Letter Recognition Data Using Neural Network
Hussein Salim Qasim · 2013
The letters dataset from the UCI repository website form a relatively complex problem to classify distorted raster images of English alphabets. In contrast to rather complex networks, difference boosting algorithm, a computationally less intensive Bayesian classifier, is found to produce comparable or better classification efficiency on this problem. The character images were, originally, based on 20 different fonts and each letter within these 20 fonts was randomly distorted to produce a file of 20,000 unique stimuli. Each stimulus was converted into 16 primitive numerical attributes (statistical moments and edge counts) which were then scaled to fit into a range of integer values from 0 through 15. We typically chose, randomly, 1,000 unique stimuli for study. We made sure that the distribution remains the same after choosing the one thousand stimuli. In this study, a neural network tool was developed for the purpose of predicting to identify each of a large number of black and white rectangular pixel displays as one of the 26 capital letters in the English alphabet. The character images were based on 20 different fonts and each letter within these 20 fonts was randomly distorted to produce a file of 1,000.