Browse Prior Art Database

A mining-based method to automatically generate action rules in Email System Disclosure Number: IPCOM000233969D
Publication Date: 2014-Jan-06
Document File: 5 page(s) / 169K

Publishing Venue

The Prior Art Database


This invention promotes a method to use text mining of historical email subjects/contents, and combine with contacts’ relationship from social network, to generate action rules automatically in email system. The user does not need to create the action rule manually. The end user can easily select the suggested action rules to organize the mail. For example, to mark the important mail, even though the sender does not mark his mail as "important", the mining model can recognize the mail base on the mail thread content. This invention is much easier, also it is more accurate.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 50% of the total text.

Page 01 of 5

A mining

A mining-

In the email system, user usually define rules to let email client/server help them to do some actions, such as "move to folder", "delegate email to others", and so on. Basically, these action rules are created and defined by user, manually and experimentally. Sometimes, user cannot know accurate relationships of email data, such as sender, receiver, categories of mail subject and etc, to identify how to define action rules. For instance, if user wants someone can be delegation to handle something when he is on vacation, he needs to know how to deal with upcoming emails, and who could be proper delegation. If user wants to move similar email to one folder, he needs to know what kinds of mails have same character. There are lots scenarios as well.

This invention promotes a method to use data mining to combine historical email and contacts' relationship to generate action rules automatically, and then user can select action rules on demand.

Mining historical email data, such as sender, receiver, subject, important flag and etc, can help user know what categories and keywords are high frequency contents, and user can also know who will be related people to these high frequency contents.

Meanwhile, this invention also combines historical email data with social network information and email system contacts to analyst the relationship of the people. After combining the people relationship, email system can determine more relationship by people connection. For example, email system has local contacts and it also can get the user relationship from social networks. After analyzing these relationships, email system can identify mail by consuming mining result. For example, customer's mails, customer support's mails or manager's mails can be recognized as important mail even though there isn't important flag on mail.

The data mining model has chronergy as well. So the model should be refreshed sometimes. For example, model refresh setting can be configured. The cycle could be biweekly or monthly, just like Archive.


--based method to automatically generate action rules in Email System

based method to automatically generate action rules in Email System

Page 02 of 5



Figure 1 indicates the topological graph, user uses PC or laptop as Mail Client to consume rules. Mail Client has synchronous local contacts with Mail Servers. Meanwhile, Mail Servers can get more data of people relationship from social network by API or something else. Mail Servers will combine these relationships as one of data sources. Another data source is the historical emails, these emails contain lots of information, such as sender, receiver (To, CC, BCC), subject, delivery option flag, response time, and etc. There are several pre-defined rule models on Mail Servers. When user wants to generate rules, Mail Servers will use these rule models to respond according to action type.


111 Topological Graph

Topological Graph

Page 03 of 5