Inter-Research > MEPS > v140 > p251-256  
Marine Ecology Progress Series

via Mailchimp

MEPS 140:251-256 (1996)  -  doi:10.3354/meps140251

Artificial neural network versus multiple linear regression: predicting P/B ratios from empirical data

Brey T, Jarre-Teichmann A, Borlich O

Traditionally, multiple linear regression models (MLR) are used to predict the somatic production/biomass (P/B) ratio of animal populations from empirical data of population parameters and environmental variables. Based on data from 899 benthic invertebrate populations, we compared the prediction of P/B by MLR models and by Artificial Neural Networks (ANN). The latter showed a slightly (about 6%) but significantly better performance. The accuracy of both approaches was low at the population level, but both MLR and ANN may be used to estimate production and productivity of larger population assemblages such as communities.

Productivity · Benthic invertebrates · Artificial Neural Network

Full text in pdf format