Browse Prior Art Database

Method to infer the accuracy of a user submitted value in an electronic form

IP.com Disclosure Number: IPCOM000239589D
Publication Date: 2014-Nov-18
Document File: 3 page(s) / 62K

Publishing Venue

The IP.com Prior Art Database

Abstract

This idea proposes a method to infer the accuracy of a user submitted field by keeping track of the (amount of) user edits that have been made to that field prior to submission of the value. The rationale is that multiple edits to a field prior to submission may indicate some amount of uncertainty on the users part on the validity or accuracy of the value. For a specific user, a high amount of edits indicate field accuracy, which can be used in fraud investigations. Across larger amounts of users, a high number of edits prior to submission indicates a usability problem or an ill-posed question.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 51% of the total text.

Page 01 of 3

Method to infer the accuracy of a user submitted value in an electronic form

Many forms that used to have to be completed on paper now have electronic equivalents. In particular, governments are rapidly digitizing their workflows, where

requests for information can now be submitted digitally instead of on paper. This requires stronger safeguards against fraud, as information can now be requested automatically, virtually anonymously and/or away from the processing location. This idea proposes to use keystroke logging to determine any input fields that may contain uncertain data. For example, a user's nationality may have been entered as "DUTCH

AMERICAN", which indicates some doubt on the user's part of what should be entered in this field.

    Keystroke logging itself has been around for multiple decades, but it has been mostly used in the context of security and security circumvention. We are unaware of any methods that combine keystroke logging in traditional web based input technologies to infer information about the data that has been entered.

    The idea proposes to store and transmit the full list of keystrokes for each input field whenever the form is submitted for processing. We can then infer additional meta information by examining the full keystroke log. This information includes but is not limited to:
- The original value that was intended to be submitted, by examining the full edit string.

- Any misspellings that may be suspect, for example a user misspelling their last name.

- Any fields whose content may be suspect, by looking at multiple erasures and re-edits.

- Any fields that may have been copied/pasted from somewhere else or completed automatically, by examining typing speed.

- Fields that may have been entered by different users, by examining the different typing patterns (typically speed related).

- Form fields or form instructions that are unclear, by looking at aggregate edits by many different users.

    Information processing systems on the receiving party's side could then markup received form entries, by flagging them as suspect based on t...