Modeling Inquiry about nichenetr

[xuansuyang]

Thank you for your response. I'll take a look at the paper later. Wishing you a happy life.

[csangara]

Hi,

The column represents the way the weights are aggregated between different rounds of parameter optimization. We used a five-fold cross-validation procedure to optimize the weights for the data sources and hyperparameters, as described in the MultiNicheNet paper (Parameter optimization via NSGA-II in Methods and Supplementary Note 2). These avg_weight and median_weight columns are respectively the average and median of the weights across all folds. Specifically, it was computed using the 25 parameter settings in each fold with the highest geometric average over the four objective functions. The final model uses the median weight, but as you can see in Supplementary Notes Figure 2.3 and 2.4, there is not much difference in performance between using the average and median.