Using message sampling to determine the most frequent words in a user mailbox

• US 8,032,537 B2
• Filed: 12/10/2008
• Issued: 10/04/2011
• Est. Priority Date: 12/10/2008
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Wilson, Michael Geoffrey Andrew, Kleewein, James C., Consul, Ashish, Gorti, Suryanarayana M.
• Primary Examiner(s): PHAM, KHANH B
• Application Number:
  US12/332,161
• Publication Number:
  US 20100145943A1
• Time in Patent Office: 1,028 Days
• Field of Search: 707/613, 707/750, 704/246, 704/251, 379/88.14
• US Class Current: 707/750
• CPC Class Codes: G06F 40/44 Statistical methods, e.g. probability models
• CPC Class Codes: G06Q 10/107 Computer-aided management of electronic mailing [e-mailing]

Abstract

A method is presented for generating a list of frequently used words for an email application on a server computer. When a request is received for a word frequency list for emails stored in a user'"'"'s mailbox, a word frequency list is returned if one exists. If the word frequency list does not exist, an asynchronous process is started on the server computer to generate a word frequency list. If the word frequency list exists but it is older than an aging limit, an asynchronous process is started on the server computer to regenerate the word frequency list. The word frequency list is stored in the user'"'"'s mailbox along with a timestamp indicating the date and time that the list was created or updated.

Claims

1. A method for generating a list of frequently used words for an email application on a server computer, the method comprising:

receiving a request on the server computer to provide a list of frequently used word in email message stored in a user'"'"'s mailbox on the server computer;

determining if a word frequency list exists in the user'"'"'s mailbox;

if a word frequency list exist, returning the word frequency list;

if a word frequency list does not exist, starting a asynchronous process on the server computer to generate a word frequency list;

if a word frequency list does exist and the age of the word frequency list is greater than an aging limit, starting the same asynchronous process on the server computer to regenerate the word frequency list;

storing the word frequency list in the user'"'"'s mailbox; and

storing a timestamp in the user'"'"'s mailbox, the time stamp indicating the date and time when the word frequency list was created or updated;

wherein the asynchronous process generates the word frequency list by performing steps comprising;

sampling email messages from one or more random memory blocks on the server computer;

parsing a predetermined number of email messages from the user'"'"'s mailbox from the one or more random memory blocks, the parsing resulting in generating one or more unique words for each email message;

generating a word frequency list for the one or more unique words, the word frequency list providing a count number of the number of emails in the predetermined number of email message in which each unique word is found.

2. The method of claim 1, wherein the sampling method is distributed random sampling.

3. The method of claim 1, wherein a subset of the predetermined email message is obtained and parsed from a first random memory block before a second random memory block is selected.

4. The method of claim 1, wherein a maximum of 10 email messages are parsed from each random memory block obtained.

5. The method of claim 1, wherein the request is from a voice mail transcription application running on the server computer.

6. The method of claim 1, wherein the voice mail transcription application uses the word frequency list to transcribe a voice mail message into text, the word frequency list being used to help distinguish between similarly sounding words in the voice mail message.

7. The method of claim 1, wherein the request is from a voice mail transcription application running on another server computer.

8. The method of claim 1, wherein the request is from an advertising application on the server computer.

9. The method of claim 1, wherein the aging limit is configurable.

10. The method of claim 1, wherein the predetermined number of email messages parsed from the user'"'"'s mailbox is configurable.

11. The method of claim 1, wherein words containing one or more number are not included in the word frequency list.

12. The method of claim 1, wherein words containing one or more non-language characters are not included in the word frequency list.

13. The method of claim 1, wherein words not containing a vowel are not included in the word frequency list.

14. A method for sampling memory on a server computer to generate a word frequency list for an email application, the method comprising:

obtaining a list of one or more server memory blocks that contain email messages for the user'"'"'s mailbox, each of the one or more server memory blocks storing at least one email message for the user;

randomly selecting a first server memory block from the list of one or more server memory blocks;

parsing one or more of the user'"'"'s email messages contained in the first server memory block, the parsing of the one or more user'"'"'s email messages breaking each email message up into unique words, the number of email messages being parsed from the first server memory block being less than or equal to a first predetermined limit;

