Spatial prediction of demersal fish diversity in the Baltic Sea: comparison of machine learning and regression-based techniques

Abstract

Marine spatial planning (MSP) is considered a valuable tool in the ecosystem-based management of marine areas. Predictive modelling may be applied in the MSP framework to obtain spatially explicit information about biodiversity patterns. The growing number of statistical approaches used for this purpose implies the urgent need for comparisons between different predictive techniques. In this study, we evaluated the performance of selected machine learning and regression-based methods that were applied for modelling fish community indices. We hypothesized that habitat features can influence fish assemblage and investigated the effect of environmental gradients on demersal fish diversity (species richness and Shannon-Weaver Index). We used fish data from the Baltic International Trawl Surveys (2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013)(2014) and maps of six potential predictors: bottom salinity, depth, seabed slope, growth season bottom temperature, seabed sediments and annual mean bottom current velocity. We compared the performance of six alternative modelling approaches: generalized linear models, generalized additive models, multivariate adaptive regression splines, support vector machines, boosted regression trees and random forests. We applied repeated 10-fold cross-validation, using accuracy as the measure of model quality. Finally, we selected random forest as the best performing algorithm and implemented it for the spatial prediction of fish diversity from the Baltic Proper to the Kattegat. To obtain information on the data reliability and confidence of the developed models, which are essential for MSP, we estimated the uncertainty of predictions with standard deviation of predictions obtained from all the trees in the ensemble random forest method. We showed how state-of-the-art predictive techniques, based on easily available data and simple Geographic Information System tools, can be used to obtain reliable spatial information about fish diversity. Our comparative work highlighted the potential of machine learning method to reduce prediction error in modelling of demersal fish diversity in the framework of MSP.

Hyperparameter tuning and performance assessment of statistical and machine-learning algorithms using spatial data

“…This is the reason why most model comparison studies with a binary response (e.g. Goetz et al (2015); Smoliński & Radtke (2016)) only use AUROC as a single error measure.…”

“…These have gained popularity due to their ability to handle high-dimensional and highly correlated data and the lack of explicit model assumptions. Some model comparison studies in the spatial modeling field suggest that machine learning models might be the better choice when the primary aim is accurate prediction (Hong et al, 2015;Smoliński & Radtke, 2016;Youssef et al, 2015). However, other studies found no major performance difference to parametric models Goetz et al, 2015).…”

Individual Tree Diameter Growth Models of Larch–Spruce–Fir Mixed Forests Based on Machine Learning Algorithms

“…Then the comparisons between the four optimal models were made by assessing the performance measures on each fold. As the performances were measured using identically resampled datasets, statistical methods for paired comparisons can be used to detect pairwise differences in the performance among models [27,48]. Therefore, paired t-test was applied to verify if the differences in performance among the four optimal models with the significant level of 0.05.…”

Reconstruction of Ocean Color Data Using Machine Learning Techniques in Polar Regions: Focusing on Off Cape Hallett, Ross Sea

“…The most critical work in this study was to select the appropriate model based on the given datasets and to determine the optimized model parameters (called the hyperparameters) in the selected model to derive the most accurate output. We used the 10-fold cross-validation method to assess the accuracy of the different models [30]. Once models with good performance were selected, the grid-search approach was used to identify hyperparameters in the model.…”

“…Machine learning is one of the reliable and cost-effective approaches that could reconstruct the inevitable gaps in the ocean color data and is a predictive model. In particular, the approach is more flexible than conventional parametric models because of its ability to handle non-linear relationships and complex interactions, which often occur in ecological data [28][29][30]. Jouini et al [9] attempted to recover significant gaps in satellite-derived chlorophyll concentration (CHL) data using Self Organizing Maps (SOM) classification (instance-based machine learning) in the western sector of the North Atlantic.…”