Techniques to select and prioritize application of junk email filtering rules

Patent metadata

Field	Value
Publication number	US-10516638-B2
Application number	US-201715835627-A
Country	US
Kind code	B2
Filing date	Dec 8, 2017
Priority date	Jun 29, 2012
Publication date	Dec 24, 2019
Grant date	Dec 24, 2019

Abstract

Techniques to select and prioritize the application of spam filtering rules in a way that reduces processing time may include receiving an email message for a recipient at a spam filter and extracting email characteristics from the message. Global filtering rule statistics and a profile for the recipient may be retrieved. The technique may include selecting a subset of rules from a set of filtering rules according to the email characteristics, the global filtering rule statistics, and/or the recipient characteristics. The subset of rules may be prioritized and applied to the message from highest priority to lowest until a determination of whether the message is spam is reached. Other embodiments are described and claimed.

First claim

The invention claimed is:

1. An apparatus, comprising:
a logic circuit; and
a memory for storing executable instructions, that when executed by the logic circuit cause the apparatus to:
receive email characteristics of a received email message;
retrieve system-wide filtering rule statistics;
retrieve a recipient profile for an intended recipient of the message, wherein the recipient profile comprises a country with which the recipient communicates by email and a language with which the recipient communicates by email;
select a subset of rules from a set of filtering rules according to the email characteristics, the system-wide filtering rule statistics, and the recipient profile;
prioritize the subset of rules according to the system-wide filtering rule statistics; and
apply the prioritized subset of rules to the message until a determination of whether the message is spam is made.

2. The apparatus of claim 1 , wherein the instructions further comprise prioritization logic, that when executed by the logic circuit cause the apparatus to:
prioritize the subset of rules according to the system-wide filtering rule statistics comprising at least one of:
a processing time of a rule;
a weight of a rule;
a score of a rule; or
an effectiveness of a rule.

3. The apparatus of claim 2 , wherein the prioritization logic is configured to assign a higher priority to a rule having a lower processing time, a higher weight, and a higher effectiveness relative to other filtering rules in the subset of rules.

4. The apparatus of claim 1 , wherein the instructions when executed by the logic circuit further cause the apparatus to:
determine a score for a rule when the rule is applied to the message;
accumulate the scores from the applying of the subset of rules to the message into a determination score; and
compare the determination score to a threshold value to make the determination of whether the message is spam.

5. The apparatus of claim 1 , the email characteristics comprising at least one of a language, a country of origin, a message size, a time of day, a recipient, or a category.

6. The apparatus of claim 5 , the recipient profile comprising a category associated with the recipient.

7. The apparatus of claim 1 , wherein the filtering rules comprise a plurality of rule types and subtypes, wherein the subset of rules is selected according to subtypes that are relevant to at least one of the email characteristics, the system-wide filtering rule statistics, or the recipient profile.

8. A computer-implemented method, comprising:
receiving an email message for a recipient at an email server;
extracting email characteristics from the received email;
retrieving filtering rule statistics;
retrieving a recipient profile for the recipient, wherein the recipient profile comprises a country of origin with which the recipient communicates by email and a language preference with which the recipient communicates by email;
selecting a subset of rules from a set of filtering rules according to the email characteristics, the filtering rule statistics, and the recipient characteristics;
prioritizing the subset of rules by assigning a higher priority to a filtering rule based upon information relative to other filtering rules in the set of filtering rules; and
applying the prioritized subset of rules from highest priority to lowest priority until a determination of whether the message is spam is reached.

9. The computer-implemented method of claim 8 , further comprising:
determining a score for a rule when the rule is applied to the message;
accumulating the scores from the applying of the subset of rules to the message into a determination score;
comparing the determination score to a threshold value after each score determination to make the determination of whether the message is spam; and
stopping the application of the subset of rules when a determination is reached that the message is spam.

10. The computer-implemented method of claim 8 , wherein the global filtering rule statistics comprise at least one of: a processing time of a rule; a weight of a rule; a score of a rule; or an effectiveness of a rule; and comprising: assigning a higher priority to a rule having a lower processing time, a higher weight, and a higher effectiveness relative to other filtering rules in the subset of rules.

11. The computer-implemented method of claim 8 , wherein the filtering rules comprise a plurality of rule types and subtypes, and the computer-implemented method further comprising: selecting the subset of rules according to subtypes that are relevant to at least one of the email characteristics, the filtering rule statistics, or the recipient profile.

12. The computer-implemented method of claim 11 , wherein the plurality of rule types comprise at least a languages rule type, a country of origin rule type, a category rule type, and a recipient rule type, wherein a filtering rule is associated with at least one type.

13. The computer-implemented method of claim 8 , further comprising applying a rule by evaluating a regular expression against a string of text in the message.

14. The computer-implemented method of claim 8 , the email characteristics comprising at least one of a language, a country of origin, a message size, a time of day, a recipient, or a category.

15. At least one non-transitory computer-readable storage medium comprising instructions that, when executed by a processor, cause a system to:
extract email characteristics from a received email message;
retrieve system-wide spam rule statistics;
retrieve a recipient profile for an intended recipient of the message, wherein the recipient profile comprises at least one of a country frequency with which the recipient communicates with a country of origin by email and a language frequency with which the recipient communicates in a language by email;
prioritize a set of filtering rules according to the email characteristics, the system-wide spam rule statistics, the recipient profile;
assign a higher priority to a filtering rule based upon at least one of a processing time, a weight, or an effectiveness relative to other filtering rules in the set of filtering rules; and
apply the prioritized set of rules sequentially to the message from highest priority to lowest, until a determination of whether the message is spam is made.

16. The non-transitory computer-readable storage medium of claim 15 , comprising instructions that when executed cause the system to:
select a subset of rules from the prioritized set of rules according to at least one of the email characteristics, the global spam rule statistics, or the recipient characteristics; and
apply the prioritized subset of rules to the email.

17. The non-transitory computer-readable storage medium of claim 16 , wherein the filtering rules comprise a plurality of types and subtypes, the medium comprising instructions that when executed cause the system to:
select the subset of rules according to subtypes that are relevant to at least one of the email characteristics, the system-wide spam rule statistics, or the recipient profile.

18. The non-transitory computer-readable storage medium of claim 15 , comprising instructions that when executed cause the system to:
determine a score for a rule when the rule is applied to the message;
accumulate the scores from the applying of the subset of rules to the message into a determination score; and
compare the determination score to a threshold value to make the determination of whether the message i

Assignees

• Microsoft Technology Licensing Llc

Inventors

• Syrowitz Clinton
• Gandhi Mauktik
• Mishra Ashish
• Sundaram Manivannan
• Zink Terence

Classifications

• G06Q10/107 Primary

  Computer-aided management of electronic mailing [e-mailing] · CPC title

• H04L51/12 Primary

  Electricity · mapped topic

• H04L51/212 Primary

  using filtering or selective blocking · CPC title