I am working on some lossless compression topic and looking on any resources or prior studies on lossless file entropy predictability.

Assuming we know type of the file (we can tell its belongs to certain group and it is not falsified). Can we predict file entropy of any file in that group of files based on its size?

Groups like: Books with English text (I am assuming the answer is yes), binary files (like from Apple App Store), maybe some web data, or text snippets for big data (like twitter data). Or even sensor data generated from particular simulation or sensors.

  • 3
    $\begingroup$ 1. Sure. What research have you done? Where have you looked, and what have you found so far? Have you spent some quality time with Google Scholar? The entropy per character of English text has been well-studied. And, you should be able to work out this information on your own, by measuring the entropy of a bunch of files of that type. 2. We prefer you ask one question per question. Asking about all file types sounds like it might potentially be a bit too broad to be a good fit here. $\endgroup$
    – D.W.
    Commented Jun 12, 2016 at 23:57


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.