Automated graph-based identification of early adopter users

• US 10,049,375 B1
• Filed: 03/23/2015
• Issued: 08/14/2018
• Est. Priority Date: 03/23/2015
• Status: Expired due to Fees

• Current Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Original Assignee: Amazon Technologies, Inc. (Amazon.com, Inc.)
• Inventors: Zappella, Giovanni, Ackermann, Marcel, Jenatton, Rodolphe, Palfrey, David Spike, Sandler, Samuel Theodore
• Primary Examiner(s): Rines, Robert D
• Assistant Examiner(s): Vizvary, Gerald C
• Application Number:
  US14/666,023
• Time in Patent Office: 1,240 Days
• Field of Search: 1 1, 370400, 703 23, 705 37, 709226, 709200
• CPC Class Codes: G06F 16/2322 using timestamps
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9574 of access to content, e.g. by caching
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

A system is disclosed that identifies early adopter users by creating a directed graph of item access information for an item category and performing a page rank type process on the item access information. This directed graph may be created in a reverse temporal order. The early adopter users can be identified as the users with nodes in the directed graph that have a threshold number or rate of incoming links directly or indirectly pointing towards the nodes. Using the early adopter users as a sample, systems herein can determine whether to recommend an item based on the popularity of the item with respect to the early adopter users. Further, systems herein can determine an inventory level to maintain for an item based on the popularity of the item with respect to the early adopter users.

Claims

1. A computer-implemented method comprising:

as implemented by one or more computing devices configured with specific computer-executable instructions,

identifying a set of items associated with an item category from a set of item categories;

accessing purchase information for the set of items, wherein the purchase information includes purchase time information identifying when an instance of an item from the set of items was purchased;

for individual items in the set of items;

identifying users who purchased the item; and

creating a computer representation of a unidirectional directed graph of the users who purchased the item based at least in part on the purchase time information, the purchase time information including a relative temporal ordering of one or more item accesses by the users, wherein a first node associated with a first user who purchased the item after a second user is connected with a directed link to a second node associated with the second user, and wherein the directed link points from the first node to the second node thereby configuring the computer representation of the unidirectional directed graph in a reverse temporal order based at least in part on the purchase time information for the item;

creating a computer representation of a composite directed graph based on the computer representation of the unidirectional directed graph for individual items in the set of items;

performing a graph-based centrality measurement process on the computer representation of the composite directed graph to generate a measure of an extent to which users from a set of users are early adopter users;

identifying one or more users as early adopter users based at least in part on the measure of the extent to which users from the set of users are early adopter users, wherein the early adopter users comprise users who purchase items within a threshold release window;

determining based at least in part on the composite directed graph that at least a threshold amount of the early adopter users accessed a digital item that is consumed via streaming; and

adjusting a number of cache servers configured to cache the digital item based at least in part on the determination that at least the threshold amount of the early adopter users accessed the digital item.

2. The computer-implemented method of claim 1, further comprising:

receiving an indication of an item purchase for an instance of a first item from the set of items by a user; and

updating the computer representation of the composite directed graph to incorporate a node associated with the user into the computer representation of the composite directed graph.

3. The computer-implemented method of claim 1, further comprising:

adding an unassigned node to the computer representation of the composite directed graph, the unassigned node not associated with a user; and

creating a link from individual nodes associated with a user in the computer representation of the composite directed graph to the unassigned node.

4. The computer-implemented method of claim 3, wherein the link from individual nodes to the unassigned node is weighted based at least in part on a number of other nodes linked to the node.

5. The computer-implemented method of claim 4, wherein the number of other nodes linked to the node includes nodes indirectly linked to the node.

6. The computer-implemented method of claim 1, further comprising:

receiving an identity of a new item associated with the item category;

determining a number of users from the early adopter users who purchased the new item within a threshold time period; and

in response to the number of users from the early adopter users satisfying a purchase threshold, including the new item in a list of recommended items.

7. The computer-implemented method of claim 1, further comprising:

receiving an identity of a new item associated with the item category;

determining a number of users from the early adopter users who purchased the new item within a threshold time period; and

maintaining an inventory amount of the new item based at least in part on the number of users from the early adopter users who purchased the new item within the threshold time period.

