Machine learning based conformal predictors for uncertainty-aware compressive strength estimation of concrete - Research

Aug 4, 2026

Estimating concrete compressive strength is crucial for predicting performance, optimising material usage, and ensuring durability and safety. Deterministic machine learning models often ignore uncertainty, although concrete is non-homogeneous with complex variability that makes precise strength prediction difficult. This work proposes an uncertainty quantification framework using conformal prediction: eight ML models combined with six conformal variants to build statistically rigorous prediction intervals. A novel Efficiency Score evaluates trade-offs between empirical coverage and interval width, identifying LightGBM with Jackknife+ as most efficient, and shows adaptation to heteroscedasticity in higher-grade concrete.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages14

Reading Time35 min

AccessFree to download

Introduction

Importance of compressive strength in concrete quality

Limitations of code-based testing and deterministic ML

Prior modelling approaches (regression and physics-based)

Motivation for uncertainty-aware ML with conformal prediction

POLITECNICO DI TORINO Repository ISTITUZIONALE Machine learning based conformal predictors for uncertainty-aware compressive strength estimation of concrete Original Machine learning based conformal predictors for uncertainty-aware compressive strength estimation of concrete / Tamuly, Pranjal; Nava, Vincenzo. -In: CONSTRUCTION AND BUILDING MATERIALS. -ISSN 0950-0618. -487:(2025) .[10.1016/j.conbuildmat.2025.141844] Availability: This version is available at: 11583/3000786 since: 2025-06-09T12:18:09Z Publisher: Elsevier Ltd Published DOI:10.1016/j.conbuildmat.2025.141844 Terms of use: This article is made available under terms and conditions as specified in the corresponding bibliographic description in the repository Publisher copyright (Article begins on next page) 04 October 2025 Construction and Building Materials 487 (2025) 141844 | Machine learning based conformal predictors for uncertainty-aware compressive strength estimation of concrete
Pranjal Tamulya ,∗, Vincenzo Nava b
a Basque Center for Applied Mathematics, Alameda de Mazarredo, 14, Bilbao, 48009, Spain b Politecnico di Torino, Corso Duca degli Abruzzi, 24, Turin, 19129, Italy | | | | --- | --- | --- | | A R T I C L E I N F O | | A B S T R A C T | | Keywords:
Conformal predictor Machine learning Compressive strength Uncertainty quantification Prediction interval | | Estimating concrete compressive strength is crucial for accurately predicting its performance, optimising material usage, and ensuring the durability and safety of the structure. Traditional machine learning (ML) models have primarily focused on deterministic predictions of compressive strength, often overlooking the uncertainty associated with these estimates. However, concrete is a non-homogeneous material with complex and variable behaviour, making it inherently difficult to predict compressive strength with precision. Therefore, incorporating uncertainty into predictive modelling is essential for producing more reliable and practical results in real-world engineering applications. This study addresses this gap by proposing a comprehensive framework for uncertainty quantification in concrete strength estimation using conformal prediction methods. In this comprehensive study, eight distinct machine learning models are systematically integrated with six conformal prediction variants to construct statistically rigorous prediction intervals. To evaluate the performance of the models holistically in engineering contexts, a novel Efficiency Score (ES) is proposed, combining empirical coverage, mean interval width, and point prediction accuracy. The findings reveal notable trade-offs between predicted interval width and empirical coverage across the model spectrum. Among the tested combinations, LightGBM coupled with Jackknife+ emerges as the most effective configuration, demonstrating the highest efficiency score. Additionally, conformal predictors exhibit satisfactory adaptation to heteroscedasticity, which arises in the predictions of higher-grade concrete (> 40 MPa). Thus, the proposed framework empowers more informed decision-making in concrete design and quality control by providing robust uncertainty bounds advancing beyond traditional deterministic point predictions to support risk-aware infrastructure development. | 1. Introduction Compressive strength serves as the primary indicator of concrete quality, influencing the safety, durability, and performance of infrastructure projects. The complex behaviour of concrete depends upon numerous interrelated factors, including cement type, water-cement ratio, aggregate properties, admixtures, curing conditions, and age. Traditionally, the compressive strength of concrete is determined according to structural design codes, involving the casting and curing of specimens, followed by testing their compressive strength after a specified period. However, this process is highly inefficient as it is timeconsuming, l

FAQs

Why is uncertainty-aware estimation of concrete compressive strength important?

Concrete strength prediction directly impacts structural safety, durability, and material efficiency. Because concrete is non-homogeneous and variable, deterministic predictions can be unreliable when uncertainty is not quantified.

How does the study construct prediction intervals?

It combines eight machine learning models with six conformal prediction variants to produce statistically rigorous prediction intervals that quantify uncertainty.

What method is reported as the most effective configuration?

LightGBM coupled with Jackknife+ shows the highest Efficiency Score, balancing empirical coverage, interval width, and point prediction accuracy.

Aug 4, 2026

Upload by Seraphina