Relation between prognostics predictor evaluation metrics and local interpretability SHAP values

2022/02/15 by Marcia L. Baptista, Márcia L. Baptista, Kai Goebel +2 · 267 citations
Computer Science · Decision Sciences · Mathematics · #Adversarial Robustness in Machine Learning #Artificial intelligence #Artificial neural network #Computer science #Data mining #Explainable Artificial Intelligence (XAI) #Field (mathematics) #Forecasting Techniques and Applications #Interpretability #Machine learning #Mathematics #Perceptron #Prognostics #Set (abstract data type)

paper · pdf · doi:10.1016/j.artint.2022.103667

published in Artificial Intelligence 306, 103667 (Elsevier BV)

openalex publication_date 2022/02/15 · openalex created_date 2022/02/24 · openalex updated_date 2026/08/05

Abstract

Maintenance decisions in domains such as aeronautics are becoming increasingly dependent on being able to predict the failure of components and systems. When data-driven techniques are used for this prognostic task, they often face headwinds due to their perceived lack of interpretability. To address this issue, this paper examines how features used in a data-driven prognostic approach correlate with established metrics of monotonicity, trendability, and prognosability. In particular, we use the SHAP model (SHapley Additive exPlanations) from the field of eXplainable Artificial Intelligence (XAI) to analyze the outcome of three increasingly complex algorithms: Linear Regression, Multi-Layer Perceptron, and Echo State Network. Our goal is to test the hypothesis that the prognostics metrics correlate with the SHAP model's explanations, i.e., the SHAP values. We use baseline data from a standard data set that contains several hundred run-to-failure trajectories for jet engines. The results indicate that SHAP values track very closely with these metrics with differences observed between the models that support the assertion that model complexity is a significant factor to consider when explainability is a consideration in prognostics.

Citations

• European Union regulations on algorithmic decision-making and a "right to explanation"

Cited by

• Semi-supervised and unsupervised learning for health indicator extraction from guided waves in aerospace composite structures
• Study of Robust Features in Formulating Guidance for Heuristic Algorithms for Solving the Vehicle Routing Problem
• Keeping Medical AI Healthy and Trustworthy: A Review of Detection and Correction Methods for System Degradation
• An Unsupervised Framework for Dynamic Health Indicator Construction and Its Application in Rolling Bearing Prognostics
• Reservoir Dynamic Interpretability for Time Series Prediction: A Permutation Entropy View. [europepmc]
• A Decision Support System for Diagnosis of COVID-19 from Non-COVID-19 Influenza-like Illness Using Explainable Artificial Intelligence. [europepmc]
• An Unsupervised Machine Learning Approach for Monitoring Data Fusion and Health Indicator Construction. [europepmc]
• Machine Learning Algorithms Predict Successful Weaning From Mechanical Ventilation Before Intubation: Retrospective Analysis From the Medical Information Mart for Intensive Care IV Database. [europepmc]
• Enhanced SARS-CoV-2 case prediction using public health data and machine learning models. [europepmc]
• Explainable prediction of node labels in multilayer networks: a case study of turnover prediction in organizations. [europepmc]
• Predicting physical functioning status in older adults: insights from wrist accelerometer sensors and derived digital biomarkers of physical activity. [europepmc]
• Cheminformatic Identification of Tyrosyl-DNA Phosphodiesterase 1 (Tdp1) Inhibitors: A Comparative Study of SMILES-Based Supervised Machine Learning Models. [europepmc]
• Multiple imputation integrated to machine learning: predicting post-stroke recovery of ambulation after intensive inpatient rehabilitation. [europepmc]
• Conversion, prediction, and application of strength and stiffness parameters for grouted reconstructed rock mass. [europepmc]
• Prediction of benign and malignant pulmonary nodules using preoperative CT features: using PNI-GARS as a predictor. [europepmc]
• Salmon Consumption Behavior Prediction Based on Bayesian Optimization and Explainable Artificial Intelligence. [europepmc]
• A stacked ensemble machine learning model for the prediction of pentavalent 3 vaccination dropout in East Africa. [europepmc]
• TransXLT: A novel ZTD prediction method with SASR-based data reconstruction. [europepmc]
• Artificial Intelligence-Driven Strategies for Targeted Delivery and Enhanced Stability of RNA-Based Lipid Nanoparticle Cancer Vaccines. [europepmc]
• Organic photovoltaic prediction model based on Bayesian optimization and explainable AI. [europepmc]
• Assessing wildfire extents in Siberian forests using machine learning. [europepmc]
• Development, validation and clinical utility of prognostic models for patients with traumatic brain injury in an Indian setting using machine learning and traditional approaches: a study protocol. [europepmc]
• Development of the machine learning and deep learning models with SHAP strategy for predicting groundwater levels in South Korea. [europepmc]
• eXplainable Artificial Intelligence (XAI): A Systematic Review for Unveiling the Black Box Models and Their Relevance to Biomedical Imaging and Sensing. [europepmc]
• Joint association between pan-immune inflammation value and physical pain with self-perception of patients in rheumatoid arthritis: a retrospective cohort study employing traditional statistics and interpretable machine learning. [europepmc]
• Artificial intelligence-driven anticancer peptide discovery. [europepmc]
• An explainable AI-driven hybrid feature selection approach for coronary artery disease diagnosis. [europepmc]
• Camping in urban parks as a public health-oriented therapeutic landscape experience accomplished through self-transformation. [europepmc]
• Childhood living conditions as predictors of self-rated health status in middle-aged and older adults: Evidence from a machine learning analysis in 27 high-income countries [europepmc]

Related

• Interpretable Prognostics with Concept Bottleneck Models
• Fusing Physics-based and Deep Learning Models for Prognostics
• Toward Decision-Oriented Prognostics: An Integrated Estimate-Optimize Framework for Predictive Maintenance
• Generalization bounds and sample complexity for remaining useful life prediction from complete degradation trajectories
• I-GLIDE: Input Groups for Latent Health Indicators in Degradation Estimation