how noise effect value of k?? what is noise?
How noise is related to knn?
hey @lgoyal50_be19 ,
Noise is a bunch of our data that behaves differently from the normal trend of our data.
So if you take a large value of K with noise in data , the algorithm lacks to find the center points accurately , as these noise data points , being more far from the actual data , reduces actual algorithm performance. Hence , as a result you get poor results.
Smaller values of K , deal with this to an extent, but in then end you will get a reduction in actual performance of algorithm.
I hope this helped you understand it much better.
Thank You .