MATLAB: How to increase knnclassif​ication_ac​curacy

Question

if i want to increase classification accuracy from 70 % to 80 % what kind of actions i must do ?

i have two excel tables , and the algorithm is as the following

class = knnclassify(sample, training, group,1,'cityblock','nearest' ); %classify (parameters optional)
[class result]
 difference = abs(class - result);
 dene = ((length(group))-(sum(difference)));
 % calculates classification acuracy 
classification_accuracy = (dene * 100/(length(group)))

how i increase the classification to 80 % ?

thanks

Answer 1

The KNN classifier is pretty much the ‘gold standard’ against which other classifiers are measured, at least in my experience. You can change distance metrics to see if you can improve the accuracy, but that is likely your only option.