Machine Learning Methods Applied to the Prediction of Pseudo-nitzschia spp. Blooms in the Galician Rias Baixas (NW Spain)

Francisco M Bellas Aláez,Jesus M Torres Palenzuela,Luis González Vilas,Evangelos Spyrakos

doi:10.3390/ijgi10040199

Francisco M Bellas Aláez, Jesus M Torres Palenzuela + Show 2 more

Open Access

https://doi.org/10.3390/ijgi10040199

Copy DOI

Abstract

This work presents new prediction models based on recent developments in machine learning methods, such as Random Forest (RF) and AdaBoost, and compares them with more classical approaches, i.e., support vector machines (SVMs) and neural networks (NNs). The models predict Pseudo-nitzschia spp. blooms in the Galician Rias Baixas. This work builds on a previous study by the authors (doi.org/10.1016/j.pocean.2014.03.003) but uses an extended database (from 2002 to 2012) and new algorithms. Our results show that RF and AdaBoost provide better prediction results compared to SVMs and NNs, as they show improved performance metrics and a better balance between sensitivity and specificity. Classical machine learning approaches show higher sensitivities, but at a cost of lower specificity and higher percentages of false alarms (lower precision). These results seem to indicate a greater adaptation of new algorithms (RF and AdaBoost) to unbalanced datasets. Our models could be operationally implemented to establish a short-term prediction system.

Highlights

Harmful algae blooms (HABs) are an increasingly frequent and intense event in coastal areas worldwide [1,2]
This study focuses on the HABs caused by Pseudo-nitzschia spp. in the Rias Baixas area in Galicia (NW Spain)
In addition to the most extended machine learning technologies for HABs prediction (SVM and neural networks (NNs)), we introduce recent ensemble learning methods (RF and AdaBoost), which are expected to show a better generalization ability by combining multiple weak learners [16]

Summary

Introduction

Harmful algae blooms (HABs) are an increasingly frequent and intense event in coastal areas worldwide [1,2]. The prediction of phytoplankton blooms includes the application of several methods which can vary in modelling approach and complexity [5]. The prediction of harmful algal events is based on statistical numerical models, such as logistic regression, regression trees or Bayesian models [6,7,8,9,10]. Most of these approaches are limited to linear systems, while HABs usually occur in complex and highly dynamic coastal environments [11,12]

Methods

Results

Discussion

Conclusion