# Depth-Based kNN The implementation of the affine-invariant depth-based kNN of Paindaveine and Van Bever (2015). ```r dknn.classify(objects, data, k, depth = "halfspace", seed = 0) ``` ## Arguments - `objects`: Matrix containing objects to be classified; each row is one $d$-dimensional object. - `data`: Matrix containing training sample where each of $n$ rows is one object of the training sample where first $d$ entries are inputs and the last entry is output (class label). - `k`: the number of neighbours - `depth`: Character string determining which depth notion to use; the default value is `"halfspace"`. Currently the method supports the following depths: "halfspace", "Mahalanobis", "simplicial". - `seed`: the random seed. The default value `seed=0` makes no changes. ## Returns List containing class labels, or character string "Ignored" for the outsiders if "Ignore" was specified as the outsider treating method. ## References Paindaveine, D. and Van Bever, G. (2015). Nonparametrically consistent depth-based classifiers. **Bernoulli** 21 62--82. ## See Also `dknn.train` to train the Dknn-classifier. `dknn.classify.trained` to classify with the Dknn-classifier. `ddalpha.train` to train the DD$\alpha$-classifier. `ddalpha.getErrorRateCV` and `ddalpha.getErrorRatePart` to get error rate of the Dknn-classifier on particular data (set `separator = "Dknn"`). ## Examples ```r # Generate a bivariate normal location-shift classification task # containing 200 training objects and 200 to test with class1 <- mvrnorm(200, c(0,0), matrix(c(1,1,1,4), nrow = 2, ncol = 2, byrow = TRUE)) class2 <- mvrnorm(200, c(2,2), matrix(c(1,1,1,4), nrow = 2, ncol = 2, byrow = TRUE)) trainIndices <- c(1:100) testIndices <- c(101:200) propertyVars <- c(1:2) classVar <- 3 trainData <- rbind(cbind(class1[trainIndices,], rep(1, 100)), cbind(class2[trainIndices,], rep(2, 100))) testData <- rbind(cbind(class1[testIndices,], rep(1, 100)), cbind(class2[testIndices,], rep(2, 100))) data <- list(train = trainData, test = testData) # Train the classifier # and get the classification error rate cls <- dknn.train(data$train, kMax = 20, depth = "Mahalanobis") cls$k classes1 <- dknn.classify.trained(data$test[,propertyVars], cls) cat("Classification error rate: ", sum(unlist(classes1) != data$test[,classVar])/200) # Classify the new data based on the old ones in one step classes2 <- dknn.classify(data$test[,propertyVars], data$train, k = cls$k, depth = "Mahalanobis") cat("Classification error rate: ", sum(unlist(classes2) != data$test[,classVar])/200) ```

dknn.classify function