# US 2021/0174239 A1 · Training classifiers in machine learning

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

- Number: US 2021/0174239 A1
- Document ID: US20210174239A1
- Status: Published application
- Kind: A1
- Application number: 16/703,371
- Filed: 2019-12-04
- Published: 2021-06-10
- Earliest claimed: 2019-12-04
- Assignees: International Business Machines Corporation
- Applicants: International Business Machines Corporation
- Inventors: Qingzi Liao; Yunfeng Zhang; Michael Desmond; Rachel Katherine Emma Bellamy
- Web page: https://patentgrep.com/patent/US20210174239A1

## CPC

- G06N 20/00: Machine learning
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS

## Abstract

An approach is provided for training classifiers used in machine learning. A corpus of training data is received. One or more clusters of the training data is generated according to features of the training data. The one or more clusters are refined using user-specified rules. One or more classifiers are trained for use in machine learning based upon the refined one or more clusters.

## Claims

### Claim 1

A method for training classifiers used in machine learning, the method comprising:
receiving, by one or more processors of a computer system, a corpus of training data;
generating, by the one or more processors of the computer system, one or more clusters of the training data according to features of the training data;
refining, by the one or more processors of the computer system, the one or more clusters using user-specified rules; and
training, by one or more processors of a computer system, one or more classifiers for use in machine learning based upon the refined one or more clusters.

### Claim 2 (depends on 1)

The method of claim 1, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are authored by the user.

### Claim 3 (depends on 1)

The method of claim 1, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are automatically extracted and suggested to users for selection, confirmation and editing.

### Claim 4 (depends on 1)

The method of claim 1, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are elicited case-based.

### Claim 5 (depends on 1)

The method of claim 1, wherein the refining further includes:
assigning subsets of the corpus of training data in or out of clusters of the one or more cluster using the user-specified rules; and
using a voting algorithm or a generative model to resolve conflicts in the assigning.

### Claim 6 (depends on 1)

The method of claim 1, wherein the refining and the training are implemented by the one or more processors of the computer system in a single process.

### Claim 7 (depends on 1)

The method of claim 1, wherein the one or more clusters represent classes of the training data.

### Claim 8

A computer system, comprising:
one or more processors;
one or more memory devices coupled to the one or more processors; and
one or more computer readable storage devices coupled to the one or more processors, wherein the one or more storage devices contain program code executable by the one or more processors via the one or more memory devices to implement a method for training classifiers in machine learning, the method comprising:
receiving, by one or more processors of a computer system, a corpus of training data;
generating, by the one or more processors of the computer system, one or more clusters of the training data according to features of the training data;
refining, by the one or more processors of the computer system, the one or more clusters using user-specified rules; and
training, by one or more processors of a computer system, one or more classifiers for use in machine learning based upon the refined one or more clusters.

### Claim 9 (depends on 8)

The computer system of claim 8, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are authored by the user.

### Claim 10 (depends on 8)

The computer system of claim 8, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are automatically extracted and suggested to users for selection, confirmation and editing.

### Claim 11 (depends on 8)

The computer system of claim 8, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are elicited case-based.

### Claim 12 (depends on 8)

The computer system of claim 8, wherein the refining further includes:
assigning subsets of the corpus of training data in or out of clusters of the one or more cluster using the user-specified rules; and
using a voting algorithm or a generative model to resolve conflicts in the assigning.

### Claim 13 (depends on 8)

The computer system of claim 8, wherein the refining and the training are implemented by the one or more processors of the computer system in a single process.

### Claim 14 (depends on 8)

The computer system of claim 8, wherein the one or more clusters represent classes of the training data.

### Claim 15

A computer program product, comprising one or more computer readable hardware storage devices storing a computer readable program code, the computer readable program code comprising an algorithm that when executed by one or more processors of a computing system implements a method for training classifiers in machine learning, the method comprising:
receiving, by one or more processors of a computer system, a corpus of training data;
generating, by the one or more processors of the computer system, one or more clusters of the training data according to features of the training data;
refining, by the one or more processors of the computer system, the one or more clusters using user-specified rules; and
training, by one or more processors of a computer system, one or more classifiers for use in machine learning based upon the refined one or more clusters.

### Claim 16 (depends on 15)

The computer program product of claim 15, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are authored by the user.

### Claim 17 (depends on 15)

The computer program product of claim 15, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are automatically extracted and suggested to users for selection, confirmation and editing.

### Claim 18 (depends on 15)

The computer program product of claim 15, wherein the refining further includes:
interacting with a user with a cluster refining rule elicitation module, wherein the user-specified rules are elicited case-based.

### Claim 19 (depends on 15)

The computer program product of claim 15, wherein the refining further includes:
assigning subsets of the corpus of training data in or out of clusters of the one or more cluster using the user-specified rules; and
using a voting algorithm or a generative model to resolve conflicts in the assigning.

### Claim 20 (depends on 15)

The computer program product of claim 15, wherein the refining and the training are implemented by the one or more processors of the computer system in a single process.
