Methods and systems for analyzing email messages

• US 8,285,806 B2
• Filed: 09/23/2011
• Issued: 10/09/2012
• Est. Priority Date: 01/16/2004
• Status: Expired due to Fees

• Current Assignee: Gozoom.com, Inc.
• Original Assignee: Gozoom.com, Inc.
• Inventors: Yu, Tonny
• Primary Examiner(s): Park, Jeong S
• Application Number:
  US13/243,827
• Publication Number:
  US 20120016945A1
• Time in Patent Office: 382 Days
• Field of Search: 709/206, 709/207
• US Class Current: 709/206
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]
• CPC Class Codes: H04L 51/212 using filtering or selective blocking

Abstract

Methods and systems for analyzing email messages are provided. In an embodiment, a method includes determining whether a received email message is considered to be the same as one or more prior email messages already classified into one of a plurality of different categories. If the received message is not considered to be the same as one of the prior email messages already classified into one of the categories, then the received message is classified into one of the categories by performing a set of computations. If the received message is considered to be the same as one of the prior email messages already classified into one of the categories, then the received message is classified into one of the categories without performing the computations, by classifying the received message into the same category as a prior email message that the received message is considered to be the same as.

Claims

1. A method for analyzing email messages, comprising:

(a) for each of a plurality of received email messages, using a plurality of modules to each independently determine a module output indicative of a level of sameness of the received email message with one or more prior email messages, and determining an overall level of sameness of the received email message with one or more prior email messages by combining the module outputs of at least two of the plurality of modules using non-zero weights assigned to the modules in dependence on performance levels of the modules;

(b) classifying each of the plurality of received email messages into one of a plurality of different categories by

(b.1) determining, based on the overall level of sameness determined for the received email message, whether the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories;

(b.2) if the received email message is not considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories, then performing a set of computations to thereby classify the received email message into one of the categories; and

(b.3) if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories, then classifying the received email message into the same category as the prior email message that the received email message is considered to be the same as, without performing the set of computations for the received email message;

wherein steps (a) and/or (b) are performed using at least one processor.

2. The method of claim 1, wherein at step (b.3), if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories, then the received email message is classified into the same category as the prior email message that the received email message is considered to be the same as, without performing the set of computations for the received email message, by re-using the set of computations and corresponding results for the prior email message that the received email message is considered to be the same as.

3. The method of claim 1, wherein:

step (b.3) includes, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as spam, then classifying the received email message as spam, without performing the set of computations for the received email message.

4. The method of claim 1, wherein:

step (b.3) includes, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as legitimate, then classifying the received email message as legitimate, without performing the set of computations for the received email message.

5. The method of claim 1, further comprising:

monitoring performance levels of the modules; and

adjusting the non-zero weights of at least two of the modules based on their performance levels.

6. The method of claim 5, wherein the adjusting the non-zero weights of at least two of the modules based on their performance levels is performed using Bayesian analysis.

7. A system for analyzing email messages, comprising:

a plurality of modules each configured to independently determine a level of sameness of each of a plurality of received email messages with one or more prior email messages; and

one or more servers configured to classify each of the plurality of received email messages into one of a plurality of different categories;

wherein, for each of the plurality of received email messages, the one or more servers is/are configured to

determine an overall level of sameness of the received email message with one or more prior email messages by combining the module outputs of at least two of the plurality of modules using non-zero weights assigned to the modules in dependence on performance levels of the modules;

determine, based on the overall level of sameness determined for the received email message, whether the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories;

perform a set of computations to thereby classify the received email message into one of the categories, if the received email message is not considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories; and

classify the received email message into the same category as the prior email message that the received email message is considered to be the same as, without performing the set of computations for the received email message, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories.

8. The system of claim 7, wherein, for each of the plurality of received email messages, the one or more servers is/are configured to

classify the received email message into the same category as the prior email message that the received email message is considered to be the same as, without performing the set of computations for the received email message, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories, by re-using the set of computations and corresponding results for the prior email message that the received email message is considered to be the same as.

