KNN ALGORITHM- A GEOMETRIC INTUITION

K Nearest Neighbors (KNN) is one useful supervised ML algorithm, generally used for classification problems but can be used for regression problems too. It is non-parametric algorithm, meaning the algorithm does no make any particular assumption about the kind of mapping function or the underlying data. For now, let us assume we are working on a binary classification problem and each observation in the data set (2-D) belongs to either positive class or negative class The main purpose of classification is given a point'xq', we have to predict the output value'yq'. Please note: There are various types of methods to calculate the distance between two points. I will try to cover them in a separate article later.