An ecologically constrained procedure for sensitivity analysis of Artificial Neural Networks and other empirical models

Abstract

Sensitivity analysis applied to Artificial Neural Networks (ANNs) as well as to other types of empirical ecological models allows assessing the importance of environmental predictive variables in affecting species distribution or other target variables. However, approaches that only consider values of the environmental variables that are likely to be observed in real-world conditions, given the underlying ecological relationships with other variables, have not yet been proposed. Here, a constrained sensitivity analysis procedure is presented, which evaluates the importance of the environmental variables considering only their plausible changes, thereby exploring only ecological meaningful scenarios. To demonstrate the procedure, we applied it to an ANN model predicting fish species richness, as identifying relationships between environmental variables and fish species occurrence in river ecosystems is a recurring topic in freshwater ecology. Results showed that several environmental variables played a less relevant role in driving the model output when that sensitivity analysis allowed them to vary only within an ecologically meaningful range of values, i.e. avoiding values that the model would never handle in its practical applications. By comparing percent changes in MSE between constrained and unconstrained sensitivity analysis, the relative importance of environmental variables was found to be different, with habitat descriptors and urbanization factors that played a more relevant role according to the constrained procedure. The ecologically constrained procedure can be applied to any sensitivity analysis method for ANNs, but obviously it can also be applied to other types of empirical ecological models.

Can machine learning extract the mechanisms controlling phytoplankton growth from large-scale observations? – A proof-of-concept study

“…ESM2Mc is a fully coupled atmosphere, ocean, sea ice model into which is embedded an ocean biogeochemical cycling module. Known as BLING (Biogeochemistry with Light, Iron, Nutrients, and Gases; Galbraith et al, 2010), this module carries a macronutrient, a micronutrient, and light as predictive variables and uses them to predict biomass using a highly parameterized ecosystem (described in more detail below). The half-saturation coefficients (K N in Eq.…”

“…As a demonstration of their capabilities, the ML methods were also applied directly to monthly-averaged output from the BLING model itself using the same predictors in scenarios 1 and 2 but using the biomass calculated from the actual BLING model. As described in Galbraith et al (2010), BLING is a biogeochemical model where biomass is diagnosed as a nonlinear function of the growth rate smoothed in time. The growth rates, in turn, have the same functional form as in scenarios 1 and 2, namely (Eq.…”

Big Earth data analytics: a survey

“…Regression (Franceschini, Tancioni, Lorenzoni, Mattei, & Scardi, 2019), Clustering (Chakraborty & Paul, 2010;Khatami, Mirghasemi, Khosravi, Lim, & Nahavandi, 2017), Classification (Hong et al, 2016), Rule induction…”

“…Derived from statistical methods, regression, classification, clustering can also be used as machine learning methods, thus the exact division between machine learning and statistical methods is not always clear. For example, Artificial Neural Networks can produce regression on approximating and predicting ecological conditions (Franceschini et al, 2019). Machine learning classifiers including Random Forest, Support Vector Machines, and Bayesian Classifiers can produce the probability of an observation belonging to a specific class of Earth process, such as landslide (Hong et al, 2016).…”

The variation differences of cultivated land ecological security between flatland and mountainous areas based on LUCC

“…Moreover, the back-propagation (BP) neural network, which can approximate any nonlinear function and is widely applied in predictions, optimizations, assessments and classifications in China, has been used less in ecological security evaluations[27]. A cultivated land ecological system is a complex multivariable nonlinear dynamic system, that has posed many limitations on traditional methods for assessing and analyzing, and the BP neural network has strong nonlinear approximation ability and the capabilities to handle unclear, disordered and complex information[28]. Its characteristics are appropriate for overcoming the shortcomings of traditional approaches, thereby reaching a higher accuracy in a short time.…”