9. The system of claim 7, wherein, for each of the plurality of received email messages, the one or more servers is/are configured to

classify the received email message as spam, without performing the set of computations for the received email message, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as spam.

10. The system of claim 7, wherein, for each of the plurality of received email messages, the one or more servers is/are configured to

classify the received email message as legitimate, without performing the set of computations for the received email message, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as legitimate.

11. The system of claim 7, wherein for each of the plurality of received email messages the one or more servers is/are configured to

monitor performance levels of the modules;

and

adjust the non-zero weights of at least two of the modules based on their performance levels.

12. The system of claim 11, wherein the one or more servers is/are configured to adjust the non-zero weights of at least two of the modules based on their performance levels using Bayesian analysis.

13. A system for analyzing email messages, comprising:

a plurality of modules each configured to independently determine a level of sameness of a received email message with one or more prior email messages;

one or more servers configured to

determine an overall level of sameness of the received email message with one or more prior email messages by combining the outputs of at least two of the plurality of sameness determining means using non-zero weights assigned to the sameness determining means in dependence on performance levels of the sameness determining means;

determine based on the overall level of sameness determined for a received email message, whether the received email message is considered to be the same as one or more prior email messages that has already been classified into one of a plurality of different categories;

classify a received email message into one of the categories by performing a set of computations, if the received email message is not considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories; and

classify a received email message into one of the categories without performing the set of computations, by classifying the received email message into the same category as a prior email message that the received email message is considered to be the same as, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories.

14. The system of claim 13, wherein the one or more servers is/are configured to classify a received email message into one of the categories without performing the set of computations by reusing the set of computations and corresponding results for the prior email message that the received email message is considered to be the same as.

15. The system of claim 13, wherein the one or more servers is/are configured to classify a received email message into one of the categories without performing the set of computations by classifying a received email message as spam if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as spam.

16. The system of claim 13, wherein the one or more servers is/are configured to classify a received email message into one of the categories without performing the set of computations by classifying a received email message as legitimate if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as legitimate.

17. The system of claim 13, the one or more servers is/are configured to

monitor performance levels of the modules ;

and

adjust the non-zero weights of at least two of the modules based on their performance levels.

18. The system of claim 17, wherein the one or more servers is/are configured to adjust the non-zero weights of at least two of the modules based on their performance levels uses Bayesian analysis.

19. A method for analyzing email messages, comprising:

(a) using a plurality of modules to each independently determine a module output indicative of a level of sameness of a received email message with one or more prior email messages;

(b) determining an overall level of sameness of the received email message with one or more prior email messages by combining the module outputs of at least two of the plurality of modules using non-zero weights assigned to the modules in dependence on performance levels of the modules;

(c) determining, based on the overall level of sameness determined for the received email message, whether the received email message is considered to be the same as one or more prior email messages that has already been classified into one of a plurality of different categories;

(d) classifying the received email message into one of the categories by performing a set of computations, if the received email message is not considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories; and

(e) classifying the received email message into one of the categories without performing the set of computations, by classifying the received email message into the same category as a prior email message that the received email message is considered to be the same as, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified into one of the categories;

wherein one or more of steps (a), (b), (c), (d) and (e) is/are performed using at least one processor.

20. The method of claim 19, wherein step (e) comprises classifying the received email message into one of the categories without performing the set of computations by re-using the set of computations and corresponding results for the prior email message that the received email message is considered to be the same as.

21. The method of claim 19, wherein step (e) comprises classifying the received email message as spam, without performing the set of computations, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as spam.

22. The method of claim 19, wherein step (e) comprises classifying the received email message as legitimate, without performing the set of computations, if the received email message is considered to be the same as one of the one or more prior email messages that has already been classified as legitimate.

23. The method of claim 19, further comprising:

monitoring performance levels of the modules; and

adjusting the non-zero weights of at least two of the modules based on their performance levels.

24. The method of claim 23, wherein the adjusting the non-zero weights of at least two of the modules based on their performance levels is performed using Bayesian analysis.