Prediction of Algae Growth: A Machine Learning Perspective

Algal blooms pose a significant threat to aquatic ecosystems and human health. To address this issue, this paper proposes a machine learning-based approach for predicting harmful algal blooms (HABs) by analyzing environmental features. Algae, as primary organic matter and oxygen producers, play a vital role in the biosphere. However, the exponential increase in algal growth worldwide poses significant challenges to economic development and long-term sustainability. The paper employs three popular machine learning algorithms: Artificial Neural Network (ANN), Gradient Boosting Decision Tree (GBDT), and Support Vector Machine (SVM) to predict algal blooms. The research utilizes real-time data from two locations: the Sassafras River in the United States Chesapeake Bay and Lake Okeechobee in Florida, USA. These locations have experienced frequent HABs due to factors like chemical runoff and nutrient-rich conditions. By analyzing the collected data, the paper identifies and selects the most important features to optimize the prediction models' accuracy. Preliminary results demonstrate promising accuracy in predicting algal growth and identifying key characteristics associated with HABs. These findings contribute to a better understanding of algal blooms and pave the way for effective mitigation strategies to combat this global environmental challenge.