Machine learning models for water safety enhancement

Area of Science:

• Environmental Science
• Radiation Biology
• Public Health

Background:

• Human exposure to natural and artificial radiation sources is a public health concern.
• Radionuclides and heavy metals in consumables can pose health risks, including cancer and genetic mutations.
• Assessing radiation and heavy metal levels in consumable mineral water is crucial for public safety.

Purpose of the Study:

• To quantify radioactive isotopes and heavy metals in mineral water from Arak City, Iran.
• To evaluate the associated health risks from consuming this mineral water.
• To validate findings using Machine Learning (ML) models.

Main Methods:

• Random sampling of mineral water from Arak City.
• Radioactive isotope analysis (Th-232, K-40, Cs-137, Ra-226) and heavy metal screening (lead, chromium).
• Calculation of annual effective doses and comparison with WHO and UNSCEAR guidelines.
• Application of Machine Learning (ML) models for prediction accuracy.

Main Results:

• Concentrations of Th-232, K-40, and Cs-137 were below WHO-established thresholds.
• Ra-226 was not detected in the samples.
• Lead and chromium were absent; annual effective doses were below UNSCEAR limits.
• ML models demonstrated high accuracy in predictions.

Conclusions:

• Mineral water consumed in Arak City presents minimal health risks regarding radioactive isotopes and heavy metals.
• Radiation exposure from mineral water consumption is significantly below international safety standards.
• The study confirms the safety of local mineral water, supported by robust ML model predictions.