Figure 10.

Comparison of numerical prediction algorithms for nanomaterial environmental impact modeling.

Notes: The list shows the performance of different algorithms for predicting the overall impact of nanomaterials on the embryonic zebrafish. The overall impact is an aggregation of individual toxic endpoints such mortality, delayed development, and morphological malformations. The prediction accuracy is measured by the correlation between the actual and predicted impact values. The abbreviations in the list are defined by the Weka software package. Refer to: http://weka.sourceforge.net/doc.dev/overview-summary.html. Weka is open source software issued under the GNU General Public License. Refer to: http://www.cs.waikato.ac.nz/ml/weka/.

Abbreviations: IBK, instance-based classifier; M5P, M5 model trees; M5Rules, generating rules from M5 model trees; KStar, k instance-based classifier; REPTree, reduced error pruning tree; LWL, locally-weighted learning; SVMreg, support vector machine for regression; SMOreg, sequential minimal optimization for regression; RBFNetwork, radial basis function network.