Outgoing message monitor

• US 8,375,052 B2
• Filed: 10/03/2007
• Issued: 02/12/2013
• Est. Priority Date: 10/03/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Bordeaux, Lucas, Hamadi, Youssef, Izadi, Shahram, Vinay, Vishwa
• Primary Examiner(s): GORTAYO, DANGELINO N
• Application Number:
  US11/866,637
• Publication Number:
  US 20090094240A1
• Time in Patent Office: 1,959 Days
• Field of Search: 707/1, 707/3, 707/6, 707/7, 707/100, 707/104.1, 707/726, 707/737, 707/749, 707/752, 707/771, 707/802, 707/769, 707/776, 707/780, 709/206
• US Class Current: 707/780
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]

Abstract

An outgoing message monitor is provided. In an embodiment, outgoing messages are monitored to detect potential errors and alerts may be triggered. Using information about a message such as an email, a first classifier classifies the email into an expected class and a second classifier classifies the email into an actual class. On the basis of a comparison of the expected and actual classes an alert may be triggered. In an embodiment, the second classifier uses information derived from text content of the email which may optionally be pre-processed. The first classifier, for example, uses other information about the email such as its intended recipients, information about the presence of attachments, information about whether the email is part of a thread and other information.

Claims

1. A computer-implemented method comprising:

receiving information about a proposed outgoing message stored in memory, the received information comprising text content of a body portion of the proposed outgoing message and metadata of the proposed outgoing message;

using a first classifier to classify the proposed outgoing message into one of a plurality of specified classes, to obtain an expected class of the proposed outgoing message based on the metadata of the proposed outgoing message and not the text content of the body portion of the proposed outgoing message, the first classifier being a multi-way classifier comprising N*(N-1)/2 sub-classifiers where N is the number of the plurality of specified classes and each sub-classifier provides a one-to-one classification determination between two specified classes, the metadata comprising one or more of;

information on whether the proposed outgoing message is part of an email thread;

information associated with a file attached to the proposed outgoing message;

information in a subject line of the proposed outgoing message;

information on voting buttons associated with the proposed outgoing message; and

information on one or more required receipts associated with the proposed outgoing message;

using a second classifier different from the first classifier to obtain an actual class of the proposed outgoing message based on the text content of the body portion of the proposed outgoing message and not the metadata of the proposed outgoing message;

comparing the expected class of the proposed outgoing message based on the metadata and the actual class of the proposed outgoing message based on the text content of the body portion; and

triggering an alert in response to the comparison failing to yield a match.

2. The computer-implemented method as claimed in claim 1 further comprising pre-processing the text content of the body portion of the proposed outgoing message and inputting results of the pre-processing into the second classifier.

3. The computer-implemented method as claimed in claim 1 further comprising-receiving feedback about the proposed outgoing message and using the feedback to update at least one of the first or second classifiers.

4. The computer-implemented method as claimed in claim 3 further comprising using a machine learning process to update the at least one of the first or second classifier.

5. The computer-implemented method as claimed in claim 1 wherein using the first classifier further comprises obtaining the expected class together with a likelihood estimate associated with the expected class.

6. The computer-implemented method as claimed in claim 1 wherein using the second classifier further comprises obtaining the actual class together with a likelihood estimate associated with the actual class.

7. The computer-implemented method as claimed in claim 1 wherein the comparing comprises using a specified threshold.

8. A hardware computer-readable storage medium storing instructions which, when executed, cause one or more processors to perform actions, the actions comprising:

receiving information about a proposed outgoing message stored in memory, the received information comprising text content of a body portion of the proposed outgoing message and metadata of the proposed outgoing message;

using a first classifier to classify the proposed outgoing message into one of a plurality of specified classes, to obtain an expected class of the proposed outgoing message based on the metadata of the proposed outgoing message and not the text content of the body portion of the proposed outgoing message, the first classifier being a multi-way classifier comprising N*(N-1)/2 sub-classifiers where N is the number of the plurality of specified classes and each sub-classifier provides a one-to-one classification determination between two specified classes, the metadata comprising one or more of;

information on whether the proposed outgoing message is part of an email thread;

information associated with a file attached to the proposed outgoing message;

information in a subject line of the proposed outgoing message;

information on voting buttons associated with the proposed outgoing message; and

information on one or more required receipts associated with the proposed outgoing message;

using a second classifier different from the first classifier to obtain an actual class of the proposed outgoing message based on the text content of the body portion of the proposed outgoing message and not the metadata of the proposed outgoing message;

comparing the expected class of the proposed outgoing message based on the metadata and the actual class of the proposed outgoing message based on the text content of the body portion; and

triggering an alert in response to the comparison failing to yield a match.

9. The hardware computer-readable storage medium as claimed in claim 8 further comprising pre-processing the text content of the body portion of the proposed outgoing message and inputting results of the pre-processing into the second classifier.

10. The hardware computer-readable storage medium as claimed in claim 8 further comprising receiving feedback about the proposed outgoing message and using the feedback to update at least one of the first or second classifiers.

11. The hardware computer-readable storage medium as claimed in claim 10 further comprising using a machine learning process to update the at least one of the first or second classifier.

12. The hardware computer-readable storage medium as claimed in claim 8 wherein using the first classifier further comprises obtaining the expected class together with a likelihood estimate associated with the expected class.

13. The hardware computer-readable storage medium as claimed in claim 8 wherein using the second classifier further comprises obtaining the actual class together with a likelihood estimate associated with the actual class.

14. The hardware computer-readable storage medium as claimed in claim 8 wherein the comparing comprises using a specified threshold.

15. A apparatus comprising:

one or more processors;

a receiver configured to receive information about a proposed outgoing message stored in memory, the received information comprising text content of a body portion of the proposed outgoing message and metadata of the proposed outgoing message;

a first classifier configured to classify the proposed outgoing message into one of a plurality of specified classes, to obtain an expected class of the proposed outgoing message based on the metadata of the proposed outgoing message and not the text content of the body portion of the proposed outgoing message, the first classifier being a multi-way classifier comprising N*(N-1)/2 sub-classifiers where N is the number of the plurality of specified classes and each sub-classifier provides a one-to-one classification determination between two specified classes, the metadata comprising one or more of;

information on whether the proposed outgoing message is part of an email thread;

information associated with a file attached to the proposed outgoing message;

information in a subject line of the proposed outgoing message;

information on voting buttons associated with the proposed outgoing message; and

information on one or more required receipts associated with the proposed outgoing message;

a second classifier different from the first classifier configured to obtain an actual class of the proposed outgoing message based on the text content of the body portion of the proposed outgoing message and not the metadata of the proposed outgoing message;

a comparing unit configured to compare the expected class of the proposed outgoing message based on the metadata and the actual class of the proposed outgoing message based on the text content of the body portion; and

a triggering unit configured to trigger an alert in response to the comparison failing to yield a match.

16. The apparatus as claimed in claim 15 further comprising a pre-processing unit configured to pre-process the text content of the body portion of the proposed outgoing message and inputting results of the pre-processing into the second classifier.

17. The apparatus as claimed in claim 15 further comprising a feedback receiver configured to receive feedback about the proposed outgoing message and an updater configured to use the feedback to update at least one of the first or second classifiers.

18. The apparatus as claimed in claim 17 further comprising a machine learning unit configured to update the at least one of the first or second classifier.

19. The apparatus as claimed in claim 15 wherein the first classifier is further configured to obtain the expected class together with a likelihood estimate associated with the expected class.

20. The apparatus as claimed in claim 15 wherein the second classifier is further configured to obtain the actual class together with a likelihood estimate associated with the actual class.