Algal Bloom Forecasting in a Classification and Regression Setting

Algal Bloom Forecasting in a Classification and Regression Setting: Implementing a UNet Architecture to evaluate the differences between both settings

Alvarez Lucendo, Rodrigo (TU Delft Electrical Engineering, Mathematics and Computer Science; TU Delft Pattern Recognition and Bioinformatics)

Lengyel, A. (mentor)
van Gemert, J.C. (mentor)
Bruintjes, R. (mentor)
Langendoen, K.G. (graduation committee)

Delft University of Technology

Computer Science and Engineering

CSE3000 Research Project

2023-02-03

Forecasting algal blooms using remote sensing data is less labour-intensive and has better cover- age in time and space than direct water sampling. The paper implements a deep learning technique, the UNet Architecture, to predict the chlorophyll concentration, which is a good indicator for al- gal bloom in the Rio Negro water reservoirs of Uruguay. The research question focuses on the dif- ferences between classification and regression in algal bloom forecasting. The experiments show that the regression implementation achieves bet- ter accuracy and lower mean squared error than the classification implementation that uses cross- entropy loss and four pre-fixed bins. Different loss functions that account for the class imbalance in the data do not improve the model’s performance. Fi- nally, a quantile-based binning strategy that consid- ers the data’s underlying distribution achieves the highest accuracy in both settings.

http://resolver.tudelft.nl/uuid:05505158-1c11-4c3c-820e-2ade68ba753a

Student theses

bachelor thesis

© 2023 Rodrigo Alvarez Lucendo