Browse Prior Art Database

Automatic email mis-sending detection and correction

IP.com Disclosure Number: IPCOM000241471D
Publication Date: 2015-May-04
Document File: 1 page(s) / 30K

Publishing Venue

The IP.com Prior Art Database

Abstract

Many email systems use a person-name-like string such as "John Smith/Beijing Dev" to refer the email address instead of the raw email address string. This design provides users a convenient way of associating the email receivers and senders with the input strings. However, the problem raises since the user may unintentionally send emails to a wrong receivers due to the similar person-name-like strings for two persons. This disclosure address this problem by analyzing the co-receiver list, thus detects and corrects potential mis-sending receivers.

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

Page 01 of 1

Automatic email mis

Automatic email mis-

The key observations are: an email is wrongly sent to unwanted persons by mistake
1) Many email tools and systems allow users to input a person name style string instead of the exact email address, to make more user-friendly
2) For such name string, many people's ID (idendity) are similar and prone to type the wrong receiver by mistake
3) It is often the case people use the wrong ID for sending email when type into the ID
As a solution, this disclosure automatically detect (warn) and optionally semi-automatically correct the susceptible IDs by generating an overall score which leverages
1) Hierarchical and structural organization, social network information (co-community score)

2) Cross-checking with similar names among relevant groups (mix-up evidence score).

3) Explore the historical topic/content sensitive activities (engagement score).

4) Content/topic/style based analysis for receiver inference (content relevance score)

The approach details are as follows:
1) Scoring and detecting mis-receiver notes ID;

1.1) Compute the distance D1 for current sample distances to the formed teams from co-receivers, and find the smallest distance as the co-community score;
1.2) The distance can be defined as the topological distance between two business units, or semantic distance by considering the business overlap or collaboration, which can also be adaptively learned;

2) Scoring the potential replacer given the detected potential mis-receivers
2.1) Compute the distance D2 for current dete...