Intrepetability of KNN -Tulasiram

Cancer is an abnormal disease.As in olden days it is very hard to detect cancer.But when the ML is used in medical domain it is found easy to detect some of the dangerous diseases like pneumonia,cancer etc…The ambit of ML is extended to detect cancer.From the past 20 years Machine Learning is been using to detect cancers.But in medical domains the black box models(models which take input and provide output without giving any reasons) are not useful as they do not give any reasons behind . It is very absurd to say to the patient that you are suffering from a cancer without a reason.This is where intrepretability matters. Intrepretability helps to win trust in algorithms that try to improve on suggestions rather than just predicting.So our model should intrepret the reason behind the output it is giving.One of such model is K-NN(K-NEAREST NEIGHBORS).


Click here to check my complete article.