A Comparative Analysis on Online Handwritten Strokes Classification Using Online Learning

Abstract

The online handwriting recognition is recognition of handwritten data through the machine using a digital pen. The online learning includes training of the classifier with test data and the test data becomes part of a training model for next test data. We have done a novel study first in this direction to experiment online learning with online handwritten strokes. The experimentation carried out with benchmarked datasets as unipen and online handwritten Gurmukhi script strokes including 12,477 and 26,572 samples, respectively. The tool used in experimentation is Libol which includes all the state of art algorithms for online learning. The results indicate that online learning could be a suitable choice for online handwriting recognition. The online learning is popular today for its use with large data and less computation time. The present study could be benefited for online handwriting recognition like applications in online learning environments.