"Can we replicate your findings?" Assessing the availability of reproducibility assets in machine learning-based concrete research

Abstract Research on machine learning (ML)-based concrete property prediction has expanded significantly, yet the availability and accessibility of research assets needed to support independent verification remain insufficiently characterized. This study investigated 419 articles indexed in the Science Citation Index Expanded (SCIE) of the Web of Science Core Collection that were retrieved with the WoS Publication Year filter set to 2025 and focused on tabular data-based ML studies predicting concrete properties from mixture-related variables. The availability and accessibility of datasets, source code, and graphical user interfaces (GUIs) were manually verified and further analyzed by journal quartile and corresponding author country. The empirical findings reveal a profoundly low asset-sharing rate; algorithmic source codes remained undisclosed in 90.7% of the literature. For datasets, full raw data sharing stood at only 21.5%, whereas 50.4% relied on external citations. Furthermore, an analysis of 85 GUI-developed publications showed that 66.0% and 53.6% of instances in Q1 and Q2 journals, respectively, exhibited a developed-but-restricted access pattern. An exploratory review of Data Availability Statements identified several stated reasons for restricting asset sharing, including institutional intellectual property protection and industrial confidentiality agreements, and noted several statements that appeared inconsistent with the data-dependent nature of the reported ML analyses.

Authors

Institutions

Publication Details

Journal
Research Integrity and Peer Review
Published
2026-11-17
DOI
https://doi.org/10.1186/s41073-026-00266-3
Primary Topic
Research Data Management Practices
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

"Can we replicate your findings?" Assessing the availability of reproducibility assets in machine learning-based concrete research

Diāna Bajāre, Jeonghyun Kim
1 links
Research Integrity and Peer Review
Research Data Management Practices
article

"Can we replicate your findings?" Assessing the availability of reproducibility assets in machine learning-based concrete research

Diāna Bajāre, Jeonghyun Kim
article en

Abstract

Abstract Research on machine learning (ML)-based concrete property prediction has expanded significantly, yet the availability and accessibility of research assets needed to support independent verification remain insufficiently characterized. This study investigated 419 articles indexed in the Science Citation Index Expanded (SCIE) of the Web of Science Core Collection that were retrieved with the WoS Publication Year filter set to 2025 and focused on tabular data-based ML studies predicting concrete properties from mixture-related variables. The availability and accessibility of datasets, source code, and graphical user interfaces (GUIs) were manually verified and further analyzed by journal quartile and corresponding author country. The empirical findings reveal a profoundly low asset-sharing rate; algorithmic source codes remained undisclosed in 90.7% of the literature. For datasets, full raw data sharing stood at only 21.5%, whereas 50.4% relied on external citations. Furthermore, an analysis of 85 GUI-developed publications showed that 66.0% and 53.6% of instances in Q1 and Q2 journals, respectively, exhibited a developed-but-restricted access pattern. An exploratory review of Data Availability Statements identified several stated reasons for restricting asset sharing, including institutional intellectual property protection and industrial confidentiality agreements, and noted several statements that appeared inconsistent with the data-dependent nature of the reported ML analyses.

Research Integrity and Peer Review
Riga Technical University (LV)
Latvijas Zinātnes Padome
Research Data Management Practices
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon