# US 2018/0025072 A1 · Clustering system, method, and program

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2018/0025072 A1
- Document ID: US20180025072A1
- Status: Published application
- Kind: A1
- Application number: 15/549,897
- Filed: 2016-01-27
- Published: 2018-01-25
- Earliest claimed: 2015-02-18
- Assignees: NEC Corporation
- Applicants: NEC Corporation
- Inventors: Ryohei FUJIMAKI; Yousuke MOTOHASHI
- Web page: https://patentgrep.com/patent/US20180025072A1

## CPC

- G06F 16/285: {Clustering or classification}
- G06F 16/2246: {Trees, e.g. B+trees}
- G06N 5/01: Dynamic search techniques; Heuristics; Dynamic trees; Branch-and-bound
- G06N 20/00: Machine learning
- G06Q 30/0201: Market modelling; Market analysis; Collecting market data
- G06N 7/01: Probabilistic graphical models, e.g. probabilistic networks
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06Q: INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR

## Abstract

A classifier 81 classifies target data into a cluster on the basis of a mixture model defined using two different types of variables that indicate features of the target data. In this classification, the classifier 81 classifies the target data into a cluster on the basis of a mixture model in which a mixing ratio of the mixture model is represented by a function of a first variable and in which the element distribution of the clusters into which the target data is classified is represented by a function of a second variable.

## Claims

### Claim 1

A clustering system comprising:
a hardware including a processor; and
a classifier, implemented by the processor, which classifies target data into a cluster based on a mixture model defined by using two different types of variables representing features of the target data,
wherein the classifier classifies the target data into a cluster based on the mixture model in which a mixing ratio of the mixture model is represented by a function of a first variable and in which an element distribution of the cluster into which the target data is classified is represented by a function of a second variable.

### Claim 2 (depends on 1)

The clustering system according to claim 1, wherein the classifier estimates a cluster to which the target data belongs and which maximizes a conditional probability based on the function of the first variable for target data whose second variable is unknown in a state represented by the first variable and estimates the estimated cluster to be a cluster into which the target data is classified.

### Claim 3 (depends on 1)

The clustering system according to claim 1, wherein:
the function of the first variable is defined by a probabilistic model represented by a tree structure in which clusters into which the target data are classified is arranged in leaf nodes and in which the conditions of the first variable are arranged in condition nodes which are other than the leaf nodes; and
the classifier estimates a cluster optimizing the probabilistic model to be a cluster into which the target data is classified.

### Claim 4 (depends on 1)

The clustering system according to claim 1, wherein:
the clustering system classifies a target store into a store cluster generated based on sales;
the clustering system includes an input unit which receives an input of demographic information indicating the features of the target store as the first variable; and
the classifier classifies the target store indicated by the input demographic information into the store cluster based on a mixture model in which the mixing ratio is represented by a function of the demographic information and in which the element distribution of the store cluster is represented by a function of product sales data of the store which is the second variable.

### Claim 5 (depends on 4)

The clustering system according to claim 4, wherein:
the function of demographic information is defined by a probabilistic model represented by a tree structure in which the store clusters into which the target stores are classified are arranged in leaf nodes and in which the conditions of the demographic information are arranged in the condition nodes; and
the classifier estimates a cluster optimizing the probabilistic model to be a store cluster into which the target store is classified.

### Claim 6

A clustering method, wherein:
a computer classifies target data into a cluster based on a mixture model defined by using two different types of variables representing features of the target data; and
the computer classifies the target data into a cluster based on the mixture model in which a mixing ratio of the mixture mode is represented by a function of a first variable and in which an element distribution of the cluster into which the target data is classified is represented by a function of a second variable in the classification.

### Claim 7 (depends on 6)

The clustering method according to claim 6, wherein the computer estimates a cluster to which the target data belongs and which maximizes a conditional probability based on the function of the first variable for target data whose second variable is unknown in a state represented by the first variable and estimates the estimated duster to be a cluster into which the target data is classified.

### Claim 8

A non-transitory computer readable information recording medium storing a clustering program, when executed by a processor, which performs a method for:
classifying target data into a cluster based on a mixture model defined by using two different types of variables representing features of the target data; and
classifying the target data into a cluster based on the mixture mode in which a mixing ratio of the mixture model is represented by a function of a first variable and in which an element distribution of the cluster into which the target data is classified is represented by a function of a second variable.

### Claim 9 (depends on 8)

The non-transitory computer readable information recording medium according to claim 8, estimating a cluster to which the target data belongs and which maximizes a conditional probability based on the function of the first variable for target data whose second variable is unknown in a state represented by the first variable and estimating the estimated cluster to be a cluster into which the target data is classified.