8. A system comprising:

an electronic data store configured to store item access information; and

an interactive computing system in communication with the electronic data store and configured to at least;

identify a set of items associated with an item category from a set of item categories;

access the item access information for the set of items, wherein the item access information includes access time information identifying when an item from the set of items was accessed;

for individual items in the set of items;

identify users who accessed the item; and

create a unidirectional directed graph of the users who accessed the item based at least in part on the time access information, the time access information including a relative temporal ordering of item access by the users, wherein the unidirectional directed graph is configured in a reverse temporal order based at least in part on the access time information for the item;

create a composite directed graph based on the unidirectional directed graph for individual items in the set of items;

perform a graph-based measurement process on the users with corresponding nodes included in the composite directed graph based at least partly on a number of incoming links to the corresponding nodes to obtain an ordered list of users;

identify a set of users from the ordered list of users as early adopter users based on an early adopter threshold; and

determine based at least in part on the composite directed graph that at least a threshold amount of the early adopter users accessed a digital item that is consumed via streaming; and

adjust a number of cache servers configured to cache the digital item based at least in part on the determination that at least the threshold amount of the early adopter users accessed the digital item.

9. The system of claim 8, wherein the early adopter threshold is based at least partly on the item category.

10. The system of claim 8, wherein the early adopter threshold varies based at least partly on a sales rate for the item category.

11. The system of claim 8, wherein the interactive computing system is further configured to at least:

add a fictional user associated with a fictional node to the composite directed graph; and

create a link from individual nodes associated with a user in the composite directed graph to the fictional node.

12. The system of claim 11, wherein the fictional user is included in the ordered list of users and wherein the interactive computing system is further configured to at least weight the link from other nodes in the composite directed graph to the fictional node such that the fictional node is ranked first after ranking the users with corresponding nodes included in the composite directed graph.

13. The system of claim 12, wherein the interactive computing system is further configured to at least filter the fictional node from the ordered list of users.

14. The system of claim 8, wherein a link between a first node and a second node in the composite directed graph is weighted based at least in part on a similarity between a first user associated with the first node and a second user associated with the second node.

15. The system of claim 8, wherein the interactive computing system is further configured to at least create a second composite directed graph associated with a second set of items from a second item category, and wherein a first item associated with a node in the composite directed graph is associated with a node in the second composite directed graph.

16. A computer-readable, non-transitory storage medium storing computer executable instructions that, when executed by one or more computing devices, configure the one or more computing devices to perform operations comprising:

identifying a set of items associated with an item category;

obtaining access time information for the set of items, the access time information including one or more access times for individual items from the set of items;

creating a composite directed graph based on the access time information the access time information including a relative temporal ordering of one or more item accesses by a set of users, wherein the composite directed graph comprises a set of nodes and a set of directed links, wherein individual nodes in the set of nodes is associated with a different user from the set of users, wherein individual users from the set of users accessed at least one item from the set of items, and wherein at least one directed link from the set of directed links points from a second node associated with a second user who accessed a second instance of an item from the set of items to a first node associated with a first user who accessed a first instance of the item from the set of items before the second user accessed the second instance;

performing a ranking process on the composite directed graph to obtain an ordered list of users;

identifying an early adopter set of users from the ordered list of users; and

determining based at least in part on the composite directed graph that at least a threshold amount of users from the early adopter set of users accessed a digital item that is consumed via streaming; and

adjusting a number of cache servers configured to cache the digital item based at least in part on the determination that at least the threshold amount of users from the early adopter set of users accessed the digital item.

17. The computer-readable, non-transitory storage medium of claim 16, wherein the composite directed graph is created from a plurality of discrete directed graphs, and wherein individual discrete directed graphs from the discrete directed graphs are associated with a different item from the set of items.

18. The computer-readable, non-transitory storage medium of claim 16, wherein the operations further comprise:

receiving an identity of a new item associated with the item category;

determining a percentage of the early adopter set of users who access the new item within a threshold time period; and

in response to the percentage of the early adopter users who access the new item satisfying an access threshold, recommending the new item to another user.

19. The computer-readable, non-transitory storage medium of claim 18, wherein the access threshold is based at least in part on the item category.

20. The computer-readable, non-transitory storage medium of claim 16, wherein the operations further comprise:

receiving an identity of a new item associated with the item category;

determining a percentage of the early adopter set of users who access the new item; and

determining an inventory threshold based at least in part on the percentage of the early adopter set of users who access the new item, wherein the inventory threshold is used to determine the number of instances of the new item to maintain in an inventory.