System and method for utilizing the content of an online conversation to select advertising content and/or other relevant information for display

• US 7,653,627 B2
• Filed: 05/13/2005
• Issued: 01/26/2010
• Est. Priority Date: 05/13/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Gao, Hongbin, Wang, Xianfang, Chen, Zheng, Najm, Tarek, Li, Li, Zhang, Benyu, Li, Ying, Zeng, Hua-Jun
• Primary Examiner(s): Jalil; Neveen Abel
• Assistant Examiner(s): RADTKE, MARK A
• Application Number:
  US11/128,788
• Publication Number:
  US 20060259473A1
• Time in Patent Office: 1,719 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: Y10S 707/99937 Sorting

Abstract

Systems and methods for analyzing the content of online conversations using data mining technologies are provided. Methods and systems for utilizing the results of data mining technology implementation to retrieve relevant advertising content and/or other relevant information for display in association with the content of an online conversation are also provided.

Claims

1. A computer-implemented method for utilizing content of an online conversation to select advertising content for display, the method comprising:

receiving, via a computing device, the content of the online conversation, wherein the online conversation comprises an instant messaging session;

extracting one or more keywords from the content of the online conversation;

characterizing the one or more keywords that are extracted from the content of the online conversation as one or more of an advertising selection keyword and a user-intention keyword,

wherein an advertising selection keyword includes a keyword that matches a word on an advertising-keyword list, and wherein a user-intention keyword indicates user interest in a topic and includes a keyword that is proximate to one or more of a trigger word and a trigger phrase, which are identified in a pre-defined list;

scoring a keyword that has been characterized as an advertising selection keyword, wherein scoring comprises calculating a measure of frequency of the advertising selection keyword;

determining that the advertising selection keyword, which has been scored, has also been characterized as a user-intention keyword;

re-weighting the advertising selection keyword, which has also been characterized as a user-intention keyword, to be given greater weight in a joint-probability-distribution algorithm that is usable to select advertising content;

selecting the advertising content to be displayed based on the joint-probability-distribution algorithm, which includes the advertising selection keyword that has been both scored and re-weighted; and

outputting in association with the content the advertising content that is selected.

2. The method of claim 1, further comprising displaying the advertising content selected in association with the content of the online conversation.

3. The method of claim 1, further comprising receiving one or more of a user profile and information regarding user behavior, wherein selecting the advertising content to be displayed comprises selecting the advertising content to be displayed based upon at least one of the one or more extracted keywords, the user profile, the information regarding user behavior, an historic click-through rate, and a monetization value.

4. The method of claim 1, wherein the joint-probability-distribution algorithm is represented by:

P(C,I,R_V,R_U|A,W,V,U)=P(C|I,A,U).P(I|R_V,R_U).P(R_U|W,U).P(R_V|W,V)

wherein P(R_V|W,V) represents a probability that an ad-word W is relevant to the online conversation V and is based on the advertising selection keyword that has been re-weighted;

wherein P(R_U|W,V) represents a probability that a user has a general interest in the advertising selection keyword and is based on one or more of a user profile and historical user behavior; and

wherein P(I|R_V,R_U)represents a probability that the user is interested in the advertising content and is based on one or more of a historical interest (R_V) of the user and a momentary interest of the user (R_U).

5. The method of claim 1, further comprising:

retrieving historical-keyword information;

calculating a mixed topic relevance score between the advertising selection keyword and the historical-keyword information; and

comparing the mixed topic relevance score to a threshold value to determine that a topic change has occurred.

6. The method of claim 5, further comprising re-weighting the advertising selection keyword based upon historical data.

7. One or more computer-readable storage media having computer-executable instructions for performing a method for utilizing content of an online conversation to select advertising content for display, the method comprising:

receiving the content of the online conversation, wherein the online conversation comprises an instant messaging session;

extracting a keyword from the content of the online conversation, wherein the keyword is an advertising selection keyword that matches a word on an advertising-keyword list;

scoring the advertising selection keyword tat has been extracted by calculating a measure of frequency;

determining that the advertising selection keyword includes a user intention keyword;

re-weighting the advertising selection keyword to be given greater weight in a joint-probability-distribution algorithm that is usable to select advertising content;

selecting the advertising content to be presented based upon the keyword that has been scored and re-weighted; and

displaying the advertising content in association with the content of the online conversation.

8. The method of claim 7, further comprising receiving one or more of a user profile and information regarding user behavior, wherein selecting the relevant information for display comprises selecting the relevant information for display based upon at least one of the keyword, the user profile, and the information regarding user behavior.

9. The method of claim 7, further comprising comparing the keyword to one or more relevant-information selection keywords, wherein selecting the relevant information for display comprises selecting the relevant information for display based upon the comparison of the keyword to the one or more relevant-information selection keywords.

10. The method of claim 7, further comprising determining whether a topic change has occurred.

11. The method of claim 10, wherein if it is determined that a topic change has occurred, the method further comprises re-weighting the one or more scored keywords based upon historical data.

12. One or more computer-readable storage media having computer-executable instructions for performing a method for utilizing content of an online conversation to select advertising content to be displayed, the method comprising:

receiving current content of an online conversation, wherein the online conversation comprises an instant messaging session;

extracting one or more keywords from the current content of the online conversation;

retrieving historical-keyword information of the online conversation, the historical-keyword information being provided by a user prior to the current content;

calculating a mixed topic relevance score between the one or more keywords from the current content and the historical-keyword information, wherein the mixed topic relevance score is usable to detect a change in a topic between the current content and the historical-keyword information;

determining whether the mixed topic relevance score is either above or below a threshold value,

wherein if the mixed topic relevance score is above the threshold value, no change is deemed to have occurred and both the one or more keywords from the current content and the historical-keyword information are used to select at least one of advertising content and other relevant information, and

wherein if the mixed topic relevance score is below the threshold value, a change is deemed to have occurred and the one or more keywords from the current content are used to select at least one of advertising content and other relevant information and the historical-keyword information is not used;

characterizing a keyword of the one or more keywords from the current content as an advertising selection keyword, which includes a keyword that matches a word on an advertising-keyword list;

scoring the keyword that has been characterized as an advertising selection keyword, wherein scoring comprises calculating a measure of frequency of the advertising selection keyword;

determining that the advertising selection keyword, which has been scored, has also been characterized as a user-intention keyword;

re-weighting the advertising selection keyword, which has also been characterized as a user-intention keyword, to be given greater weight in a joint-probability-distribution algorithm that is usable to select advertising content;

selecting the advertising content to be displayed based which includes the advertising selection keyword that has been both scored and re-weighted; and

retrieving the advertising content to be displayed based on the joint-probability-distribution algorithm; and

displaying advertising content that was retrieved.