Automated Question Mining via Keyword Co-occurrence

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current question mining methods for intelligent customer service systems face challenges in generating high-quality, diverse, and logically accurate question texts due to limitations in data augmentation methods like EDA and BERT-related models, which require human intervention and result in semantic errors or insufficient generalization.

Innovation Solution

An automated question mining method that utilizes a pre-built standard question database to mine keywords based on importance and co-occurrence information, allowing for the generation of target question texts that accurately reflect intent categories without human participation, thereby expanding the standard question database efficiently.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If data augmentation methods like EDA and BERT-related models are used to generate question texts, then the diversity of question texts is improved, but semantic errors occur and human intervention is required

Engineering Contradiction:

Improvediversity of question textsVSAvoidsemantic accuracy

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

The system automatically mines keywords from the standard question database and generates target question texts without requiring human intervention. The keyword mining module autonomously identifies important words based on importance degrees, and the question text generation module automatically creates diverse question texts by combining keywords with co-occurrence words, eliminating the need for manual verification while maintaining semantic accuracy.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent introduces an intermediary keyword mining process between the standard question database and the generated question texts. By extracting keywords and their co-occurrence relationships as intermediate representations, the system ensures that generated questions maintain semantic coherence with the original intent categories while achieving diversity through various combination strategies.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If manual question mining is performed to ensure high quality, then the accuracy of intent categorization is improved, but the workload for users increases

Engineering Contradiction:

Improveaccuracy of intent categorizationVSAvoidworkload for users

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system performs automatic keyword mining and question text generation without requiring user intervention. The keyword mining module automatically calculates importance degrees of words in the standard question database, identifies keywords and co-occurrence words, and generates target question texts autonomously, completely eliminating the manual workload while maintaining high accuracy in intent categorization.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical process of question mining with an automated computational system. Instead of users manually analyzing and creating question texts, the system uses algorithms to mine keywords, calculate co-occurrence relationships, and generate diverse question texts automatically, substituting human effort with automated processing while preserving or improving accuracy.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If automated question generation is implemented, then the productivity is improved, but the semantic generalization ability may be insufficient

Engineering Contradiction:

Improvequestion generation efficiencyVSAvoidsemantic generalization ability

Core Design Contradiction:

ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent introduces keywords and co-occurrence words as intermediary elements that bridge the standard question database and generated question texts. By mining keywords based on importance degrees and identifying their co-occurrence relationships, the system ensures that generated questions maintain strong semantic connections to the original intent categories, achieving both high productivity and semantic generalization ability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system dynamically adjusts the composition of generated question texts by varying the combination of keywords and co-occurrence words. By changing parameters such as the number of co-occurrence words included and the selection criteria for keyword combinations, the system generates diverse question texts with different semantic nuances while maintaining generalization ability across various intent categories.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250013675A1Question mining method, electronic device, and non-transiroty storage media

Publication Date: 2025.01.09 MASHANG CONSUMER FINANCE CO LTD

AI summary

A question-mining method includes obtaining a pre-built standard question database, where the standard question database includes a first standard question text, the first standard question text corresponds to a first intent category, and the first standard question text comprises a plurality of words; mining keywords of the first intent category from the plurality of words according to an importance degree of each word of the first standard question text to the first intent category, wherein the plurality of words include the keywords and non-keywords; determining a co-occurrence word of the keywords according to co-occurrence information of the keywords and the non-keywords in the standard question database; and mining a target question text from a pre-obtained target text set according to the co-occurrence word of the keywords.