KEYWORD USAGE SCORE BASED ON FREQUENCY IMPULSE AND FREQUENCY WEIGHT

• US 20080301117A1
• Filed: 06/01/2007
• Published: 12/04/2008
• Est. Priority Date: 06/01/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Li, Hua, Chen, Zheng, Wang, Jian, Zeng, Hua-Jun, Hu, Jian, Peng, Han
• Granted Patent: US 7,644,075 B2
• Time in Patent Office: Days
• US Class Current: 707/5
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A method and system for assessing keyword usage based on frequency of usage of the keywords during various periods is provided. A keyword usage measurement system is provided with the frequency of keywords during various periods. The measurement system then calculates a recent usage score for a keyword by combining a frequency impulse score for the keyword with a frequency weight for the keyword. The frequency impulse score for a keyword indicates whether a recent change in the frequency of the keyword has occurred. The frequency weight for a keyword indicates a recent measure of the frequency of the keyword.

Claims

1. A method in a computing device for assessing keyword usage, the method comprising:

providing frequency of keywords during various periods;

calculating a frequency impulse score for a keyword, the frequency impulse score indicating amount of recent change in frequency of the keyword;

calculating a frequency weight for the keyword, the frequency weight indicating a recent measure of the frequency of the keyword; and

combining the frequency impulse score and frequency weight of the keyword to give the recent usage score for the keyword for a recent period.

2. The method of claim 1 wherein a recent period refers to a certain period relative to prior periods.

3. The method of claim 1 wherein a magnitude of a positive recent usage score indicates amount of increase in keyword usage and a magnitude of a negative recent usage score indicates amount of decrease in keyword usage.

4. The method of claim 1 wherein the frequency impulse score is calculated according to the following:

$I_n=\left(f_n-\underset{i=1}{\overset{n-1}{\sum <br><br>}}<br><br>f_i*{\mathrm{\alpha <br><br>}}_i\right)/f_n$

5. The method of claim 1 wherein the frequency impulse score is a difference between a ratio of a recent frequency of the keyword to a recent total frequency for multiple keywords and a ratio of a combination of less recent frequencies of the keyword to less recent total frequencies for multiple keywords.

6. The method of claim 5 wherein the ratio of the combination applies a decay factor based on recency of the period.

7. The method of claim 1 wherein the frequency weight is based on a logarithm of the recent frequency of the keyword.

8. The method of claim 7 wherein the frequency weight is also based on a logarithm of less recent frequencies of the keyword by applying a decay factor based on recency of the period.

9. The method of claim 1 wherein the frequency weight is calculated according to the following:

$W_n=\log \left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>{\mathrm{freq}}_i*{\mathrm{\alpha <br><br>}}_i\right)$

10. The method of claim 1 wherein the recent usage score is calculated according to the following:

$U_n=\left(\left(f_n-\underset{i=1}{\overset{n-1}{\sum <br><br>}}<br><br><br><br>f_i*{\mathrm{\alpha <br><br>}}_i\right)/f_n\right)*\log \left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>{\mathrm{freq}}_i*{\mathrm{\alpha <br><br>}}_i\right)$

11. The method of claim 1 including ranking documents of a search based at least in part on the recent usage scores of terms of the documents.

12. The method of claim 1 including selecting keywords for placement of advertisements based at least in part on the recent usage scores of the keywords.

13. A computer-readable medium encoded with instructions for controlling a computing device to assess keyword usage, by a method comprising:

providing a search log for each of a plurality of periods, a search log of a period indicating search requests submitted to a search engine service during that period, each search request including keywords;

calculating a frequency impulse score for a keyword, the frequency impulse score indicating a change in frequency of the keyword occurring in search requests of a search log for a recent period;

calculating a frequency weight for the keyword, the frequency weight indicating a measure of the frequency of the keyword occurring in search requests of the search log for the recent period; and

multiplying the frequency impulse score and frequency weight of the keyword to give a recent usage score of the keyword.

14. The computer-readable medium of claim 13 wherein the frequency impulse score is a difference between a ratio of a frequency of the keyword occurring in search requests of the search log for the recent period to a total frequency for multiple keywords occurring in search requests of the search log for the recent period and a ratio of a combination of frequencies of the keyword occurring in search requests of search logs for less recent periods to total frequencies for multiple keywords occurring in search requests of search logs for less recent periods.

15. The computer-readable medium of claim 14 wherein the ratio of the combination applies a decay factor based on recency of the search log.

16. The computer-readable medium of claim 13 wherein the frequency weight is based on a logarithm of the frequency of the keyword occurring in search requests of the search log for the recent period and a logarithm of the frequency of the keyword occurring in search requests of search logs for less recent periods by applying a decay factor based on recency.

17. The computer-readable medium of claim 13 wherein the recent usage score is calculated according to the following:

$U_n=\left(\left(f_n-\underset{i=1}{\overset{n-1}{\sum <br><br>}}<br><br><br><br>f_i*{\mathrm{\alpha <br><br>}}_i\right)/f_n\right)*\log \left(\underset{i=1}{\overset{n}{\sum <br><br>}}<br><br>{\mathrm{freq}}_i*{\mathrm{\alpha <br><br>}}_i\right)$

18. A computing system for ranking documents, comprising:

a data store that indicates frequency of keywords;

a component that calculates a recent usage score for a keyword by combining a frequency impulse score for the keyword and a frequency weight of the keyword, the frequency impulse score indicating a recent change in frequency of the keyword and the frequency weight indicating a recent measure of the frequency of the keyword; and

a component that ranks documents of a search result of a search request based at least in part on the recent usage score of the keyword.

19. The computing system of claim 18 wherein the rank of a document is modified based on the recent usage score of the keyword when the document contains a word relating to the keyword.

20. The computing system of claim 19 wherein the rank is modified only when the search request also contains a word relating to the keyword.