creating a word frequency list for the email messages parsed, the word frequency list including each unique word obtained during the parsing of the one or more user'"'"'s email messages that meet predetermined criteria for unique words, each unique word being assigned a count in the word frequency list, the count being updated for each parsed email in which the unique word is found;

storing the word frequency list on the server computer; and

if the number of parsed emails is less than a second predetermined limit, randomly selecting a second server memory block from the list of one or more server memory blocks, parsing one or more user'"'"'s email messages contained in the second server memory block, the number of email messages being parsed from the second server memory block being less than or equal to the first predetermined limit, the parsing breaking up each email message into unique words that meet predetermined criteria for unique words, updating the count in the word frequency list for each unique word already included in the word frequency list, the count being updated for each parsed email in which the unique word is found, adding each unique word to the word frequency list that is not currently included in the word frequency list and storing a count for each of these unique words in the word frequency list.

15. The method of claim 14, further including randomly selecting one or more additional server memory blocks from the list of one or more server memory blocks, parsing email messages from the one or more additional server memory blocks until the second predetermined limit of email messages is reached, one or more email messages being parsed from each of the one or more additional server memory blocks up to the first predetermined limit for each additional server memory block;

updating the count of unique words included in the word frequency list for each email message in which the unique word is contained and adding each unique word to the word frequency list that is not already included in the word frequency list and storing a count for each of these unique words.

16. The method of claim 14, wherein the second predetermined limit is configurable.

17. The method of claim 14, wherein the size of the buffer memory corresponds to the size of an I/O read operation for the server computer.

18. The method of claim 14, wherein the predetermined criteria for unique words excludes words containing one or more numbers, containing non-language characters and not containing at least one vowel.

19. A computer-readable storage medium comprising instructions that, when executed by a server computer, cause the server computer to:

receive a request to provide a list of frequently used words in email messages stored in a user'"'"'s mailbox on the server computer;

determine if a word frequency list exists in the user'"'"'s mailbox;

if a word frequency list exists, return the word frequency list;

if a word frequency list does not exist, start an asynchronous process on the server computer to generate a word frequency list;

if a word frequency list does exist and the age of the word frequency list is greater than an aging limit, start the same asynchronous process on the server computer to regenerate the word frequency list;

wherein the asynchronous process that generates the word frequency list includes the steps of;

requesting a list of one or more server memory blocks that contain email messages for the user'"'"'s mailbox, each of the one or more server memory blocks storing at least one email message for the user;

randomly selecting a first server memory block from the list of one or more server memory blocks;

reading into a buffer memory the contents of the first server memory block;

parsing one or more of the email messages from the buffer memory, the parsing of the one or more email messages breaking each email message up into unique words, the number of email messages being parsed from the buffer memory being less than or equal to a first predetermined limit;

creating a word frequency list for the email messages parsed, the word frequency list including each unique word obtained during the parsing of the one or more user'"'"'s email messages that meet predetermined criteria for unique words, each unique word being assigned a count in the word frequency list, the count being updated for each parsed email in which the unique word is found;

storing the frequency count of each unique word on the server computer;

if the number of parsed emails is less than a second predetermined limit, randomly selecting a second server memory block from the list of one or more server memory blocks, reading into a buffer memory the contents of the second server memory block, the number of email messages being parsed from the buffer memory being less than or equal to the first predetermined limit, the parsing breaking up each email message into unique words that meet predetermined criteria for unique words, updating the count in the word frequency list for each unique word already included in the word frequency list, the count being updated for each parsed email in which the unique count is found, adding each unique word to the word frequency list that is not currently included in the word frequency list and storing a count for each of these unique words in the word frequency list; and

if the number of parsed emails is less than the second predetermined limit, randomly selecting one or more additional server memory blocks from the list of one or more server memory blocks, parsing email messages from the one or more additional server memory blocks until the second predetermined limit of email messages is reached, one or more email messages being parsed from each of the one or more additional server memory blocks up to the first predetermined limit for each additional server memory block, the parsing breaking up each email message into unique words that meet predetermined criteria for unique words, updating the count of unique words included in the word frequency list for each email message in which the unique word is contained and adding each unique word to the word frequency list that is not already included in the word frequency list and storing a count for each of these unique words.