Mapping product identification information to a product
View Patent ↗Mapping product identification data to product information. Generating classifiers operative to classify products referred to in product identification data. Receiving client product identification data. Applying at least one generated classifier on the received product identification data. Using at least one classification determined by the applied classifier(s) and a subset of the client product identification data to query the product database. Qualifying the results of the query using at least one classification.
1. A method for mapping product identification data to product information contained in a product information database, the product information database further comprising product classifications, the method comprising:
receiving product identification data;
predicting at least one first classification of the product identification data;
generating a query based on a subset of the received product identification data and at least one predicted first classification;
searching the product information database using the query;
qualifying one or more results of the query at least in part using the at least one predicted first classification;
determining a mapping prediction for each of the one or more results of the query, the mapping prediction including a first label and a first probability,
the first label including a proposition that the product identification data indicates a product that is out of scope of the product information database,
the first probability including a likelihood that the proposition is true;
if the first probability exceeds a first uncertainty threshold, identifying a product associated with the received product identification data as out of scope of the product information database; and
mapping the received product identification data to a result qualifying at or above a qualification threshold.
2. The method of claim 1 , wherein predicting at least one first classification of the product identification data comprises:
determining a plurality of predictions, each prediction including
a second label and a second probability,
the second label including a proposition regarding classification of the received product identification data,
the second probability including the likelihood that the proposition is true; and
for the highest second probability exceeding a second uncertainty threshold, predicting the classification associated with the proposition.
3. The method of claim 2 , wherein the uncertainty threshold is determined based at least in part on a ratio of the highest first probability and the second highest first probability.
4. The method of claim 1 , wherein the at least one first classification of the product identification data includes at least one of:
a category classification of the product identification data,
an accessory classification of the product identification data,
a pack classification of the product identification data, or
a third party classification of the product identification data.
5. The method of claim 1 , wherein qualifying the results of the query comprises determining an extent to which product identification data is found in the results of the query.
6. The method of claim 1 further comprising presenting the received product information and the one or more results of the query for resolution by a user if the one or more results do not qualify at or above the qualification threshold.
7. A non-transitory computer readable medium that stores instructions for mapping product identification data to product information contained in a product information database, the product information database further comprising product classifications, the instructions being executable by one or more processors to perform steps comprising:
receiving product identification data;
predicting at least one first classification of the product identification data;
generating a query based on a subset of the received product identification data and at least one predicted first classification;
searching the product information database using the query;
qualifying one or more results of the query at least in part using the at least one predicted first classification; and
determining a mapping prediction for each of the one or more results of the query, the mapping prediction including a first label and a first probability,
the first label including a proposition that the product identification data indicates a product that is out of scope of the product information database,
the first probability including a likelihood that the proposition is true;
if the first probability exceeds a first uncertainty threshold, identifying a product associated with the received product identification data as out of scope of the product information database; and
mapping the received product identification data to a result qualifying at or above a qualification threshold.
8. The non-transitory computer readable medium of claim 7 , wherein the instructions for predicting at least one first classification of the product identification data include instructions for:
determining a plurality of predictions, each prediction including
a second label and a second probability,
the second label including a proposition regarding classification of the received product identification data,
the second probability including the likelihood that the proposition is true; and
for the highest probability exceeding a second uncertainty threshold, predicting the classification associated with the proposition.
9. The non-transitory computer readable medium of claim 8 , wherein the uncertainty threshold is determined at least in part on a ratio of the highest second probability and the second highest second probability.
10. The non-transitory computer readable medium of claim 7 , wherein the at least one first classification of the product identification data includes at least one of:
a category classification of the product identification data,
an accessory classification of the product identification data,
a pack classification of the product identification data, or
a third party classification of the product identification data.
11. The non-transitory computer readable medium of claim 7 , wherein the instructions for qualifying the results of the query include instructions for determining an extent to which product identification data is found in the results of the query.
12. The non-transitory computer readable medium of claim 7 further comprising instructions for presenting the received product information and the one or more results of the query for resolution by a user if the one or more results do not qualify at or above the qualification threshold.
13. A system for mapping product identification data to product information, the system comprising:
a product information database including a plurality of product classifications; and
at least one tangible computer readable medium storing instructions-executable by a processor to perform operations that include:
receive product identification data;
predict at least one first classification of the product identification data;
generate a query based on a subset of the received product identification data and at least one predicted first classification;
search the product information database using the query;
qualify one or more results of the query at least in part using the at least one predicted first classification;
determine a mapping prediction for each of the one or more results of the query, the mapping prediction including a first label and a first probability,
the first label including a proposition that the product identification data indicates a product that is out of scope of the product information database,
the first probability including a likelihood that the proposition is true;
if the first probability exceeds a first uncertainty threshold, identify a product associated with the received product identification data as out of scope of the product information database; and
map the received product identification data to a result qualifying at or above a qualification threshold.
14. The system of claim 13 , wherein the processor is configured to predict the at least one first classification of the product identification data by:
determining a plurality of predictions, each prediction including a second label and a second probability,
the second label including a proposition regarding classification of the received product identification data,
the second probability including the likelihood that the proposition is true; and
for the highest probability exceeding a second uncertainty threshold, predicting the classification associated with the proposition.
15. The system of claim 14 , wherein the uncertainty threshold is determined at least in part on a ratio of the highest first probability and the second highest first probability.
16. The system of claim 13 , wherein the at least one first classification of the product identification data includes at least one of:
a category classification of the product identification data,
an accessory classification of the product identification data,
a pack classification of the product identification data, or
a third party classification of the product identification data.
17. The system of claim 13 , wherein the processor is further configured to determine an extent to which product identification data is found in the results of the query.
18. The system of claim 13 wherein the processor is further configured to present the received product information and the one or more results of the query for resolution by a user if the one or more results do not qualify at or above the qualification threshold.