Category predictions for user behavior

Patent metadata

Field	Value
Publication number	US-9767417-B1
Application number	US-201414303516-A
Country	US
Kind code	B1
Filing date	Jun 12, 2014
Priority date	Jun 12, 2014
Publication date	Sep 19, 2017
Grant date	Sep 19, 2017

Abstract

Techniques for providing category predictions may be provided. For example, a process may attempt to improve a user experience when the user provides a search query. The process can predict the category associated with the search query, even when the category is not a keyword in the search query. Once the category is determined, data may be provided for the particular category, including data that enables an adjustment of a user experience. For example, when the category is apparel, the user experience may include an image-heavy layout and, when the category is books, the user experience may provide more text.

First claim

What is claimed is:

1. A computer-implemented method, comprising:
receiving a search query from a user, the search query including a keyword associated with a first category;
accessing a data set that includes one or more items of the first category, one or more categories of the items, and historical data representing one or more actions of other users associated with the items;
determining, by a computer system, at least one related category related to the data set from the one or more categories of items;
determining that a frequency of the at least one related category is greater than a first threshold, the frequency determined by historical data representing selection of an item in the at least one related category by the other users after the other users have submitted the search query and received and accessed the data set;
determining, by the computer system, a conversion rate for the at least one related category by the other users, the conversion rate representing an intended or actual purchase of the item in the related category, the conversion rate greater than a second threshold;
when the frequency of the at least one related category is greater than the first threshold and the conversion rate is greater than the second threshold, providing data associated with the at least one related category to the user, the data provided including a navigation strip that is tailored to the at least one related category but not the first category; and
adjusting a user experience based at least in part on the data and on the frequency, the user experience being adjusted at least by providing a different layout including the navigation strip on a user interface to browse through the data set that corresponds with the at least one related category but not the first category.

2. The computer-implemented method of claim 1 , wherein the keyword is associated with an item identifier, and wherein the item identifier identifies a unique item.

3. The computer-implemented method of claim 2 , wherein the conversion rate is associated with a number of times the item identifier was ordered by other users during a time frame.

4. The computer-implemented method of claim 2 , wherein the conversion rate is associated with a number of times the item identifier was selected by other users during a time frame.

5. The computer-implemented method of claim 1 , wherein adjusting the user experience provides a recommendation for items associated with at least one related the category.

6. The computer-implemented method of claim 1 , wherein adjusting the user experience provides a display of best-selling items from the at least one related category on a user interface of a computing device.

7. The computer-implemented method of claim 1 , wherein adjusting the user experience provides an advertisement for a second item associated with the at least one related category.

8. The method of claim 1 , wherein the different layout displays more text and smaller images than a previous layout.

9. The method of claim 1 , wherein the different layout includes a first presentation format that corresponds with the first category and a second presentation format that corresponds with the at least one related category.

10. The method of claim 1 , wherein the conversion rate is determined from interactions with other users and the computer system, wherein the computer system receives the search query and also provides an electronic marketplace to offer the item for purchase.

11. One or more computer-readable non-transitory storage media collectively storing computer-executable instructions that, when executed by one or more computer systems, configure the one or more computer systems to collectively perform operations comprising:
receiving a search query from a user, the search query including a keyword associated with a first category;
accessing a data set that includes one or more items of the first category or one or more categories of the items;
determining at least one related category related to the data set from the one or more categories of the items;
determining that a frequency of the at least one related category is greater than a first threshold, the frequency determined by historical data representing selection of an item in the at least one related category by the other users after the other users have submitted the search query and received and accessed the data set;
when the frequency of the at least one related category is greater than the first threshold, providing data associated with the at least one related category to the user, the data provided including a navigation strip that is tailored to the at least one related category but not the first category; and
adjusting a user experience based at least in part on the data and on the frequency, the user experience being adjusted at least by providing a different layout including the navigation strip on a user interface to browse through the data set that corresponds with the at least one related category but not the first category.

12. The one or more computer systems of claim 11 , wherein the data set associated with the search query comprises click data, an item identifier, and a number of times the item identifier was ordered during a time frame.

13. The one or more computer systems of claim 11 , wherein the category is determined through a process, wherein the process includes one of the following: simple prediction, smoothed browse node prediction, expanded query group, N-Gram generative model, estimated probability as prior, or interpolation between multiple prediction models.

14. The one or more computer systems of claim 13 , wherein the simple prediction uses a formula:

P
⁡

(

Click
x

|
Q

)

=

∑
N

⁢

w
x

⁢

Click
x

∑
N

⁢

w
x

wherein P is a probability, Click x is an observation of a click in browse node x, N is be a number of observations of query Q, and weight for an impression is w x .

15. The one or more computer systems of claim 13 , wherein the smoothed browse node prediction uses a formula:

P
⁡

(

Click
x

|
Q

)

=

∑
N

Assignees

• Amazon Tech Inc

Inventors

• Hoover Thomas Jay

Classifications

• G06F17/30598

  Physics · mapped topic

• G06N7/02 Primary

  using fuzzy logic (computing arrangements based on biological models G06N3/00; computing arrangements using knowledge-based models G06N5/00) · CPC title

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/285

  Clustering or classification · CPC title