Opinion search engine

Patent metadata

Field	Value
Publication number	US-9443245-B2
Application number	US-56870209-A
Country	US
Kind code	B2
Filing date	Sep 29, 2009
Priority date	Sep 29, 2009
Publication date	Sep 13, 2016
Grant date	Sep 13, 2016

Abstract

A computer-readable storage medium having stored thereon computer-executable instructions which, when executed by a computer, cause the computer to implement an opinion search engine. The instructions to implement an opinion search engine cause the computer to collect opinion data about one or more objects from the Internet, extract metadata about the opinion data from the opinion data, remove duplicate metadata from the metadata to generate a resulting metadata, categorize the resulting metadata for similar objects according to one or more taxonomies from one or more websites on the Internet and rank the similar objects based on the categorized metadata.

First claim

What is claimed is:

1. A method implemented by a processing unit of a computing system, the method comprising:
collecting opinion data about a user-specified product category of interest from multiple different opinion-providing webpages having different link paths;
extracting metadata about the opinion data from the opinion data;
evaluating the different link paths of the different opinion-providing webpages to determine product category taxonomies of the different opinion-providing webpages;
determining that a first link path of a first opinion-providing webpage uses a first term to refer to a specific product category and a second link path of a second opinion-providing webpage uses a second term to refer to the specific product category;
creating a product category taxonomy of hierarchical product categories that includes the specific product category;
categorizing the metadata according to the hierarchical categories of the product category taxonomy; and
ranking different products discussed in the opinion data based at least on the categorized metadata.

2. A method implemented by a processing unit of a computing system, the method comprising:
collecting opinion data from one or more networks, the opinion data reflecting opinions about a product; and
determining a brand name of the product using a finite state machine comprising multiple state sequences reflecting product names, wherein the determining the brand name further comprises:
creating the finite state machine such that the finite state machine accepts other products but not the product,
finding modifications to the finite state machine that cause the finite state machine to accept the product, and
updating the finite state machine with the modifications such that the finite state machine accepts the product.

3. The method of claim 2 , further comprising:
finding a minimal set of modifications to the finite state machine that will accept the product,
the updating comprising applying the minimal set of modifications to the finite state machine.

4. A computer system, comprising:
at least one processing unit; and
at least one memory comprising program instructions executable by the at least one processing unit, wherein the program instructions, when executed by the at least one processing unit, cause the at least one processing unit to:
receive, from a user, a user-specified product category of interest;
collect opinion data that discusses multiple products from the user-specified product category of interest from multiple different opinion-providing websites having different site maps;
extract opinion metadata from the opinion data, the opinion metadata relating to the multiple products discussed on the multiple different opinion-providing websites;
based at least on the different site maps of the different opinion-providing websites, determine a product category taxonomy for the multiple products;
organize the opinion metadata according to the product category taxonomy; and
evaluate the multiple products based at least on the organized opinion metadata.

5. The system of claim 4 , wherein the multiple different opinion-providing websites include:
a blog,
a forum, and
a newsgroup.

6. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine that a first user of a first opinion-providing website referred to a manufacturer's name and version number of a specific product in an individual first review; and
determine that a second user of a second opinion-providing website also refers to the specific product in an individual second review that does not mention the manufacturer's name and does not mention the version number.

7. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine hotness scores for the multiple products by:
determining a number of query requests relating to individual products made by various other users to an Internet search engine, and
identifying certain web pages that the various users clicked after receiving Internet search results relating to the individual products from the Internet search engine,

wherein the individual products are evaluated by the system based at least on the hotness scores.

8. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine a hotness score for an individual product by:
determining a number of reviews concerning the individual product that were added to an individual opinion-providing website during a certain time period,
determining another number of subsequent reviews concerning the individual product that were added to the individual opinion-providing website during a subsequent time period after the certain time period, and
determining a trend in numbers of reviews for the individual product based on the number of reviews and the another number of subsequent reviews added to the individual opinion-providing website,

wherein the individual product is evaluated by the system based at least on the hotness score.

9. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine a hotness score for an individual product by:
determining a number of reviewers that have reviewed the individual product on an individual opinion-providing website, and
determining demographic information of the reviewers, including at least:
genders of the reviewers,
age of the reviewers, and
location of the reviewers,

wherein the individual product is evaluated by the system based at least on the hotness score.

10. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine a unified opinion score for an individual product that was reviewed on the multiple different opinion-providing websites by:
determining different rating scores given to the individual product by different users of the multiple different opinion-providing websites,
and
combining the different rating scores to determine the unified opinion score,

wherein the individual product is evaluated by the system based at least on the unified opinion score.

11. The system of claim 10 , wherein the program instructions further cause the at least one processing unit to:
determine the unified opinion score based at least on a total number of users that provided reviews about the another individual product.

12. The system of claim 10 , wherein the program instructions further cause the at least one processing unit to:
distinguish between the multiple different users to identify an editor of an individual opinion-providing website, a product specialist that submitted a review to the individual opinion-providing website, and other reviewing users; and
accord different levels of authority to the editor, the product specialist, and the other reviewing users when determining the unified opinion score.

13. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
determine a total number of positive reviews and a total number of negative reviews for an individual product that was reviewed on the multiple different opinion-providing websites; and
and
rank the individual product relative to other products based at least on the total number of positive reviews for the individual product and the total number of negative reviews for the individual product.

14. The system of claim 4 , wherein the program instructions further cause the at least one processing unit to:
receive a

Assignees

• Sun Jian-Tao
• Ni Xiaochuan
• Xu Peng
• Wang Gang
• Tang Ke
• Chen Zheng
• Microsoft Technology Licensing Llc

Inventors

• Sun Jian-Tao
• Ni Xiaochuan
• Xu Peng
• Wang Gang
• Tang Ke
• Chen Zheng

Classifications

• G06F16/38

  Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually · CPC title

• G06F16/355

  Creation or modification of classes or clusters · CPC title

• G06Q30/02 Primary

  Marketing; Price estimation or determination; Fundraising · CPC title

• G06F16/313

  Selection or weighting of terms for indexing · CPC title

• G06F16/9535

  Search customisation based on user profiles and personalisation · CPC title