# US 2025/0053747 A1 · Classification of user sentiment based on machine learning

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

- Number: US 2025/0053747 A1
- Document ID: US20250053747A1
- Status: Published application
- Kind: A1
- Application number: 18/928,557
- Filed: 2024-10-28
- Published: 2025-02-13
- Earliest claimed: 2020-06-18
- Assignees: Home Depot Product Authority, LLC
- Applicants: Home Depot Product Authority, LLC
- Inventors: Haozheng Tian; James Morgan White
- Web page: https://patentgrep.com/patent/US20250053747A1

## CPC

- G06F 40/30: Semantic analysis
- G06F 40/247: Thesauruses; Synonyms
- G06N 3/0442: characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
- G06N 3/045: Combinations of networks
- G06N 3/08: Learning methods
- G06N 3/09: Supervised learning
- G06V 30/19093: {Proximity measures, i.e. similarity or distance measures}
- G06V 30/19107: {Clustering techniques}
- G06F 40/268: Morphological analysis
- G06F 40/284: Lexical analysis, e.g. tokenisation or collocates
- G06N 20/00: Machine learning
- G06F: ELECTRIC DIGITAL DATA PROCESSING (computer systems based on specific computational models G06N)
- G06N: COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06V: IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING

## Abstract

A system and method for machine learning classification of user sentiment is disclosed. The method includes storing including a plurality of category information. The plurality of category information includes a set of domain-specific category information. The method further includes extracting a plurality of aspects from textual data. The method further includes generating a sentiment by a machine learning model. The method further includes receiving the plurality of aspects and the set of domain-specific category information. The method further includes generating a sentiment based on the plurality of aspects and the set of domain-specific category information.

## Claims

### Claim 1

A system for performing classification of user sentiment, the system comprising:
a processor; and
a non-transitory computer readable media having stored thereon instructions that are executable by the processor to perform operations comprising:
obtain a plurality of category information including a set of domain-specific category information based on text data;
determine a plurality of clusters from the text data, each cluster of the plurality of clusters comprising one or more words from the text data;
extract a plurality of aspects from the plurality of clusters, the extracting comprising:
determine a most frequent word as an explicit aspect of the one or more words of a cluster of the plurality of clusters,
determine a part of speech associated with the one or more words of the cluster,
responsive to the part of speech, determine a set of synonyms of the one or more words of the cluster associated with an adjective part of speech,
responsive to the adjective part of speech, generate lemmatized noun forms of the set of synonyms of the cluster,
determine a match of the lemmatized noun forms with the explicit aspect of the cluster, and
select the explicit aspect based on the match having a greatest number of occurrences to represent an aspect of an implicit cluster, and
generate, by a machine learning model, a sentiment based on the plurality of aspects and the set of domain-specific category information,
wherein the plurality of aspects comprises implicit aspects and explicit aspects of the plurality of clusters.

### Claim 2 (depends on 1)

The system of claim 1, further comprising:
a sentiment analysis module configured for communication with a search interface,
wherein the search interface being configured to receive user inputs corresponding to the text data.

### Claim 3 (depends on 2)

The system of claim 2, the operations further comprising:
obtain the user inputs entered at the search interface as the text data; and
update the plurality of category information based on the sentiment and the plurality of aspects.

### Claim 4 (depends on 2)

The system of claim 2, further comprising:
a data store comprising the plurality of category information including the set of domain-specific category information, and wherein the plurality of category information and the set of domain-specific category information being obtained from the data store based on the text data.

### Claim 5 (depends on 2)

The system of claim 2, wherein the sentiment analysis module comprising:
an explicit extraction engine,
wherein the explicit extraction engine being configured to extract the explicit aspects from the plurality of clusters.

### Claim 6 (depends on 5)

The system of claim 5, wherein the sentiment analysis module further comprising:
an implicit extraction engine,
wherein the implicit extraction engine being configured to extract the implicit aspects from the plurality of clusters.

### Claim 7 (depends on 1)

The system of claim 1, wherein the machine learning model comprises:
a gate mechanism,
wherein the gate mechanism is configured to regulate flow of the plurality of category information in the machine learning model, wherein the gate mechanism filters a portion of the plurality of category information and passes the set of domain-specific category information.

### Claim 8 (depends on 7)

The system of claim 7, wherein the gate mechanism is further configured to filter the portion of the plurality of category information and passes a subset of domain-specific category information based on a weight value applied to the set of domain-specific category information.

### Claim 9 (depends on 8)

The system of claim 8, wherein an output of the gate mechanism comprises one or more bins of equal range, each of the one or more bins having a respective value range representative of the portion of the plurality of category information corresponding to the subset of domain-specific category information that is passed through the gate mechanism.

