Clear Filters
Clear Filters

How can i get KNN classifier for test and training purpose?

1 view (last 30 days)
I want to study about KNN classifier and i would like to implement it..Where can i get source code..

Answers (1)

Star Strider
Star Strider on 8 Mar 2017
If you do not have the Statistics and Machine Learning Toolbox and with it the knnsearch (link) function, a simple KNN classifier is straightforward to write.
The Code
V = randi(50, 1, 3); % Vector — Create Data
M = randi(50, 15, 3); % Matrix — Create Data
dif = bsxfun(@minus, V, M); % Subtract Vector from Matrix
D2 = sqrt(sum(dif.^2,2)); % Euclidean Distance Metric
[Ds,Ix] = sort(D2,'ascend'); % Sort Ascending
k = 5; % Number Of Neighbours
KNN = M(Ix(1:k),:); % K-th Nearest Neighbours
This finds the ‘nearest neighbours’ in ‘M’ to the vector ‘V’. The data in ‘V’ and ‘M’ must have the same column size.
  2 Comments
Star Strider
Star Strider on 8 Mar 2017
No. This is code I wrote earlier to illustrate how a k-th nearest neighbour classifier works.
The link I provided in my Answer is to the documentation for the MATLAB function. You must have the Statistics and Machine Learning Toolbox installed to use the knnsearch function.

Sign in to comment.

Categories

Find more on Statistics and Machine Learning Toolbox in Help Center and File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!