The complexity of a natural language itself is very challenging as the natural language is not free from ambiguity problem. It is almost impossible to identify that the given text is having sense or not. In today's scenario it becomes even much important to detect that input is given by human or a machine. A valid input with sense is needed everywhere from Social media platforms to Business Intelligence. This Classification algorithm aims to detect whether the given input text is valid, or randomly typed in a keyboard. It returns a percentage value where a lower one means valid text, and a higher value means random text. The approach is based on identifying that the amount of unique chars, amount of vowels of letters, the word/char ratio (in %) are in a usual range. Then it further calculates "deviation score" to compute the accuracy of given input.

Data mining text mining text classification sentence validation pattern learning