Systems and methods for determining price bands and user price affinity predictions using machine learning architectures and techniques
A system comprising one or more processors and one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform operations comprising: evaluating, using a price band determination model, degrees of expensiveness of items relative to each other in item type categories; generating, using the price band determination model, price bands associated with item type categories; assigning each of the items to a respective one of the price bands associated with a respective one of the item type categories; and presenting, to one or more end-user applications, at least one other item corresponding to at least one of the price bands associated with at least one of the item type categories. Other embodiments are disclosed.
1. A system comprising:
one or more processors; and
one or more non-transitory computer-readable media storing computing instructions that, when executed on the one or more processors, cause the one or more processors to perform operations comprising:
evaluating, using a price band determination machine learning model, degrees of prices of items relative to each other in item type categories;
generating, using the price band determination machine learning model, price bands associated with the item type categories, wherein the price band determination machine learning model is trained using training procedures to generate the price bands by utilizing at least a clustering model, and wherein the clustering model comprises at least one of a K-means clustering model or a Gaussian mixture model;
assigning each of the items to a respective one of the price bands associated with a respective one of the item type categories, wherein the assigning further comprises using a similarity model to:
analyze historical usage data to determine (a) correlations between purchased items in separate item type categories and (b) corresponding price bands for the purchased items,
compute item-item price band similarity scores across item type categories, and
combine input features including price affinity predictions and the item-item price band similarity scores to generate ranking results that are used to generate item recommendation results; and
presenting, to one or more end-user applications, at least one other item corresponding to at least one of the price bands associated with at least one of the item type categories, wherein the item recommendation results comprise the at least one other item.
2. The system of claim 1 , wherein generating the price bands further comprises:
analyzing price band activity data associated with an item of the item type categories, wherein the price band activity comprises interactions by a user; and
generating activity patterns of the user based on the interactions.
3. The system of claim 1 , wherein the instructions, when executed on the one or more processors, cause the one or more processors to further perform an operation comprising generating, using an affinity prediction model, price affinity predictions for users, wherein the price affinity predictions for the users predict preferences for the users with respect to the price bands associated with the item type categories.
4. The system of claim 1 , wherein the instructions, when executed on the one or more processors, cause the one or more processors to further perform an operation comprising ranking, using a ranking engine using machine learning, outputs in an ordered sequence.
5. The system of claim 1 , wherein the one or more end-user applications utilize price indicators associated with the degrees of prices of the items to generate outputs, and wherein the outputs generated by the one or more end-user applications comprise:
recommendation results ranked based at least in part on the price indicators;
search results ranked in an ordered list based at least in part on the price indicators; or advertisements selected based at least in part on the price indicators.
6. The system of claim 5 , wherein:
the price indicators of the items are stored in metadata associated with the items; and
the price indicators of the items identify the price bands that are associated with the items.
7. The system of claim 1 , wherein the price band determination machine learning model comprises a transaction balancing model to identify respective price range boundaries associated with respective price bands of the price bands for respective ones of the item type categories, and wherein the transaction balancing model further identifies the respective price range boundaries for the respective price bands of the price bands in a manner that accounts for transaction volumes associated with respective ones of the items.
8. The system of claim 1 , wherein the clustering model identifies respective price range boundaries associated with the price bands.
9. The system of claim 1 , wherein the price band determination machine learning model executes a statistical model to identify respective price range boundaries associated with the price bands, and wherein the statistical model comprises a range-based statistical model or a percentile-based statistical model.
10. A method implemented via execution of computing instructions configured to run at one or more processors and configured to be stored at non-transitory computer-readable media, the method comprising:
evaluating, using a price band determination machine learning model, degrees of prices of items relative to each other in item type categories;
generating, using the price band determination machine learning model, price bands associated with the item type categories, wherein the price band determination machine learning model is trained using training procedures to generate the price bands by utilizing at least a clustering model, and wherein the clustering model comprises at least one of a K-means clustering model or a Gaussian mixture model;
assigning each of the items to a respective one of the price bands associated with a respective one of the item type categories, wherein the assigning further comprises using a similarity model to:
analyze historical usage data to determine (a) correlations between purchased items in separate item type categories and (b) corresponding price bands for the purchased items,
compute item-item price band similarity scores across item type categories, and
combine input features including price affinity predictions and the item-item price band similarity scores to generate ranking results that are used to generate item recommendation results; and
presenting, to one or more end-user applications, at least one other item corresponding to at least one of the price bands associated with at least one of the item type categories, wherein the item recommendation results comprise the at least one other item.
11. The method of claim 10 , wherein generating the price bands further comprises:
analyzing price band activity data associated with an item of the item type categories, wherein the price band activity comprises interactions by a user; and
generating activity patterns of the user based on the interactions.
12. The method of claim 10 , wherein the instructions, when executed on the one or more processors, cause the one or more processors to further perform an operation comprising generating, using an affinity prediction model, price affinity predictions for users, wherein the price affinity predictions for the users predict preferences for the users with respect to the price bands associated with the item type categories.
13. The method of claim 10 , wherein the instructions, when executed on the one or more processors, cause the one or more processors to further perform an operation comprising ranking, using a ranking engine using machine learning, outputs in an ordered sequence.
14. The method of claim 10 , wherein the one or more end-user applications utilize price indicators associated with the degrees of prices of the items to generate outputs, and wherein the outputs generated by the one or more end-user applications comprise:
recommendation results ranked based at least in part on the price indicators;
search results ranked in an ordered list based at least in part on the price indicators; or
advertisements selected based at least in part on the price indicators.
15. The method of claim 14 , wherein:
the price indicators of the items are stored in metadata associated with the items; and
the price indicators of the items identify the price bands that are associated with the items.
16. The method of claim 10 , wherein the price band determination machine learning model comprises a transaction balancing model to identify respective price range boundaries associated with respective price bands of the price bands for respective ones of the item type categories, and wherein the transaction balancing model further identifies the respective price range boundaries for the respective price bands of the price bands in a manner that accounts for transaction volumes associated with respective ones of the items.
17. The method of claim 10 , wherein the clustering model identifies respective price range boundaries associated with the price bands.
18. The method of claim 10 , wherein the price band determination machine learning model executes a statistical model to identify respective price range boundaries associated with the price bands, and wherein the statistical model comprises a range-based statistical model or a percentile-based statistical model.
19. A non-transitory computer-readable medium storing computing instructions that, when executed by one or more processors, cause the one or more processors to perform operations comprising:
evaluating, using a price band determination machine learning model, degrees of prices of items relative to each other in item type categories;
generating, using the price band determination machine learning model, price bands associated with the item type categories, wherein the price band determination machine learning model is trained using training procedures to generate the price bands by utilizing at least a clustering model, and wherein the clustering model comprises at least one of a K-means clustering model or a Gaussian mixture model;
assigning each of the items to a respective one of the price bands associated with a respective one of the item type categories, wherein the assigning further comprises using a similarity model to:
analyze historical usage data to determine (a) correlations between purchased items in separate item type categories and (b) corresponding price bands for the purchased items,
compute item-item price band similarity scores across item type categories, and
combine input features including price affinity predictions and the item-item price band similarity scores to generate ranking results that are used to generate item recommendation results; and
presenting, to one or more end-user applications, at least one other item corresponding to at least one of the price bands associated with at least one of the item type categories, wherein the item recommendation results comprise the at least one other item.
20. The non-transitory computer-readable medium of claim 19 , wherein the operations further comprise:
ranking, using a ranking engine using machine learning, to rank outputs in an ordered sequence.