Search research articles

Search research articles

Related Experiment Video

Updated: Oct 20, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Towards Interpretable Deep Learning: A Feature Selection Framework for Prognostics and Health Management Using Deep

Joaquín Figueroa Barraza¹, Enrique López Droguett², Marcelo Ramos Martins¹

¹LabRisco-Analysis, Evaluation and Risk Management Laboratory, Department of Naval Architecture and Ocean Engineering, University of São Paulo, São Paulo 05508-030, Brazil.

Sensors (Basel, Switzerland)

|September 10, 2021

Summary

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Exploiting the Capabilities of Bayesian Networks for Engineering Risk Assessment: Causal Reasoning through Interventions.

Risk analysis : an official publication of the Society for Risk Analysis·2021

Same author

A Methodology for Assessing the Probability of Occurrence of Undesired Events in the Tietê-Paraná Inland Waterway Based on Expert Opinion.

Risk analysis : an official publication of the Society for Risk Analysis·2020

See all related articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

This study introduces a novel feature selection technique for deep learning models in prognostics and health management (PHM). The method enhances model interpretability without sacrificing performance, improving deployment readiness.

Area of Science:

Engineering
Computer Science

Background:

Deep Learning (DL) models have advanced prognostics and health management (PHM) for diagnostics, prognostics, and anomaly detection.
A key challenge is the interpretability tradeoff, where complex DL models offer high accuracy but lack transparency, hindering deployment.

Purpose of the Study:

To address the accuracy/interpretability tradeoff in DL-based PHM.
To propose an embedded feature selection (FS) technique within deep neural networks for PHM.

Main Methods:

Developed a novel FS layer integrated into deep neural networks.
Trained the FS layer concurrently with the network to assess input feature importance.
Introduced a new metric, the ranking quality score (RQS), to evaluate performance evolution based on feature ranking.

Keywords:

deep learning deep neural networks feature selection interpretable AI prognostics and health management

More Related Videos

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

Related Experiment Videos

Last Updated: Oct 20, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches

Published on: December 15, 2023

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Author Spotlight: Impact of Intergenic Interactions on Disease-Identifying Dark Biomarkers

Published on: March 1, 2024

Main Results:

The proposed embedded FS technique achieved higher RQS compared to other methods.
The technique maintained performance levels consistent with DL models lacking the FS layer.
Demonstrated effectiveness across three case studies in health state diagnostics, prognostics, and remaining useful life prediction.

Conclusions:

The proposed embedded FS technique effectively balances accuracy and interpretability in DL-based PHM.
This approach facilitates the deployment of more transparent and reliable PHM models.
The RQS metric provides a valuable tool for evaluating feature selection strategies in PHM.