snn (version 1.1)

mybnn: Bagged Nearest Neighbor Classifier

Description

Implement the bagged nearest neighbor classification algorithm to predict the label of a new input using a training data set.

Usage

mybnn(train, test, ratio)

Arguments

train

Matrix of training data sets. An n by (d+1) matrix, where n is the sample size and d is the dimension. The last column is the class label.

test

Vector of a test point. It also admits a matrix input with each row representing a new test point.

ratio

Resampling ratio.

Value

It returns the predicted class label of the new test point. If input is a matrix, it returns a vector which contains the predicted class labels of all the new test points.

Details

The bagged nearest neighbor classifier is asymptotically equivalent to a weighted nearest neighbor classifier with the i-th weight a function of the resampling ratio, the sample size n, and i. See Hall and Samworth (2005) for details. The tuning parameter ratio can be tuned via cross-validation, see cv.tune function for the tuning procedure.

References

Hall, P. and Samworth, R. (2005). Properties of Bagged Nearest Neighbor Classifiers. Journal of the Royal Statistical Society, Series B, 67, 363-379.

Examples

	# Training data
	set.seed(1)
	n = 100
	d = 10
	DATA = mydata(n, d)

	# Testing data
	set.seed(2015)
	ntest = 100  
	TEST = mydata(ntest, d)
	TEST.x = TEST[,1:d]
	
	# bagged nearest neighbor classifier
	mybnn(DATA, TEST.x, ratio = 0.5)