Information processing apparatus, information processing method, and non-transitory computer readable storage medium

Patent metadata

Field	Value
Publication number	US-10467650-B2
Application number	US-201514847918-A
Country	US
Kind code	B2
Filing date	Sep 8, 2015
Priority date	Sep 19, 2014
Publication date	Nov 5, 2019
Grant date	Nov 5, 2019

Abstract

An information processing apparatus includes: a model generation unit configured to generate a prediction model for calculating a prediction value related to a probability that a user performs an action on the Internet by operating a user terminal, for each group created by grouping users based on user information; a model selection unit configured to select a prediction model suited to the user from the prediction models; and a prediction value calculation unit configured to calculate the prediction value by using the prediction model selected by the model selection unit.

First claim

What is claimed is:

1. An information processing apparatus comprising:
a processor programmed to:
generate a plurality of prediction models for groups created by grouping a plurality of users based on user information, each of the prediction models calculating a prediction value related to a probability that a user performs an action on the Internet by operating a user terminal;
for each of the plurality of users:
acquire a history of actions of a corresponding user,
acquire a feature value based on the acquired history of actions for the corresponding user,
test the plurality of prediction models by inputting the feature value of the corresponding user into each of the plurality of prediction models and determining the predicted value for each of the plurality of prediction models, and
assign a prediction model suited to the corresponding user from the plurality of prediction models, based on a comparison between the predicted value of the assigned prediction model and an actual value associated with the history of actions of the corresponding user;

from a target user of the plurality of users:
acquire user information including a history of actions;
select a prediction model assigned to the target user from the prediction models; and
calculate a prediction value for the target user via the selected prediction model.

2. The information processing apparatus according to claim 1 , wherein the processor generates a common model common among the groups, as the prediction model.

3. The information processing apparatus according to claim 1 , wherein the processor is further programmed to:
obtain a history of actions that the user performed on the Internet by operating the user terminal to test the prediction models, and
assign a prediction model suited to the target user, wherein
the processor selects the assigned prediction model.

4. The information processing apparatus according to claim 3 , wherein the processor is further programmed to:
obtain features value acquired from the history of the actions that the user performed on the Internet by operating the user terminal to test the prediction models, the features value including at least one of: a user attribute, a similarity between a delivery target page and an advertisement, a relativity between the user attribute and the advertisement, information on the advertisement, or a past delivery record,
assign a prediction model suited to the target user,
generate classification information in which the obtained features value is associated with the assigned prediction model, and
store the classification information in a memory, and
after the classification information is stored in the memory for a predetermined period, select a prediction model assigned to the target user based on the stored related information.

5. The information processing apparatus according to claim 1 , wherein the processor is further programmed to:
assign a prediction model suited to the user based on the related information indicating a relationship between the features value related to the actions that the user performed on the Internet by operating the user terminal, and the prediction model, wherein
the processor selects the assigned prediction model.

6. The information processing apparatus according to claim 1 , wherein the processor is further programmed to:
select a prediction model suited to the user based on the related information indicating the relationship between the features value related to the actions that the user performed on the Internet by operating the user terminal, and the prediction model.

7. The information processing apparatus according to claim 1 , wherein the processor is further programmed to:
store, in a memory, an action history being a history of actions that the user performed on the Internet by operating the user terminal, wherein
the processor associates the action history with the selected prediction model and stores the action history in the memory, and
the processor generates the prediction model based on a corresponding action history stored in the memory.

8. The information processing apparatus according to claim 1 , wherein the processor generates the prediction model for each group created by grouping the users based on user attributes.

9. The information processing apparatus according to claim 1 , wherein the processor is further programmed to:
group the users based on the user information, via a Dirichlet process being a statistical technique, including:
setting a value of a parameter to determine the degree of gathering in the Dirichlet process in several levels, and
performing several types of groupings; and

generate the prediction model on a group by group basis for each of the several types of groupings.

10. An information processing method to be executed by a computer, comprising:
generating a plurality of prediction models for groups created by grouping a plurality of users based on user information, each of the prediction models calculating a prediction value related to a probability that a user performs an action on the Internet by operating a user terminal;
for each of the plurality of users:
acquiring a history of actions of a corresponding user,
acquiring a feature value based on the acquired history of actions for the corresponding user,
testing the plurality of prediction models by inputting the feature value of the corresponding user into each of the plurality of prediction models and determining the predicted value for each of the plurality of prediction models, and
assigning a prediction model suited to the corresponding user from the plurality of prediction models, based on a comparison between the predicted value of the assigned prediction model and an actual value associated with the history of actions of the corresponding user;

from a target user of the plurality of users:
acquiring user information including a history of actions;
selecting a prediction model assigned to the target user from the prediction models; and
calculating a prediction value for the target user via the selected prediction model.

11. A non-transitory computer-readable storage medium with an executable program stored thereon, wherein the program instructs a computer to perform:
generating a plurality of prediction models for groups created by grouping a plurality of users based on user information, each of the prediction models calculating a prediction value related to a probability that a user performs an action on the Internet by operating a user terminal;
for each of the plurality of users:
acquiring a history of actions of a corresponding user,
acquiring a feature value based on the acquired history of actions for the corresponding user,
testing the plurality of prediction models by inputting the feature value of the corresponding user into each of the plurality of prediction models and determining the predicted value for each of the plurality of prediction models, and
assigning a prediction model suited to the corresponding user from the plurality of prediction models, based on a comparison between the predicted value of the assigned prediction model and an actual value associated with the history of actions of the corresponding user;

from a target user of the plurality of users:
acquiring user information including a history of actions;
selecting a prediction model assigned to the target user from the prediction models; and
calculating a prediction value for the target user via the selected prediction model.

Assignees

• Yahoo Japan Corp

Inventors

• Tsubouchi Kota

Classifications

• G06N5/025

  Extracting rules from data · CPC title

• G06N20/00

  Machine learning · CPC title

• G06Q30/0242 Primary

  Determining effectiveness of advertisements · CPC title