Back to results
US 10,475,101 B1Granted
Determining potential causes of an issue associated with recommendations and changing recommendation filter settings based on the outcome of an action
AssigneeAmazon Technologies, Inc.
InventorsGualtiero Forte, Grant Michael Emery, Madhu Madhava Kurup, Ram Prasad Venkatesan
Life of this patent
Estimating the term…
Abstract
Techniques for analyzing recommendations may be described. In particular, an item offered from a network-based resource may be selected based at least in part on past orders for the item. Recommended items offered from the network-based resource may also be identified. A determination may be made as to whether the recommended items may include the item. Based on this determination, an indication of an issue associated with recommending the item may be generated. Based on the indication, a workflow to identify a potential cause of the issue may be initiated.
Claims20
A computer-implemented method, comprising:
- selecting, by a computer system associated with an electronic marketplace, an item offered from a web site of the electronic marketplace, the item selected based at least in part on a number of orders for the item exceeding a threshold;
- selecting, by the computer system, a plurality of items based at least in part on associations with the item, the plurality of items offered from the electronic marketplace, each association generated based at least in part on a first user order for the item and for one or more items of the plurality of items, the associations stored in a data store and indicating that the item is recommendable for ordering based at least in part on a second user order for any other item of the plurality of items;
- providing, by the computer system, identifiers of the plurality of items to a recommendation engine of the web site, the recommendation engine hosted on the computer system;
- receiving, from the recommendation engine, a list of recommended items determined based at least in part on the identifiers, the list of recommended items generated by the recommendation engine based at least in part on the associations stored in the data store;
- determining, by the computer system, that an identifier of the item is absent from the list of recommended items;
- generating, by the computer system, an indication of an issue associated with recommending the item based at least in part on the identifier of the item being absent from the list of recommended items;
- determining, by the computer system from potential causes identified in a workflow, that a recommendation filter setting is a potential cause of the issue, the workflow identifying a selection of the potential cause from the potential causes based at least in part on a comparison of a frequency of occurrence of the issue to a frequency threshold, the workflow further identifying an action to be performed to test the potential cause, the action comprising:
- updating the recommendation filter setting to a second recommendation filter setting on the recommendation engine based on an application programming interface (API) call,
- replicating the list of recommended items by using the second recommendation filter setting on the recommendation engine, and
- determining whether the identifier of the item is present in the replicated list of recommended items;
- and
- changing, by the computer system, the recommendation filter setting to the second recommendation filter setting based at least in part on an outcome of performing the action identified in the workflow indicating that the identifier of the item is present in the replicated list of recommended items.
9 more elements
One or more non-transitory computer-readable media comprising instructions that, when executed with one or more processors, cause a computing system to perform operations comprising at least:
- selecting an item offered from a web site of a marketplace based at least in part on past orders for the item;
- identifying, from a data store and based at least in part on a recommendation engine of the web site, recommended items offered from the web site of the marketplace, the recommended items recommendable by the recommendation engine based at least in part on a request for information about the item;
- determining whether the recommended items comprise the item;
- generating an indication of an issue associated with recommending the item based at least in part on whether the recommended items comprise the item;
- initiating a workflow to identify a potential cause of the issue based at least in part on the indication of the issue;
- determining, from potential causes associated with the workflow, that a filter setting of the recommendation engine is the potential cause of the issue, the workflow indicating a selection of the potential cause from the potential causes based at least in part on a comparison of a frequency of occurrence of the issue to a frequency threshold, the workflow further indicating an action to be performed to test the potential cause, the action comprising:
- updating the filter setting to a second filter setting on the recommendation engine based on an application programming interface (API) call,
- replicating a list of recommended items by using the second filter setting on the recommendation engine, and
- determining whether the item is identified in the list of recommended items;
- and
- changing the filter setting of the recommendation engine based at least in part on an outcome of performing the action.
8 more elements
A system comprising:
- one or more processors;
- and
- one or more computer-readable media comprising instructions that, when executed with the one or more processors, cause the system to at least:
- select an item offered from a network-based resource based at least in part on determining that a number of past orders for the item exceeds a threshold;
- identify, from a data store and based at least in part on a recommendation engine of the network-based resource, recommended items offered at the network-based resource, the recommended items recommendable by the recommendation engine based at least in part on a request for information about the item;
- generate an indication of an issue associated with recommending the item based at least in part on determining that the recommended items exclude the item;
- initiate a workflow to identify a potential cause of the issue based at least in part on the indication of the issue;
- determine, from potential causes associated with the workflow, that a filter setting of the recommendation engine is the potential cause of the issue, the workflow indicating a selection of the potential cause from the potential causes based at least in part on a comparison of a frequency of occurrence of the issue to a frequency threshold, the workflow further indicating an action to be performed to test the potential cause, the action comprising:
- updating the filter setting to a second filter setting on the recommendation engine based on an application programming interface (API) call,
- replicating a list of recommended items by using the second filter setting on the recommendation engine, and
- determining whether the item is identified in the list of recommended items;
- and
- change the filter setting of the recommendation engine based at least in part on an outcome of performing the action.
10 more elements
Dependencies are inferred from the claim text.