingG function.

Value

A list object of class "selfTrainingG" containing:

model

The final base classifier trained using the enlarged labeled set.

instances.index

The indexes of the training instances used to train the model. These indexes include the initial labeled instances and the newly labeled instances. Those indexes are relative to the y argument.

Examples

library(ssc)

## Load Wine data set
data(wine)

cls <- which(colnames(wine) == "Wine")
x <- wine[, -cls] # instances without classes
y <- wine[, cls] # the classes
x <- scale(x) # scale the attributes

## Prepare data
set.seed(20)
# Use 50% of instances for training
tra.idx <- sample(x = length(y), size = ceiling(length(y) * 0.5))
xtrain <- x[tra.idx,] # training instances
ytrain <- y[tra.idx]  # classes of training instances
# Use 70% of train instances as unlabeled set
tra.na.idx <- sample(x = length(tra.idx), size = ceiling(length(tra.idx) * 0.7))
ytrain[tra.na.idx] <- NA # remove class information of unlabeled instances

# Use the other 50% of instances for inductive testing
tst.idx <- setdiff(1:length(y), tra.idx)
xitest <- x[tst.idx,] # testing instances
yitest <- y[tst.idx] # classes of testing instances

## Example: Training from a set of instances with 1-NN (knn3) as base classifier.
gen.learner <- function(indexes, cls)
  caret::knn3(x = xtrain[indexes, ], y = cls, k = 1)
gen.pred <- function(model, indexes)
  predict(model, xtrain[indexes, ]) 

md1 <- selfTrainingG(y = ytrain, gen.learner, gen.pred)

cls1 <- predict(md1$model, xitest, type = "class")
table(cls1, yitest)

## Example: Training from a distance matrix with 1-NN (oneNN) as base classifier.
dtrain <- as.matrix(proxy::dist(x = xtrain, method = "euclidean", by_rows = TRUE))
gen.learner <- function(indexes, cls) {
  m <- ssc::oneNN(y = cls)
  attr(m, "tra.idxs") <- indexes
  m
}

gen.pred <- function(model, indexes)  {
  tra.idxs <- attr(model, "tra.idxs")
  d <- dtrain[indexes, tra.idxs]
  prob <- predict(model, d, distance.weighting = "none")
  prob
}

md2 <- selfTrainingG(y = ytrain, gen.learner, gen.pred)
ditest <- proxy::dist(x = xitest, y = xtrain[md2$instances.index,],
                      method = "euclidean", by_rows = TRUE)
cls2 <- predict(md2$model, ditest, type = "class")
table(cls2, yitest)