Clustering and recommending items based upon keyword analysis
AssigneeAmazon Technologies, Inc.
InventorsAaron James Dykstra, Saurabh Nangia, Stephen B. Ivie, David Michael Hurley
Life of this patent
Estimating the term…
Abstract
A multi-level approach for generating suggestions and/or recommendations associated with one or more items is provided. A group of related items may be identified, for example, by generating a collaborative filtering graph or other graph that includes items and connections between the items. Description data associated with items included in the group may be evaluated in order to determine respective keywords associated with the items and/or a list of common keywords representative of the group of items or a subset of the group of items. Based at least in part upon the list of common keywords, at least one suggestion may be generated.
Claims29
A system, comprising:
- at least one memory configured to store computer-executable instructions;
- at least one processor configured to access the at least one memory and execute the computer-executable instructions to:
- generate a collaborative filtering graph comprising:
- a plurality of nodes, wherein a node corresponds to an item offered in an electronic catalog;
- and
- a plurality of edges, wherein an edge connects a pair of nodes of the plurality of nodes that have at least one co-occurred user interaction;
- determine, based at least in part on the plurality of edges, a subset of the plurality of nodes of the collaborative filtering graph that are related;
- determine, for each of at least two nodes of the subset, a list of keywords comprising one or more keywords from a description associated with the item corresponding to the node;
- determine, based at least in part on evaluation of the keywords determined for the at least two nodes of the subset, a list of common keywords comprising one or more keywords shared by the lists of keywords for the at least two nodes of the subset, wherein the subset is smaller than a whole set of the plurality of nodes;
- and
- generate, based at least in part upon the determined list of common keywords, at least one suggestion.
A method, comprising:
- under control of one or more computer systems configured with executable instructions,
- identifying a plurality of related items of a graph, the graph comprising:
- a plurality of items offered in an electronic catalog;
- and
- a plurality of edges, wherein an edge connects a pair of items that have at least one co-occurred user interaction, and
- wherein the identifying is based at least in part on the plurality of edges;
- determining, for each of two or more items of the plurality of related items, a list of keywords comprising one or more keywords from a description associated with the item;
- determining, based at least in part upon an evaluation of the list of keywords for the two or more items of the plurality of related items, a list of common keywords comprising one or more keywords shared by the lists of keywords for the two or more items of the plurality of related items;
- and
- generating, based at least in part upon the determined list of common keywords, at least one suggestion.
One or more non-transitory computer-readable media storing computer-executable instructions that, when executed by at least one processor, configure the at least one processor to:
- generate a collaborative filtering graph comprising:
- a plurality of nodes, wherein a node corresponds to an item offered in an electronic catalog;
- and
- a plurality of edges, wherein an edge connects a pair of nodes of the plurality of nodes that have at least one co-occurred user interaction;
- determine, based at least in part on the plurality of edges, a subset of the plurality of nodes of the collaborative filtering graph that are related;
- determine, for each of at least two nodes of the subset, a list of keywords comprising one or more keywords from a description associated with the item corresponding to the node;
- determine, based at least in part on evaluation of the keywords determined for the at least two nodes of the subset, a list of common keywords comprising one or more keywords shared by the lists of keywords for the at least two nodes of the subset, wherein the subset is smaller than a whole set of the plurality of nodes;
- and
- generate, based at least in part upon the determined list of common keywords, at least one suggestion.
Dependencies are inferred from the claim text.