Sign in Subscribe
Machine learning

Sustainable Land Management: New Research on Soil Organic Carbon Prediction using Machine Learning

Strategic Insights

— 1 min read

Researchers from Imam Mohammad Ibn Saud Islamic University have made significant contributions to sustainable land management by developing a novel integration of the Ninja Optimization Algorithm (NiOA) for simultaneous feature selection and hyperparameter optimization. This innovative approach enables accurate prediction of soil organic carbon (SOC) levels, which play a crucial role in global carbon cycling, climate regulation, and soil fertility.

Key Takeaways:

  • The researchers used 80% of the dataset for training and 20% for testing, achieving a mean squared error (MSE) of 0.00513 with the baseline Support Vector Machine (SVM) model.
  • The binary NiOA (bNiOA) for feature selection reduced the MSE to 0.00011.
  • Full NiOA-based hyperparameter tuning improved the MSE further to 7.52 x 10-7, corresponding to a 99.98% reduction in prediction error.
  • The proposed NiOA-enhanced framework demonstrates potential in advancing SOC modeling, offering a scalable, interpretable, and high-precision solution for data-scarce environments.
  • The research aims to support sustainable land management and climate change adaptation strategies.
  • Anis Ben Ghorbal led the research team consisting of Azedine Grine, Marwa M. Eid, and El-Sayed M. El-kenawy.

Statistics:

  • 80% of the dataset was allocated for training, and 20% for testing.
  • The baseline SVR model achieved an MSE of 0.00513.
  • The bNiOA reduced the MSE to 0.00011.
  • Full NiOA-based hyperparameter tuning improved the MSE to 7.52 x 10-7.
  • The proposed framework demonstrated a 99.98% reduction in prediction error.

Sources:

  • "Sustainable soil organic carbon prediction using machine learning and the ninja optimization algorithm" (Frontiers in Environmental Science, 2025,13)
  • Frontiers in Environmental Science (http://www.frontiersin.org/environmental_science)
  • Frontiers Media S.A. (publisher)
  • DOI: 10.3389/fenvs.2025.1630762 (free version available at https://doi-org.sdpl.idm.oclc.org/)
  • NewsRx LLC (Copyright 2025)

Read more