Review Papers
Synthesizes ML methods across domains, highlighting best practices in validation, explainability, and reproducible workflows.
This project produces review papers that synthesize machine learning methods across disciplines (e.g., soil science, remote sensing, geohazards, archaeology). The reviews organize advances across the full modelling workflow—data, features, algorithms, validation under spatial/temporal dependence, interpretability, and reproducibility. Emphasis is placed on best practices and common pitfalls (data leakage, weak validation, imbalance, inconsistent reporting). Outputs provide roadmaps that help researchers select methods and design credible, comparable studies.
Main focus
- Translate ML advances across disciplines into practical guidance.
- Promote robust validation, explainability, and reproducible workflows.
- Identify trends, gaps, and recommendations for responsible ML adoption.
Objectives
- Systematically review ML pipelines and methods across target disciplines.
- Compare validation and reporting practices to highlight best-practice standards.
- Deliver methodological roadmaps and recommendations for cross-domain transfer.
Graphical abstract
:::: {.row .justify-content-sm-center} ::: {.col-sm-10 .mt-3 .mt-md-0}
::: ::::
::: caption Graphical abstract summarizing the workflow and key results. :::
Papers (selected)
Paper 1 — Machine Learning in Archaeological Practice
Paper 2 — Soil Moisture Retrieval from Sentinel-1
Paper 3 — Digital Soil Mapping in Iran
Paper 4 — Conventional vs Digital Soil Mapping in Iran
Paper 5 — Explainable AI in Geohazards: A Systematic Review
