MaxEnt software has three Replicated Run Type with names: Bootstrap; Subsample, and crossvalidate. In my analyses, I obtain better AUC value when I use Bootstrap replicated run type instead of Subsample, and crossvalidate; So, can I do my analyses with this run type or I have to use crossvalidate?
Please can anyone help me?
Why not test all three? My preference has become the bootstrap for its depiction of Standard Deviation (our bear Majella and ASF China articles)
I'm not a fan of these run types because it's not really 'replication'. But AUC is still a valuable parameter
I tried all the repetitions for my thesis and I also reviewed the resources. The bootstrap was almost better than the rest. You can also determine the best according to the purpose of the your research and the definition of each type of repetition.
- Badges
- Science topic
- Similar topics
- Statistics
- Time Series