Efficient Intrusion Detection with KNN Classification and DS Theory - Proceedings of All India Seminar on Biomedical Engineering 2012 (AISOBE-2012) - page 172

Fig. 1 Decision of nearest
neighbor

Fig. 2 Majority voting

scheme

Fig. 3 Weighted-sum voting

scheme

not far from the truth). This is in contrast to other techniques like SVM where you
can discard all nonsupport vectors without any problems. Most of the lazy algo-
rithms—especially KNN—makes decisions based on the entire training data set (in
the best case a subset of them).
There are various methods which can be used to determine the nearest neighbor.
Figure 1 shows the way in which decision is taken to decide the category of a new
point.
Figures 2 and 3 shows various methods for deciding the nearest neighbor.
k-NN is a kind of example-based text categorization algorithm. However, the
determination of the k has not yet obtained a good solution. Moreover, the good
selection of k most similar texts also has a bigger effect on the categorization
results. In addition, k-NN cannot effectively solve the problem of overlapped
category borders.