Cognitive fashion-ability score driven fashion merchandising acquisition

• US 10,755,229 B2
• Filed: 04/11/2018
• Issued: 08/25/2020
• Est. Priority Date: 04/11/2018
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Sewak, Mohit
• Primary Examiner(s): Wilder, A. Hunter
• Application Number:
  US15/950,692
• Publication Number:
  US 20190318299A1
• Time in Patent Office: 867 Days
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06Q 10/087 Inventory or stock management, e.g. order filling, procurement or balancing against orders
• CPC Class Codes: G06Q 30/0631 Item recommendations

Abstract

A first plurality of f-scores for a first plurality of products in an inventory are determined. The first plurality of f-scores are discretized to generate a plurality of groups, and a forecast is generated for each of the groups based on historical data associated with the first plurality of products. Projected gaps in the inventory are identified based on the forecasts. A second plurality of f-scores are determined for a second plurality of products, where each of the second plurality of products is not in the inventory. For each of the second plurality of products, a corresponding group in the plurality of groups is identified based on the second plurality of f-scores, an at least one product in the second plurality of products is selected to order, based on determining that the identified corresponding group aligns with at least one of the one or more projected gaps in inventory.

Claims

1. A system comprising:

one or more computer processors; and

a memory containing a program which when executed by the one or more computer processors performs an operation, the operation comprising;

determining a first plurality of f-scores for a first plurality of products based on processing a first plurality of images using one or more convolutional neural networks (CNNs), wherein each of the first plurality of images corresponds to one of the first plurality of products, wherein each of the first plurality of products is in an inventory, and wherein each of the first plurality of f-scores represent aggregated N-dimensional vectors corresponding to N fashion attributes and are used to compare the first plurality of products to determine their objective similarity based on the N fashion attributes;

discretizing the first plurality of f-scores to generate a plurality of groups;

generating, for each of the plurality of groups, a forecast based on historical data associated with each of the first plurality of products in each respective group;

identifying one or more projected gaps in the inventory, based on the generated forecasts;

determining a second plurality of f-scores for a second plurality of products based on processing a second plurality of images with the one or more CNNs, wherein each of the second plurality of products is not in the inventory;

identifying, for each of the second plurality of products, a corresponding group in the plurality of groups based at least in part on the second plurality of f-scores;

selecting at least one product in the second plurality of products to order, based at least in part on determining that the identified corresponding group aligns with at least one of the one or more projected gaps in inventory; and

initiating an order for the selected at least one product in the second plurality of products.

2. The system of claim 1, wherein selecting at least one product in the second plurality of products is further based on one or more determined interaction effects between the first plurality of products, wherein the interaction effects are determined based on the associated historical data.

3. The system of claim 1, wherein selecting at least one product in the second plurality of products is further based on a determined price elasticity of the identified corresponding group of the at least one product, wherein the price elasticity of the corresponding group is determined based on the associated historical data.

4. The system of claim 1, wherein selecting at least one product in the second plurality of products is further based on a determined fulfillment time for the at least one product.

5. The system of claim 1, the operation further comprising:

determining that at least one of the one or more projected gaps cannot be satisfactorily filled by the second plurality of products; and

generating a recommendation to design and develop a new product that satisfies the at least one projected gap.

6. The system of claim 1, the operation further comprising:

determining that a public event has occurred, wherein the public event involves a third plurality of products;

determining an f-score for each of the third plurality of products, based on processing a third plurality of images, wherein each of the third plurality of images corresponds to one of the third plurality of products;

identifying, for each of the third plurality of products, a corresponding group in the plurality of groups based on the third plurality of f-scores; and

generating, for each of the plurality of groups, a revised forecast based at least in part on each identified corresponding group in the plurality of groups and further based on historical trends from one or more previous public events.

7. A computer program product comprising a non-transitory computer-readable storage medium having computer-readable program code embodied therewith, the computer-readable program code executable by one or more computer processors to perform an operation comprising:

determining a first plurality of f-scores for a first plurality of products based on processing a first plurality of images using one or more convolutional neural networks (CNNs), wherein each of the first plurality of images corresponds to one of the first plurality of products, wherein each of the first plurality of products is in an inventory, and wherein each of the first plurality of f-scores represent aggregated N-dimensional vectors corresponding to N fashion attributes and are used to compare the first plurality of products to determine their objective similarity based on the N fashion attributes;

discretizing the first plurality of f-scores to generate a plurality of groups;

generating, for each of the plurality of groups, a forecast based on historical data associated with each of the first plurality of products in each respective group;

identifying one or more projected gaps in the inventory, based on the generated forecasts;

determining a second plurality of f-scores for a second plurality of products based on processing a second plurality of images with the one or more CNNs, wherein each of the second plurality of products are not in the inventory;

identifying, for each of the second plurality of products, a corresponding group in the plurality of groups based at least in part on the second plurality of f-scores;

selecting at least one product in the second plurality of products to order, based at least in part on determining that the identified corresponding group aligns with at least one of the one or more projected gaps in inventory; and

initiating an order for the selected at least one product in the second plurality of products.

8. The computer program product of claim 7, wherein selecting at least one product in the second plurality of products is further based on one or more determined interaction effects between the first plurality of products, wherein the interaction effects are determined based on the associated historical data.

9. The computer program product of claim 7, wherein selecting at least one product in the second plurality of products is further based on a determined price elasticity of the identified corresponding group of the at least one product, wherein the price elasticity of the corresponding group is determined based on the associated historical data.

10. The computer program product of claim 7, wherein selecting at least one product in the second plurality of products is further based on a determined fulfillment time for the at least one product.

11. The computer program product of claim 7, the operation further comprising:

determining that at least one of the one or more projected gaps cannot be satisfactorily filled by the second plurality of products; and

generating a recommendation to design and develop a new product that satisfies the at least one projected gap.

12. The computer program product of claim 7, the operation further comprising:

determining that a public event has occurred, wherein the public event involves a third plurality of products;

determining an f-score for each of the third plurality of products, based on processing a third plurality of images, wherein each of the third plurality of images corresponds to one of the third plurality of products;

identifying, for each of the third plurality of products, a corresponding group in the plurality of groups based on the third plurality of f-scores; and

generating, for each of the plurality of groups, a revised forecast based at least in part on each identified corresponding group in the plurality of groups and further based on historical trends from one or more previous public events.