Forum Data Clustering for Product Issue Identification

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing electronic forums lack a systematic way to identify and triage product issues, leading to a large number of redundant threads and making it difficult for organizations to proactively address customer concerns.

Innovation Solution

The proposed solution involves analyzing forum data by clustering related threads based on strong and weak relationships, using indicators such as keyword similarity and user activity, to identify significant product issues, which can then be addressed through targeted support and solutions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If users create new threads for every issue without systematic clustering, then users can easily post about their issues, but the number of redundant threads increases making issue identification intractable

Engineering Contradiction:

Improveease of postingVSAvoidthread management complexity

Core Design Contradiction:

Ease of operationVSDevice complexity

Solution Approach 1:

The system automatically clusters threads and identifies issues without requiring manual intervention from users or administrators. The clustering algorithm autonomously groups related threads based on semantic similarity and relationship indicators, enabling self-service issue identification that reduces manual management complexity while maintaining ease of posting.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces manual thread triaging and issue identification processes with automated computational clustering mechanisms. By using algorithms that analyze thread relationships, keywords, and metadata, the system substitutes human labor with automated systems to manage the complexity of large-scale thread datasets.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If forums become purely reactive venues without proactive issue identification, then users can freely discuss issues, but organizations cannot proactively address customer concerns

Engineering Contradiction:

Improveforum flexibilityVSAvoidissue resolution efficiency

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system performs preliminary analysis of forum threads to identify issues before they escalate or before organizations can proactively respond. By continuously monitoring and clustering threads to identify patterns and emerging issues, the system enables proactive issue identification and resolution while maintaining the flexible, user-driven nature of forum discussions.

Inventive Principle:

Principle #10Preliminary action

3Adaptability or versatility

If hundreds or thousands of threads are created for a smaller number of issues, then users can express diverse perspectives, but identifying and triaging particular issues becomes intractable

Engineering Contradiction:

Improvediscussion diversityVSAvoidtime for issue triaging

Core Design Contradiction:

Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent segments the large set of threads into meaningful clusters based on semantic similarity and relationship indicators. By dividing the data into manageable groups that represent distinct issues, the system reduces the complexity of triaging from analyzing thousands of individual threads to managing a smaller number of clustered issue groups, significantly reducing time for issue identification while preserving discussion diversity within each cluster.

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS8296278B2Identifying product issues using forum data

Publication Date: 2012.10.23 MICROSOFT TECHNOLOGY LICENSING LLC

AI summary

Product issues are identified through an analysis of forum data stored in a forum database. Forum threads are identified within the forum data and clustered together by grouping related forum threads. Once the forum threads have been clustered, the clustered forum threads can be analyzed to identify product issues. Once the product issues have been identified, steps may be taken in an attempt to resolve the identified issues.