The use of the k nearest neighbor method to classify the representative elements

Abstract

The use of Statistics in Chemistry has grown significantly with advances in computation. Nowadays it is easier to deal with a large data set and extract relevant chemical information. This paper describes the use of k nearest neighbor method to classify the representative elements as metal or nonmetal according to their periodic properties: atomic radius, ionization energy, electron affinity and electronegativity. The method requires a very simple mathematical background and can be easily performed and understood. The algorithm classifies an object into a distinct class when there are two or more groups of objects of known class and takes into account the distances among the objects. The pedagogical objective is to present an interdisciplinary activity in which Statistics can be used to make comparisons in Chemistry.