Training and estimation of selection behavior of target

Patent metadata

Field	Value
Publication number	US-11423324-B2
Application number	US-201715440726-A
Country	US
Kind code	B2
Filing date	Feb 23, 2017
Priority date	Feb 23, 2017
Publication date	Aug 23, 2022
Grant date	Aug 23, 2022

Abstract

A training method is provided. The training method includes clustering, by a processor, a plurality of items that each have an item attribute value, according to the item attribute value. The training method further includes generating, by the processor, for each item, a cluster attribute value corresponding to a cluster associated with the item. The training method also includes training, by the processor, an estimation model for estimating selection behavior of a target with respect to a choice set including two or more items, based on the cluster attribute value associated with each item included in the choice set, by using training data that includes a group of a choice set of items presented to the target and an item selected by the target from among the choice set.

First claim

What is claimed is:

1. A computer program product for training an estimation model, the computer program product comprising a non-transitory computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computer to cause the computer to perform a method comprising:
clustering, by a processor, a plurality of items that each have an item attribute value, according to the item attribute values, wherein the items in each cluster have a characteristic in common represented by the item attribute value and each cluster is associated with a choice set from a plurality of choice sets;
generating, by the processor, for each item, a cluster attribute value corresponding to a cluster associated with the item; and
training, by the processor, an estimation model for estimating selection behavior of a target with respect to the choice set including two or more items, based on the cluster attribute value associated with each item included in the choice set, by using training data that includes a group of choice sets of differing items from the choice set presented to the target and an item selected by the target from among the choice set, wherein each cluster attribute value is a difference between an item attribute of each item in the choice set and an average vector of the item attribute values of all the items in the choice set.

2. The non-transitory computer-readable storage medium of claim 1 , wherein
the estimation model estimates a selection behavior of the target based on a weighted sum obtained by weighting, with different weights, a plurality of attribute values for a plurality of attributes including an item attribute and a cluster attribute, and
the plurality of attributes include attributes corresponding to different weights for different clusters as the cluster attribute.

3. The non-transitory computer-readable storage medium of claim 1 , wherein the clustering includes setting, according to instructions, at least one of a number of clusters into which the plurality of items are to be classified, priority of a plurality of item attributes, and which of a plurality of clustering algorithms is to be used.

4. The non-transitory computer-readable storage medium of claim 1 , wherein the choice set is a subset of items of the plurality of items.

5. The non-transitory computer-readable storage medium of claim 4 , wherein the program instructions executable by a computer further cause the computer to perform a method comprising estimating the selection behavior of a target with respect to a new choice set, using the estimation model, the estimation model being trained to account for cognitive bias, based on the cluster attribute value associated with each item included in the new choice set.

6. An apparatus comprising a processor configured to:
cluster a plurality of items that each have an item attribute value, according to the item attribute value, wherein the items in each cluster have a characteristic in common represented by the item attribute value and each cluster is associated with a choice set from a plurality of choice sets;
generate for each item, a cluster attribute value corresponding to a cluster associated with the item; and
train an estimation model for estimating selection behavior of a target with respect to the choice set including two or more items, based on the cluster attribute value associated with each item included in the choice set, by using training data that includes a group of choice sets of differing items from the choice set presented to the target and an item selected by the target from among the choice set, wherein each cluster attribute value is a difference between an item attribute of each item in the choice set and an average vector of the item attribute values of all the items in the choice set.

7. The apparatus of claim 6 , wherein the choice set is a subset of items of the plurality of items.

8. The apparatus of claim 7 , wherein the processor is further configured to estimate the selection behavior of a target with respect to a new choice set, using the estimation model, the estimation model being trained to account for cognitive bias, based on the cluster attribute value associated with each item included in the new choice set.

9. An apparatus comprising:
a processor configured to:
acquire for each of a plurality of items that each have an item attribute value and are clustered according to the item attribute values, wherein the items in each cluster have a characteristic in common represented by the item attribute value and each cluster is associated with a choice set from a plurality of choice sets, a cluster attribute value corresponding to a cluster associated with the item;
acquire an estimation model for estimating selection behavior of a target with. respect to a choice set including two or more items, based on the cluster attribute value associated with each item included in the choice set;
acquire an additional choice set of items presented to the target from among the plurality of items;
train, by the processor, the estimation model to estimate selection behavior of the target with respect to the additional choice set of items, by using the estimation model, based on the cluster attribute value associated with each item included in the additional choice set, wherein each cluster attribute value is a difference between an item attribute value of each item in the additional choice set and an average vector of the item attribute values of all the items in the additional choice set.

10. The apparatus of claim 9 , wherein the processor is further configured to:
acquire an item attribute value of an additional item;
determine which cluster among a plurality of clusters the additional item is associated with, by using the item attribute value of the additional item; and
generate, for the additional item, the cluster attribute value corresponding o the cluster associated with the additional item,
wherein the processor is further configured to estimate the selection behavior of the target by estimating the selection behavior of the target for the additional choice set that includes the additional item.

11. The apparatus of claim 10 , wherein the processor is further configured to re-cluster the plurality of items including the additional item, in response to a determination that the additional item is not associated with any of the plurality of clusters.

12. The apparatus of claim 11 , wherein the processor is further configured to generate a cluster attribute value corresponding to the cluster associated with each item among the plurality of items, after the re-clustering, and to retrain the estimation model using training data that includes a group of choice sets of differing items from the choice set of items presented to the target and an item selected by the target from among the choice set of items.

13. The apparatus of claim 9 , wherein the processor is further configured to generate a choice set of items for enabling the target to select a certain item, using the estimation model.

14. The apparatus of claim 9 , wherein the choice set is a subset of items of the plurality of items.

15. The apparatus of claim 14 , wherein the processor is further configured to estimate the selection behavior of a target with respect to a new choice set, using the estimation model, the estimation model being trained to account for cognitive bias, based on the cluster attribute value associated with each item included in the new choice set.

Assignees

• IBM

Inventors

• Morimura Tetsuro
• Obara Yachiko
• Osogami Takayuki

Classifications

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06Q30/02

  Marketing; Price estimation or determination; Fundraising · CPC title

• G06F16/355

  Creation or modification of classes or clusters · CPC title

• G06N20/00 Primary

  Machine learning · CPC title

• G06N5/022

  Knowledge engineering; Knowledge acquisition · CPC title