| This research presents a Soft Computing analysis of the dissolved oxygen, in the water resources of Mesochora area, in the Trikala prefecture of central Greece. Machine learning and computational intelligence techniques are utilized. This research is based on the limits set by the Norwegian Institute for Water Re-search (NIVA). Data was classified in five quality classes ("Poor", "Insuffi-cient", "Moderate", "Good" and "High") by employing popular machine learning algorithms namely k-Nearest Neighbors, Logistic Regression, Decision Tree, Random Forest, Ensemble Classifier and Support Vector Machines. Moreover, SHAP values were applied to offer model interpretation and to reveal the relative contribution of each input feature. The results have showen high classification ac-curacy and they provided useful clues for the successful optimization of water re-sources monitoring and management. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.