# US 2019/0095823 A1 · Mining textual feedback

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

- Number: US 2019/0095823 A1
- Document ID: US20190095823A1
- Status: Published application
- Kind: A1
- Application number: 16/205,619
- Filed: 2018-11-30
- Published: 2019-03-28
- Earliest claimed: 2014-09-30
- Assignees: eBay Inc.
- Applicants: eBay Inc.
- Inventors: Samaneh Abbasi Moghaddam; Marco Pennacchiotti; Thomas Normile
- Web page: https://patentgrep.com/patent/US20190095823A1

## CPC

- G06F 16/355: Creation or modification of classes or clusters
- G06F 16/3322: {using system suggestions (G06F16/3325 takes precedence)}
- G06F 16/334: Query execution (filtering based on additional data G06F16/335)
- G06F 16/345: {Summarisation for human users}
- G06F 16/353: into predefined classes
- G06F 40/30: Semantic analysis
- G06N 20/00: Machine learning
- G06N 99/00: Subject matter not provided for in other groups of this subclass
- G06Q 30/0282: Rating or review of business operators or products
- 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

Methods, systems, and apparatus for mining feedback are described. A set of one or more lexical patterns associated with one or more of a suggestion and a defect report are determined and the set of one or more lexical patterns are matched against a plurality of feedback items to generate a distance learning training set. A distance learning technique is applied to the distance learning training set to generate a distance learning model and the distance learning model is used to identify one or more candidate feedback items of the plurality of feedback items, each of which is one or more of a candidate suggestion and a candidate defect report.

## Claims

### Claim 1

A system for processing a plurality of feedback items, the system comprising:
one or more hardware processors and a memory to store instructions that, when executed by the one or more hardware processors causes the one or more hardware processors to perform operations comprising:
obtaining at least one of a suggestion or a defect report;
determining a set of one or more lexical patterns associated with the at least one of the suggestion or the defect report;
identifying, using the set of one or more lexical patterns, one or more candidate feedback items of the plurality of feedback items that are one or more of a candidate suggestion and a candidate defect report; and
causing generation of a user interface to be displayed on a client device that includes a first pane and a second pane, wherein the first pane includes at least one selectable item corresponding to at least one topic, wherein receiving a selection of the at least one selectable item causes the second pane to display the identified one or more candidate feedback items that are grouped by the at least one topic.

### Claim 2 (depends on 1)

The system of claim 1, wherein the set of one or more lexical patterns are determined by applying a first machine-learning algorithm to a seed set of feedback items.

### Claim 3 (depends on 2)

The system of claim 2, wherein the first machine-learning algorithm is a bootstrapping algorithm.

### Claim 4 (depends on 1)

The system of claim 1, wherein the operations further comprise:
obtaining one or more of manually labeled relevant candidate suggestions and manually labeled relevant candidate defect reports.

### Claim 5 (depends on 4)

The system of claim 4, wherein the operations further comprise:
applying a first machine-learning technique to a subset of the relevant candidate suggestions to generate a classifying model for suggestions and applying a second machine-learning technique to a subset of the relevant candidate defect reports to generate a classifying model for defect reports.

### Claim 6 (depends on 5)

The system of claim 5, wherein the operations further comprise:
classifying each of one or more of the relevant candidate suggestions as one of a valid suggestion and an invalid suggestion using the classifying model for suggestions and classifying each of one or more of the relevant candidate defect reports as one of a valid defect report and an invalid defect report using the classifying model for defect reports.

### Claim 7 (depends on 1)

The system of claim 1, wherein the operations further comprise: identifying each of one or more snippets as one or more of a relevant suggestion snippet and a relevant defect snippet.

### Claim 8 (depends on 7)

The system of claim 7, wherein the operations further comprise:
applying a first snippet extraction machine-learning algorithm to a subset of the relevant suggestion snippets to generate a suggestion snippet extraction model; and
applying a second snippet extraction machine-learning algorithm to a subset of the relevant defect snippets to generate a defect snippet extraction model.

### Claim 9 (depends on 8)

The system of claim 8, wherein the operations further comprise:
applying the suggestion snippet extraction model to valid suggestions to generate one or more valid suggestion snippets; and
applying the defect snippet extraction model to valid defect reports to generate one or more valid defect report snippets.

### Claim 10 (depends on 9)

The system of claim 9, wherein the operations further comprise: grouping the valid suggestion snippets and the valid defect report snippets by type into one or more groups.

### Claim 11 (depends on 10)

The system of claim 10, wherein the operations further comprise: applying a topic modeling technique to identify a top number snippets for each group of the one or more groups of valid suggestion snippets and valid defect report snippets based on a probability of co-occurrence with other topics in the group.

### Claim 12 (depends on 11)

The system of claim 11, wherein the operations further comprise: generating a gist for each group that summarizes a topic for the group.

### Claim 13 (depends on 12)

The system of claim 12, wherein the operations further comprise: applying statistical measures and linguistic processing to create an explicative sentence that summarizes the topic associated with each group.

### Claim 14

A method for processing a plurality of feedback items, the method comprising:
obtaining at least one of a suggestion or a defect report;
determining, by at least one of one or more hardware processors, a set of one or more lexical patterns associated with the at least one of the suggestion or the defect report;
identifying, using the set of one or more lexical patterns, one or more candidate feedback items of the plurality of feedback items that are one or more of a candidate suggestion and a candidate defect report; and
causing generation of a user interface to be displayed on a client device that includes a first pane and a second pane, wherein the first pane includes at least one selectable item corresponding to at least one topic, wherein receiving a selection of the at least one selectable item causes the second pane to display the identified one or more candidate feedback items that are grouped by the at least one topic.

### Claim 15 (depends on 14)

The method of claim 14, wherein the set of one or more lexical patterns are manually determined.

### Claim 16 (depends on 14)

The method of claim 14, wherein the set of one or more lexical patterns are determined by applying a first machine-learning algorithm to a seed set of feedback items.

### Claim 17 (depends on 14)

The method of claim 14, further comprising: manually labeling each of the one or more candidate feedback items as one or more of a relevant candidate suggestion and a relevant candidate defect report.

### Claim 18 (depends on 17)

The method of claim 17, further comprising:
applying a first machine-learning technique to a subset of the relevant candidate suggestions to generate a classifying model for suggestions and applying a second machine-learning technique to a subset of the relevant candidate defect reports to generate a classifying model for defect reports.

### Claim 19 (depends on 18)

The method of claim 18, further comprising:
classifying each of one or more of the relevant candidate suggestions as one of a valid suggestion and an invalid suggestion using the classifying model for suggestions and classifying each of one or more of the relevant candidate defect reports as one of a valid defect report and an invalid defect report using the classifying model for defect reports.

### Claim 20

A non-transitory computer-readable medium embodying instructions that, when executed by one or more processors of a machine, cause the machine to perform operations comprising:
obtaining at least one of a suggestion or a defect report;
determining a set of one or more lexical patterns associated with at least one of the suggestion or the defect report;
identifying, using the set of one or more lexical patterns, one or more candidate feedback items of the plurality of feedback items that are one or more of a candidate suggestion and a candidate defect report; and
causing generation of a user interface to be displayed on a client device that includes a first pane and a second pane, wherein the first pane includes at least one selectable item corresponding to at least one topic, wherein receiving a selection of the at least one selectable item causes the second pane to display the identified one or more candidate feedback items that are grouped by the at least one topic.
