Skip to main content
. 2012 May 7;7(5):e35850. doi: 10.1371/journal.pone.0035850

Figure 4. Multi-objective model optimization: extreme solutions and trade-offs.

Figure 4

A) Binarized response for the neuron in Fig.1A. B) Predicted response for a model belonging to the optimal Pareto front. The predictive performances of this model represent a suitable trade-off between Inline graphic and Inline graphic. C) Predicted response for a model belonging to the optimal Pareto front. This model has the best predictive performances for Inline graphic at the expense of a significant worsening in Inline graphic. D) Average response for the true neuron (blue) and the model in (B) and (C) (respectively red and green lines). E-H) Same as (A-D) for a different cell. For the model in (F) Inline graphic and Inline graphic are respectively 0.70 and 0.41. For the model in (G) Inline graphic = 0.64 and Inline graphic = 0.61. The smallest Inline graphic achievable was 0.36 (but for Inline graphic = 0.74).