Browse Prior Art Database

Intelligent prevention of incorrect selection of SMS recipient

IP.com Disclosure Number: IPCOM000225762D
Publication Date: 2013-Mar-04
Document File: 3 page(s) / 190K

Publishing Venue

The IP.com Prior Art Database

Abstract

It is too easy to mistakenly send a mobile phone text message intended for one recipient, to another recipient instead. This is due to the fact that the sender can mistakenly choose the wrong recipient while scrolling through their many contacts. This can have a range of consequences, from simple miscommunication/embarrassment to litigation. Although there is a potential solutions centered around vector analysis, one of the main challenges encountered when attempting to devise a solution to this problem using this methods, center around performance. Such a comprehensive analysis requires significant processing power, particularly when potential solutions can require large individual word histories to be stored for each contact, which would then need to be analysed before a text could be sent. This is a large processing overhead especially when compared to the lightweight SMS function it would be designed to support. While it would be possible to implement vector-word analysis on a powerful smartphone, or using a data connection to ?the cloud?, this would require a dependency, There are a couple of techniques that could be used to implement word vector analysis with a phone, but they have their disadvantages: A powerful smartphone may have the power to handle the processing in a reasonable time, but this would adversely affect the already limited battery-life of these phones, and potentially impact other applications running simultaneously on the phone. The processing could be offloaded to ?the cloud?, but this would incur a delay, require a data connection, incur some cost for the processing use up (currently limited) data allowances and also impact adversely affect the already limited battery-life of these phones. Of course, as many phones in use are not high-powered date-enabled smartphones, these phones would not be able to use either of these approaches to running word vector analysis. A system is proposed to intelligently analyse significant sections of the communication to determine the intended recipient for the message, thus reducing time and performance.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 48% of the total text.

Page 01 of 3

Intelligent prevention of incorrect selection of SMS recipient

An system is proposed which outlines a method for analysing discrete portions of the text message about to be sent (the beginning and the end of the text message)

    This would employ a classifier algorithm that works at a character level to analyse the first n chars, the last n chars of the message to be sent as well as statistical information on number of characters used in the message for this recipient.

    When the text message recipient is selected a comparison is performed between the first n chars, the last n chars of the message to be sent, against the first n chars, the last n chars of the messages contained in the history of SMS messages (or even the first n chars and last n chars of the SMS messages) for the intended recipient, eg.


1. New message is created

"Hey toots blahblah...........................................blahblah l8rz XXXX"


2. User selects Mr J Smith as the recipient


3. Method compares first n chars in the message (n - set in user preferences) 'Hi toots blahblah' and the last n chars in the message 'blahblah l8rz XXX' to the history of SMS messages for contact J Smith where the first n chars have typically been ' Hello John, blahblahblah' and the last n chars have typically been 'blahblah regards Jim'


4. Method compares the hail and sign of the message to the stored history of hails and sign offs for all contacts to find a match.


4. Method then displays a warning 'please recheck SMS contact - you have selected Mr J Smith, did you mean The Wife'


5. User then selects the presented alternative recipient


6. Method re-checks and confirms the pattern match

7. SMS is sent to recipient


As is shown in the above use case, if there is significant variance between the

comparison characters in the new message and the history messages of the selected recipient, then a warning is shown on the screen informing the user that they may have selecting the incorrect recipient. The user can then choose the correct contact

    This approach has the advantage of being reasonably language-agnostic. For most western languages, this approach should work equally well without any modifications for the user's language. It should be even more effective if the user uses different languages in messages to different people.

    The idea works by implementing an algorithm which is applied to a text message that a user wishes to send.

This algorithm is based on three sets of features:

1. the first n characters of a message - a count should be maintained for the occurrence of each character in each of the first n positions. This would allow us to classify messages according to particular greetings (e.g. "Hi James", "Ello!".

2. the last m characters of a message - a count should be maintained for the occurrence of each character in each of the last m positions. This would allow us to classify messages according to particular endings (e.g. "..... l8rs. XXX", " love
Jim" "reg's Jim",

3. a probabil...