Method for improving quality of search results by avoiding indexing sections of pages
A method and apparatus for improving search results is provided. The method works by delineating sections of a document that are not relevant to the main content. The document content is subjected to ranking analysis in entirety. In response to a query results are recalled omitting terms included in the no-recall sections. Terms in the no-recall sections are not used in titles and abstracts of the results. The results are ordered at least in part by the rankings attributed to the identified no-recall sections.
1. A method, comprising:
ranking a plurality of documents recalled by a search engine for a query;
wherein the plurality of documents contain certain documents, each document of said certain documents containing at least one section that is not used by said search engine for recall and one or more sections that are used by said search engine for recall;
wherein ranking a plurality of documents includes ranking said plurality of documents based, at least in part, on the at least one section of said certain documents not used by said search engine to recall documents; and
wherein the method is performed by one or more computing devices.
2. The method of claim 1 , wherein said at least one section of said certain documents not used by said search engine to recall documents are demarcated by a tag of an element having a name from a set of one or more names.
3. The method of claim 1 , wherein at least one document of said certain documents does not comprise data that contains element tags.
4. The method of claim 1 , further including said search engine generating data representing results of said query, wherein:
said data includes an abstract describing each document of said plurality of documents; and
for each respective abstract of each document of said certain documents, said abstract excludes terms from the respective at least one section not used by said search engine to recall said each document.
5. The method of claim 1 , said search engine using a search engine index to recall said plurality of documents; and wherein said search engine index does not index any term in said certain documents that is found only in one or more sections of said certain documents established as a no-recall section.
6. The method of claim 1 , wherein ranking a plurality of documents recalled by a search engine for a query depends in part on a vector space model associated with each of said documents.
7. The method of claim 1 , wherein ranking a plurality of documents recalled by a search engine for a query depends in part on a language model associated with each of said documents.
8. A computer-readable storage medium that stores instructions which, when executed by one or more processors, cause performance of:
ranking a plurality of documents recalled by a search engine for a query;
wherein the plurality of documents contain certain documents, each document of said certain documents containing at least one section that is not used by said search engine for recall and one or more sections that are used by said search engine for recall; and
wherein ranking a plurality of documents includes ranking said plurality of documents based, at least in part, on the at least one section of said certain documents not used by said search engine to recall documents.
9. The computer-readable storage medium of claim 8 wherein said at least one section of said certain documents not used by said search engine to recall documents are demarcated by a tag of an element having a name from a set of one or more names.
10. The computer-readable storage medium of claim 8 wherein at least one document of said certain documents does not comprise data that contains element tags.
11. The computer-readable storage medium of claim 8 further comprising instructions for including said search engine generating data representing results of said query, wherein:
said data includes an abstract describing each document of said plurality of documents; and
for each respective abstract of each document of said certain documents, said abstract excludes terms from the respective at least one section not used by said search engine to recall said each document.
12. The computer-readable storage medium of claim 8 , said search engine using a search engine index to recall said plurality of documents; and wherein said search engine index does not index any term in said certain documents that is found only in one or more sections of said certain documents established as a no-recall section.
13. The computer-readable storage medium of claim 8 , wherein ranking a plurality of documents recalled by a search engine for a query depends in part on a vector space model associated with each of said documents.
14. The computer-readable storage medium of claim 8 , wherein ranking a plurality of documents recalled by a search engine for a query depends in part on a language model associated with each of said documents.