Fulfilling queries using specified and unspecified attributes

摘要

A query is received and processed to determine one or more specified and unspecified attributes in the query. The specified and unspecified attributes may correspond to attributes of one or more items. A graph is generated for the items and includes a node for each item and an edge between each unique pair of nodes. Each node is assigned a cost based on a distance between the specified attributes of the query and the attributes of the item associated with the node. Each edge is assigned a weight based on a distance between the unspecified attributes associated with the nodes of the node pair corresponding to the edge. A set of nodes from the graph is selected by minimizing a total cost of the nodes while maximizing a dispersion of the nodes based on the edge weights.

主权项

1. A method comprising:
receiving identifiers of a plurality of items by a computing device, wherein each item is associated with one or more attributes; receiving a query at the computing device; determining one or more specified attributes of the received query using a correlation of terms of a plurality of previously received queries and attributes associated with the items of the plurality of items by the computing device; determining one or more items from the plurality of items that are associated with attributes that include the determined one or more specified attributes by the computing device; determining one or more attributes associated with the determined one or more items that are not part of the determined one or more specified attributes as one or more unspecified attributes of the received query by the computing device; selecting a subset of the items from the plurality of items based on the one or more attributes associated with each item and the one or more specified attributes and unspecified attributes from the received query by the computing device by:
for each item in the plurality of items, associating a cost with the item based on a distance between the attributes of the item that match the one or more specified attributes from the received query;for each pair of items of the plurality of items, associating a weight with the pair based on a distance between the attributes of the items of the pair of items that match the one or more unspecified attributes from the received query; andselecting the subset of items from the plurality of items using a function that minimizes a total cost of the costs associated with the items in the subset while maximizing a total weight of the weights associated with the pairs of items in the subset; and providing identifiers of the items in the subset of the items by the computing device.