Abstract
Heliciculture, the cultivation of edible land snails, is gaining prominence as a sustainable and economically viable agricultural sector. Despite this growth, productivity in snail farming remains inconsistent due to variable environmental conditions and empirical management practices. This study applies machine learning (ML) techniques for the first time to predict production outcomes in Cornu aspersum maximum farms across Greece. Data were collected from 30 operational farms using structured questionnaires, covering biological, environmental, and management-related variables. A suite of supervised ML algorithms, Support Vector Machine (SVM), Neural Network, Stochastic Gradient Descent, and Linear Regression, were trained and evaluated using stratified 5-fold cross-validation. SVM was the most accurate model, achieving 85% predictive accuracy. SHAP (SHapley Additive exPlanations) analysis identified snail
density, feed quantity, and mortality rate as the three most critical variables influencing production output. Additionally, hierarchical clustering distinguished two farm clusters with distinct performance patterns, suggesting divergent production strategies. These results demonstrate the potential of ML to inform precision management in heliciculture. By identifying key predictors of productivity, this study provides a decision-support framework for optimizing farm inputs and enhancing operational sustainability. Results highlight the value of digital tools in transforming traditional snail farming into a data-driven and efficient system.
Keywords:
Cornu aspersum maximum , Heliciculture , Machine Learning, Predictive Modeling, Production Optimization, Snail FarmingReferences
Issue
Copyright & License
This work is licensed under a Creative Commons Attribution 4.0 International License.