Prediction of forest aboveground biomass: an exercise on avoiding overfitting

  • Authors:

  • Sara Silva;Vijay Ingalalli;Susana Vinga;João M. B. Carreiras;Joana B. Melo;Mauro Castelli;Leonardo Vanneschi;Ivo Gonçalves;José Caldas

  • Affiliations:

  • INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal, CISUC, Universidade de Coimbra, Coimbra, Portugal;INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal;INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal, FCM, Universidade Nova de Lisboa, Lisboa, Portugal;Instituto de Investigação Científica Tropical, Lisboa, Portugal;Instituto de Investigação Científica Tropical, Lisboa, Portugal;INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal, ISEGI, Universidade Nova de Lisboa, Lisboa, Portugal;ISEGI, Universidade Nova de Lisboa, Lisboa, Portugal, INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal;CISUC, Universidade de Coimbra, Coimbra, Portugal;INESC-ID, IST, Universidade Técnica de Lisboa, Lisboa, Portugal

  • Venue:
  • EvoApplications'13 Proceedings of the 16th European conference on Applications of Evolutionary Computation
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Mapping and understanding the spatial distribution of forest aboveground biomass (AGB) is an important and challenging task. This paper describes an exercise of predicting the forest AGB of Guinea-Bissau, West Africa, using synthetic aperture radar data and measurements of tree size collected in field campaigns. Several methods were attempted, from linear regression to different variants and techniques of Genetic Programming (GP), including the cutting edge geometric semantic GP approach. The results were compared between each other in terms of root mean square error and correlation between predicted and expected values of AGB. None of the methods was able to produce a model that generalizes well to unseen data or significantly outperforms the model obtained by the state-of-the-art methodology, and the latter was also not better than a simple linear model. We conclude that the AGB prediction is a difficult problem, aggravated by the small size of the available data set.