Classification of Microcystins Based on Temperature Data in North Dakota Lake

Microcystins (MC), the toxins produced by certain species of algae (cyanobacteria) in lakes, rivers, and reservoirs pose a serious threat to society and the environment. Accurate prediction of microcystin presence in lakes can result in considerable economic savings. Prediction of microcystins is often challenging due to limitations in datasets. This paper analyzed the highly correlated variables for higher microcystins and finally, the threshold levels of water temperature for higher microcystins in Bowman-Haley Dam, North Dakota was determined. Multiple Machine Learning Classification models such as Random Forest (RF), Logistic Regression (LR), Naive Bayes (NB), Support Vector Machine (SVM), and KNN were applied to find the optimal temperature threshold for higher microcystin class. Moreover, class imbalance issues in microcystin data were addressed using multiple algorithms such as SMOTE, ADASYN, and SMOTE_ENN. The density plot distribution graph also revealed a strong association between water temperature and microcystin levels. The optimal threshold of water temperature for higher microcystins in North Dakota lakes is found to be 20.21 degrees C.