Ranking Search Results Returned from Multiple Search Engines based on a Semantic Model
Publication Date: 2012-Apr-11
The IP.com Prior Art Database
If we use multiple search engines (each with a different ranking criteria) and combine search results, then a new re-ranking criterion will be needed for the results, in order to have coherent ranking. This invention is a system that ranks, based on a semantic model (a Topic Map), the most relevant documents to a user query, within a set of documents returned from a multiple of search engines. The proposed system uses a document annotator to annotate the text of the set of documents returned from the search engines, with the topics and their associations as represented in a Topic Map. Then a scoring engine uses these annotations, as related to the topics identified in the user query, to compute a novel score that is used to rank the documents with respect to their relevance to this user query.
Page 01 of 6
Ranking Search Results Returned from Multiple Search Engines based on a
The rankings for the documents returned from a search engine in response to a query may not be the same rankings that other people see. This ranking should reflect the user interest within a specific domain. Furthermore, if we use multiple search engines (each with a different ranking criteria) and combine search results, then a new re-ranking criterion will be needed for the documents in order to have coherent ranking for these documents. This re-ranking criterion should be based on a semantic model that represents the user interest in a specific domain,
There exists an extensive prior art related to ranking search results, for example:
- based on personal interests: A search engine may try to re-rank results for a search to a specific searcher based upon past searches and other tracked activity on the web from that person.
- based upon country specific results It's possible that a searcher may wish to see results biased towards sites coming from a specific country. Someone could possibly explicitly choose a preference for a specific country.
- looking at population or audience segmentation information: This method may look at things such as location, other individual demographic information, and information about groups which a searcher is associated with to help rank pages.
- Reordering based upon topic familiarity: A patent filing from Yahoo! that describes one way to do this, allows searchers to use an interface to choose results that are introductory and ones that are advanced, and a few degrees between.
- Re-ranking based upon editorial content: A granted Google patent describes re-ranking of search results based upon whether or not certain pages have been determined to be favored or unfavored.
- Re-ranking based upon conceptually related information including time-based and use-based factors: It involves grouping together concepts, and looking how those change over time and how different people participate in those changes.
The prior art did not include an accurate way that makes use of the semantic
relationship between a user query and the search results. Therefore, we need a new method to re-rank search results from multiple search engines, based upon a semantic model representing certain domain in order to get a ranking that accurately reflects the user interests.
This paper proposes a system that ranks, based on a semantic model, the most relevant documents to a user query, within a set of documents returned from a multiple of search engines.
The proposed system comprises the following components:
- A document annotator
- A scoring engine
- A semantic model, a topic map, representing knowledge pertaining to a certain domain.
(An "Annotator" is a text analytics engine that ultimately annotates the text of a document with additional related information that it discovers in the semantic model).
Page 02 of 6
The proposed system uses the...