### Claim 10

A method for machine learning classification of user sentiment, the method comprising:
obtaining a plurality of category information including a set of domain-specific category information based on text data;
determining a plurality of clusters from the text data, each cluster of the plurality of clusters comprising one or more words from the text data;
extracting a plurality of aspects from the plurality of clusters, the extracting comprising:
determining a most frequent word as an explicit aspect for the one or more words of a cluster,
determining a part of speech associated with the one or more words of the cluster,
responsive to the part of speech, determining a set of synonyms of the one or more words of the cluster associated with an adjective part of speech,
responsive to the adjective part of speech, generating lemmatized noun forms of the set of synonyms of the cluster,
determining a match of the lemmatized noun forms with the explicit aspect of the cluster, and
selecting the explicit aspect based on the match having a greatest number of occurrences to represent an aspect of an implicit cluster; and
generating, by a machine learning model, a sentiment based on the plurality of aspects and the set of domain-specific category information,
wherein the plurality of aspects comprises implicit aspects and explicit aspects of the plurality of clusters.

### Claim 11 (depends on 10)

The method of claim 10, further comprising:
obtaining user inputs entered at a search interface as the text data; and
updating the plurality of category information based on the sentiment and the plurality of aspects.

### Claim 12 (depends on 10)

The method of claim 10, wherein the plurality of category information and the set of domain-specific category information being obtained from a data store based on the text data.

### Claim 13 (depends on 10)

The method of claim 10, wherein an explicit extraction engine being configured to extract the explicit aspects from the plurality of clusters.

### Claim 14 (depends on 13)

The method of claim 13, wherein an implicit extraction engine being configured to extract the implicit aspects from the plurality of clusters.

### Claim 15 (depends on 10)

The method of claim 10, wherein generating the sentiment based on the plurality of aspects and the set of domain-specific category information further comprising:
filtering, by a gate mechanism, the plurality of category information in the machine learning model,
wherein the gate mechanism is configured to apply a weight to the set of domain-specific category information, and wherein the gate mechanism passes a subset of domain-specific category information based on the applied weight.

### Claim 16 (depends on 15)

The method of claim 15, wherein an output of the gate mechanism comprises one or more bins of equal range, each of the one or more bins having a respective value range representative of a portion of the plurality of category information corresponding to the subset of domain-specific category information that is passed through the gate mechanism.

### Claim 17

A method for gated machine learning of user sentiment classification, the method comprising:
obtaining a plurality of category information including a set of domain-specific category information based on text data;
determining a plurality of clusters from the text data, each cluster of the plurality of clusters comprising one or more words from the text data;
determining a most frequent word as an explicit aspect for the one or more words of a cluster;
determining a part of speech associated with the one or more words of the cluster;
responsive to the part of speech, determining a set of synonyms of the one or more words of the cluster associated with an adjective part of speech;
responsive to the adjective part of speech, generating lemmatized noun forms of the set of synonyms of the cluster;
determining a match of the lemmatized noun forms with the explicit aspect of the cluster;
selecting the explicit aspect based on the match having a greatest number of occurrences to represent an aspect of an implicit cluster;
extracting a plurality of aspects corresponding to implicit aspects and explicit aspects from the plurality of clusters;
filtering, by a gate mechanism, the plurality of category information based on a weight applied to the plurality of category information;
passing, by the gate mechanism, a portion of the plurality of category information corresponding to the set of domain-specific category information based on the applied weight; and
generating, by a machine learning model, a sentiment based on the plurality of aspects and the set of domain-specific category information.

### Claim 18 (depends on 17)

The method of claim 17, further comprising:
classifying the one or more words of a cluster of the plurality of clusters;
obtaining the set of domain-specific category information based on the classification of the one or more words;
associating, by the gate mechanism, gate values with the set of domain-specific category information based on contextual information of the one or more words;
applying, by the gate mechanism, the weight to the gate values associated with the set of domain-specific category information;
filtering, by the gate mechanism, the set of domain-specific category information in the machine learning model based on the weights applied to the gate values;
updating the plurality of category information based on the sentiment and the plurality of aspects;
updating the weight of the gate mechanism based on the sentiment and the plurality of aspects,
wherein the gate mechanism passes a subset of the domain-specific category information based on the weights applied to the gate values of the set of domain-specific category information.

### Claim 19 (depends on 17)

The method of claim 17, wherein an output of the gate mechanism comprises one or more bins of equal range, each of the one or more bins having a respective value range representative of a portion of the plurality of category information corresponding to the subset of the domain-specific category information that is passed through the gate mechanism.

### Claim 20 (depends on 17)

The method of claim 17, wherein the plurality of aspects comprises implicit aspects and explicit aspects of the plurality of clusters.
