Query suggestions
Briefly, embodiments of methods and/or systems for providing enhancements to search query processing are disclosed.
1. An apparatus, comprising:
a processor executing content, code, and/or instructions stored in memory to:
identify one or more query suggestions as semantically relevant to at least one query via click through feedback to be based, at least in part, on nodes of a weighted graph;
represent the nodes of the weighted graph as the one or more query suggestions, wherein nodes comprising values of a first sign to represent more desirable query suggestions of the one or more query suggestions and nodes comprising values of a sign opposite to the first sign to represent less desirable query suggestions of the one or more query suggestions;
indicate edges of the weighted graph as a correlation between two nodes; and
partition the weighted graph into at least one region to represent the more desirable query suggestions, and at least one other region to represent less desirable query suggestions based, at least in part, on a computed median value of the nodes of the weighted graph.
2. The apparatus of claim 1 , wherein the processor executing content, code, and/or instructions stored in memory is additionally to:
transmit output signals via a network to a client device.
3. The apparatus of claim 1 , wherein processor executing content, code, and/or instructions stored in memory is additionally
to determine the one or more query suggestions via a semi-definite programming (SDP) approach to process an objective function.
4. The apparatus of claim 3 , the processor executing content, code, and/or instructions stored in memory is additionally to employ the semi-definite programming (SDP) approach to perform the partitioning of the weighted graph.
5. The apparatus of claim 3 , the processor executing content, code, and/or instructions stored in memory is additionally to:
transmit output signals to comprise the one or more query suggestions.
6. The apparatus of claim 5 , wherein the processor executing content, code, and/or instructions stored in memory is additionally to transmit the output signals via a network to a client device.
7. A method comprising:
executing, by processor, computer content, code, and/or instructions stored in memory to provide:
identifying one or more query suggestions as semantically relevant to at least one query via click through feedback based, at least in part, on nodes of a weighted graph, the nodes of the weighted graph to represent a plurality of query suggestions; and
partitioning the weighted graph into at least one region to represent more desirable query suggestions of the plurality of query suggestions and at least one region to represent less desirable query suggestions of the plurality of query suggestions based, at least in part, on a computed median value of the nodes of the weighted graph, wherein
nodes comprising values of a first sign to represent the more desirable query suggestions and nodes comprising values of a second sign opposite to the first sign to represent less desirable query suggestions.
8. The method of claim 7 , and further comprising transmitting output signals comprising the one or more identified query suggestions via a network to a client device.
9. The method of claim 7 , further comprising processing of an objective function utilizing a semi-definite programming (SDP) approach to identify query suggestions semantically related to terms of the at least one query.
10. The method of claim 9 , wherein the SDP approach comprises the partitioning of the weighted graph.
11. The method of claim 10 , and further comprising:
transmitting output signals comprising the one or more query suggestions.
12. The method of claim 11 , wherein the transmitting comprises transmitting the output signals via a network to a client device.
13. An article comprising:
a non-transitory storage medium having stored thereon content, code, and/or instructions capable of being executed by at least one computing device coupled to a memory to identify one or more query suggestions as semantically relevant to at least one query via click through feedback based, at least in part, on nodes of a weighted graph, the nodes of the weighted graph to represent query suggestions, wherein
the weighted graph to be partitioned into at least one region to represent more desirable query suggestions and at least one region to represent less desirable query suggestions based, at least in part, on a computed a median value of the nodes of the weighted graph, wherein nodes comprising values of a first sign to represent the more desirable query suggestions and nodes comprising values of a sign opposite to the first sign to represent less desirable query suggestions.
14. The article of claim 13 , wherein the content, code, and/or instructions to determine the one or more query suggestions are to use a semi-definite programming (SDP) approach to process an objective function.
15. The article of claim 14 , the content, code, and/or instructions further to employ the SDP approach to perform the partitioning of the weighted graph.