KODAMA (version 0.0.1)

knn.probability: KNN Prediction Probability Routine using Pre-Calculated Distances

Description

K-Nearest Neighbor prediction probability method which uses the distances calculated by knn.dist. For predictions (not probabilities) see knn.predict.

Usage

knn.probability(train, test, y, dist.matrix, k=1, ties.meth="min")

Arguments

train

indexes which specify the rows of the dist.matrix to use as training set.

test

indexes which specify the rows of the dist.matrix to use as test set.

y

a vector of labels.

dist.matrix

the output from a call to knn.dist.

k

the number of nearest neighbors to consider.

ties.meth

method to handle ties for the k-th neighbor, the default is "min" which uses all ties, alternatives include "max" which uses none if there are ties for the k-th nearest neighbor, "random" which selects among the ties randomly and "first" which uses the ties in their order in the data.

Value

A matrix of prediction probabilities whose number of colimns is the number of test cases and the number of rows is the number of levels in the responses.

Details

Predictions are calculated for each test case by aggregating the responses of the k-nearest neighbors among the training cases and using the classprob. k may be specified to be any positive integer less than the number of training cases, but is generally between 1 and 10. The indexes for the training and test cases are in reference to the order of the entire data set as it was passed to knn.dist. The ties are handled using the rank function. Further information may be found by examining the ties.method there.

See Also

knn.dist,knn.predict

Examples

# the iris example used by knn(class)
library(class)
data(iris3)
train <- rbind(iris3[1:25,,1], iris3[1:25,,2], iris3[1:25,,3])
test <- rbind(iris3[26:50,,1], iris3[26:50,,2], iris3[26:50,,3])
cl <- factor(c(rep("s",25), rep("c",25), rep("v",25)))
# how to get predictions from knn(class)
pred <- knn(train, test, cl, k = 3, prob=TRUE)
# display the confusion matrix
table(pred,cl)
# view probabilities (only the highest probability is returned)
attr(pred,"prob")
# how to get predictions with knn.dist and knn.predict
x <- rbind(train,test)
kdist <- knn.dist(x)
pred <- knn.predict(1:75, 76:150, cl, kdist, k=3)
# display the confusion matrix
table(pred,cl)
# view probabilities (all class probabilities are returned)
knn.probability(1:75, 76:150, cl, kdist, k=3)
# to compare probabilites, rounding done for display purposes
p1 <- knn(train, test, cl, k = 3, prob=TRUE)
p2 <- round(knn.probability(1:75, 76:150, cl, kdist, k=3), digits=2)
table( round(attr(p1,"prob"), digits=2), apply(p2,2,max) )
# note any small differences in predictions are a result of
# both methods breaking ties in majority class randomly