Comparative web search
First Claim
1. A computer-implemented method performed by a computer for ranking comparative web results comprising the steps of:
- receiving a first search query;
receiving a second search query that is related to the first search query;
processing the first search query to produce first search results, the first search results having elements that are relevant to the first query, and processing the second search query to produce second search results, the second search results having elements that are relevant to the second query; and
generating by the computer a plurality of comparative search result groups that each include an element from the first search results and an element from the second search results bycalculating a score for pairs of elements with a first element of the pair being from the first search results and a second element of the pair being from the second search results, each score for a pair indicating likeliness that the elements form a comparative pair corresponding to the first and second queries, and being based on similarity between content of the elements of the pair, relevance of the first element to the first query, and relevance of the second element to the second query, the score being calculated according to the following equation;
fq1q2(p1,p2)=α
·
R(p1,q1)+β
·
R(p2,q2)+γ
·
S(p1\q1,p2\q2), p1∈
SR1,p2∈
SR2wherein SR1 represents the first search results for the first search query q1, SR2 represents the second search results for the second search query q2, p1 represents an element from SR1, p2 represents an element from SR2, R represents relevance of an element to a query, S represents similarity between two elements, p1\q1 represents text content of p1 after removing terms of q1, p2\q2 represents text content of p2 after removing terms of q2, and α
, β
, and γ
represents constants; and
repeatedly selecting in score order a previously unselected pair of elements neither of whose elements are currently in a comparative search result group and creating a comparative search result group with the elements of the selected pair.
3 Assignments
0 Petitions
Accused Products
Abstract
Methods and systems are provided for performing a comparative search. In one example, the comparative search is performed over a network, such as the web, or a database. In one exemplary implementation, a user transmits a plurality of queries which represent the topics that a user wants to compare, and a computing system can automatically retrieve and rank web pages or documents based on both their relevance to queries and the comparative contents they contain. In one such example, the comparative pages are displayed in a pair or other form of a grouping. In another example, comparative results having similar contents may be clustered into meaningful themes.
29 Citations
16 Claims
-
1. A computer-implemented method performed by a computer for ranking comparative web results comprising the steps of:
-
receiving a first search query; receiving a second search query that is related to the first search query;
processing the first search query to produce first search results, the first search results having elements that are relevant to the first query, and processing the second search query to produce second search results, the second search results having elements that are relevant to the second query; andgenerating by the computer a plurality of comparative search result groups that each include an element from the first search results and an element from the second search results by calculating a score for pairs of elements with a first element of the pair being from the first search results and a second element of the pair being from the second search results, each score for a pair indicating likeliness that the elements form a comparative pair corresponding to the first and second queries, and being based on similarity between content of the elements of the pair, relevance of the first element to the first query, and relevance of the second element to the second query, the score being calculated according to the following equation;
fq1q2(p1,p2)=α
·
R(p1,q1)+β
·
R(p2,q2)+γ
·
S(p1\q1,p2\q2), p1∈
SR1,p2∈
SR2wherein SR1 represents the first search results for the first search query q1, SR2 represents the second search results for the second search query q2, p1 represents an element from SR1, p2 represents an element from SR2, R represents relevance of an element to a query, S represents similarity between two elements, p1\q1 represents text content of p1 after removing terms of q1, p2\q2 represents text content of p2 after removing terms of q2, and α
, β
, and γ
represents constants; andrepeatedly selecting in score order a previously unselected pair of elements neither of whose elements are currently in a comparative search result group and creating a comparative search result group with the elements of the selected pair. - View Dependent Claims (2, 3, 4, 5, 6, 7)
-
-
8. A computer-implemented method for obtaining comparative web results from a computing device operatively connected to a network through a browser comprising the steps of:
-
transmitting a first search query to obtain a first search result; transmitting a second search query that is related to the first search query to obtain a second search result; receiving at the computing results the plurality of comparative search result groups that each include a first element from the first search result and a second element from the second search result, the plurality of the comparative search result groups being generated based on the following equation;
fq1q2(p1,p2)=α
·
R(p1,q1)+β
·
R(p2,q2)+γ
·
S(p1\q1,p2\q2), p1∈
SR1,p2∈
SR2
wherein SR1 represents the first search results for the first search query q1, SR2 represents the second search results for the second search query q2, p1 represents an element from SR1, p2 represents an element from SR2, R represents relevance of an element to a query, S represents similarity between two elements, p1\q1 represents text content of p1 after removing terms of p1, p2\q2 represents text content of p2 after removing terms of q2, and α
, β
, and γ
represents constantswherein the elements are grouped based on relevance of the first element to the first search query and relevance of the second element to the second search query. - View Dependent Claims (9, 10, 11, 12)
-
-
13. A computing device for ranking comparative web results, the device having a processor, and a computer-readable medium, the computer-readable medium comprising computer-readable instructions that when executed perform the method of:
-
receiving a first search query; receiving a second search query that is related to the first search query; processing the first search query to produce first search results and processing the second search query to produce second search results; and generating a plurality of comparative search result groups that each include an element from the first search results and an element from the second search results based on the following equation;
fq1q2(p1,p2)=α
·
R(p1,q2)+β
·
R(p2,q2)+γ
·
S(p1\q1,p2\q2), p1∈
SR1,p2∈
SR2
wherein SR1 represents the first search results for the first search query q1, SR2 represents the second search results for the second search query q2, p1 represents an element from SR1, p2 represents an element from SR2, R represents relevance of an element to a query, S represents similarity between two elements, p1\q1 represents text content of p1 after removing terms of q1, p2\q2 represents text content of p2 after removing terms of q2, and α
, β
, and α
represents constants;wherein the elements are grouped based on similarity between the content of the elements. - View Dependent Claims (14, 15, 16)
-
Specification