Online identification of English plain text using artificial neural network

Abstract

In online communication, most of the time plain English characters are transmitted, while a few are encrypted. Thus there is a need for an automatic recognizer of plain English text (based on the characteristics of the English Language) without using a dictionary. It works for continuous text without word break-up (text without blank spaces between words). We propose a very efficient artificial neural network-based technique by selecting relevant or important features using Joint Mutual Information for online recognition of English plain text which can recognize English text from English like or random data.