Hi Satya,

to use fitcknn you have to put all the cases (from both groups) in one matrix (Tbl input parameter). Each row of that matrix corresponds to one case and each column with a prediction variable. Then create an Y array with the response variable (0 for healthy- 1 for diseased for example). The ith element of Y gives you the class of the ith row in Tbl,

Anyhow you should have a look at fitcknn documentation where other input options  are explained.

https://es.mathworks.com/help/stats/classificationknn-class.html 

Once you have obtained the model using fitcknn you can use predict() to classify new cases.

https://es.mathworks.com/help/stats/classificationknn.predict.html

Hope this helps,

Jaime

Dear Satya,

Would you like to collaborate? I can help you apply KNN using MATLAB on the data set if you would. 

Pascal

I would highly appreciate your interest in helping me. But collaborate in the sense we work in a project together? I am currently doing a project which is in the finishing stage. If we can go with that then I would be happy to collaborate with you. Anyway I would send you my data set to your e-mail. Please send me your e-mail @ Pascal

§  data1 = xlsread('DATASET2');

§  data = data1(:,1:end-1);

§  Rec = {'Engineering','Medicine/surgery','Pharmacy','Law', 'account','architecture','Quantity','Mass communication','too poor','good at all'};

§  normdata= (data - min(data(:)))./ (max(data(:))-min(data(:)));

§  normnew= ( new - min(data(:)))./ (max(data(:))-min(data(:)));

§  a=[];

§  v=[];

§  for i = (1:799)

§  distance = pdist2(normdata(i,:),normnew,'euclidean');

§  a(i)= distance;

§  end

§  k = sort(a);

§  knd = k(1:100);

§  for i = knd

§     vote= find(a == i);

§     v = [v vote];

§  end

§  n = 1;

§  for i = v

§      if i>=1 && i100 && i200 && i300 && i400 && i500 && i600 && i700 && i800 && i

Replace the DATASET2 with your data set.My dataset had 999 examples. I use 799 for the case base repository and 200 for testing. The Rec are the descriptions of the output. I have about ten categories of output in the dataset. Look through the lines of code and ask me questions wherever you are confused.

Thank you so much for your help. I will certainly look forward to you in case of any confusion @ Pascal