Natural-language-classification

Single layer neural network composed of n perceptrons with sigmoid unipolar activation function (n is the number of languages to classify). Each perceptron is trained separately to recognise one language on the basis of A-Z letters frequency in input text. Decision which language is given is made using selector maximum. Training and test datasets consist of Divine Comedy by Dante Alighieri split into 100 lines files with letters outside English alphabet being removed.