MATLAB Answers


Can K-nearest neighbor classify more than two classes?

Asked by gugu
on 17 Dec 2017
Latest activity Commented on by gugu
on 18 Dec 2017
I want to know that K-nearest neighbor classifier can classify more than two classes? I don't understand much about this classifier.


In general "knn" methods are able to find more than 2 classes. A look into a textbook, your script, WikiPedia of the Matlab documentation should help you to learn more about this topic. If a detail about Matlab is not clear to you, please ask specifically, otherwise answering is hard.
Ok. I'll have a go at it. Actually, I want to classify three classes and now I'm currently using multi-svm. It doesn't meet my requirements. So, I want to change the classifier. Thank you for your answer.

Sign in to comment.

1 Answer

Answer by the cyclist
on 17 Dec 2017

Yes, it can. There is an example of a 3-class classification in the documentation for fitcknn, in the Statistics and Machine Learning Toolbox.
load fisheriris
X = meas;
Y = species;
% X is a numeric matrix that contains four petal measurements for 150 irises.
% Y is a cell array of character vectors that contains the corresponding iris species.
% Train a 5-nearest neighbors classifier. It is good practice to standardize noncategorical predictor data.
Mdl = fitcknn(X,Y,'NumNeighbors',5,'Standardize',1)


Show 1 older comment
Let me ask you one question. Does the result of the classifier depend on the number of training data? I mean the larger the training data, the better the result. Is that so?
Not necessarily. Denser data may make it more accurate, but just expanding the feature space so that it's bigger will not affect the results. For example if you had 500 training points in the range x=4-5, and y = 4-5, and you increased it to 2000 training points in that same area would make it more accurate. However if you still had 500 points there and just added another few thousand points in the range 0-10 in both x and y, would not necessarily make your classifications in the range 4-5 any better than what they were, since it would be using the same points as before.
It's clear on my mind. Thank you so much.

Sign in to comment.