MétaCan
Menu
Back to cohort
Record W4414015740 · doi:10.11159/cist25.131

A Case Study on Fault Type Classification and Action Prediction Using Unstructured Text Data from Automotive Parts Manufacturers

2025· article· en· W4414015740 on OpenAlex

Why this work is in the frame

A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.

venuePublished in a venue whose home country is Canada.
no affNo Canadian affiliation: this work is invisible to an affiliation-only frame.
No Canadian affiliation. An affiliation-only frame, the usual design, would never have seen this work. It is one of the works that make the case for inverting the frame.

Bibliographic record

VenueProceedings of the World Congress on Electrical Engineering and Computer Systems and Science · 2025
Typearticle
Languageen
FieldComputer Science
TopicTechnology and Data Analysis
Canadian institutionsnot available
FundersMinistry of SMEs and Startups
KeywordsAutomotive industryUnstructured dataComputer scienceFault (geology)Type (biology)Action (physics)Data miningArtificial intelligenceNatural language processingBig dataEngineering

Abstract

fetched live from OpenAlex

Effective equipment management in manufacturing requires accurately identifying malfunctions and defects by analyzing their causes and patterns.This involves classifying failure types, monitoring trends, and implementing appropriate resolutions to improve operational efficiency.Precise classification plays a crucial role in enabling data-driven decisionmaking and proactive maintenance.Manufacturing Execution Systems (MES) collect extensive data on equipment failures, incorporating both structured machine-generated logs and unstructured operator-reported text, providing valuable insights into failure mechanisms and responses.This study utilizes real-world manufacturing data to develop machine learning models for classifying equipment failure types and recommending response strategies.Specifically, classification models based on Bidirectional Encoder Representations from Transformers (BERT) and Random Forest were trained to identify ten distinct failure categories.Additionally, T5-based models were constructed and optimized to generate effective response strategies.A comprehensive performance evaluation was conducted to assess the effectiveness of these models in real-world applications.Natural Language Processing (NLP), a key component of artificial intelligence, enables the extraction of insights from unstructured text, bridging the gap between human language and computational analysis.While early NLP approaches relied on statistical methods, recent advances in deep learning have significantly enhanced its capabilities.These advancements are particularly valuable in manufacturing, where unstructured textual data is abundant but often underutilized in failure analysis.Equipment downtime, a major challenge in manufacturing, is categorized as either planned or unplanned.Unplanned downtime is particularly problematic due to its unpredictability and negative impact on productivity.Effective management requires integrating structured and unstructured data to diagnose failure causes and prioritize resolutions efficiently.This research proposes a deep learning-based framework that leverages unstructured failure logs from an automotive parts manufacturing facility to classify equipment failure types and predict optimal response strategies.By integrating advanced NLP and deep learning techniques, this study introduces a standardized, data-driven approach to equipment failure classification and resolution.The proposed methodology has the potential to reduce unplanned downtime and enhance manufacturing efficiency, contributing to the development of intelligent maintenance systems applicable across various industrial environments.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesnone
Consensus categoriesnone
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Simulation or modeling · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.933
Threshold uncertainty score0.355

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.001
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0010.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0000.000

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.032
GPT teacher head0.273
Teacher spread0.241 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it