I have generated a Maxent model for species distribution using only presence points. I need to evaluate the model and I don't know How to calculate the AUC values manually. I heard that the AUC values generated by the Maxent are not trusted so would like to calculate it by the manual method.
I presume you mean area under the ROC curve (i.e. sensitivity versus 1–specificity plot), a measure of the accuracy of a quantitative diagnostic test by rank-correlation. You can calculate AUC in R or manually using the area of a trapezoid, but first normalise the data. If you achieve AUC value near to one then your developed method is statistically sound.
Thanks Andrew for your answer. Can you please suggest me any package in R which I can use for the calculation of AUC value? I am new to Maxent modeling and have been told to calculate the AUC by the manual method. I heard that the AUC value generated by the Maxent is not reliable. My current AUC value is 0.98(calculated by Maxent).
Are you sure Maxent is not reliable? I don't see any reason why they are not reliable. I would trust that value! Trust your statistical tests and software lol :)
Please check Biomod r package: https://stat.ethz.ch/pipermail/r-sig-ecology/2016-July/005405.html
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package