PoisotLab/SpeciesDistributionToolkit.jl: v1.8.0

SpeciesDistributionToolkit v1.8.0

Diff since v1.7.2

Merged pull requests:

• Add a function to use the distance in degrees to generate pseudo-absences (#527) (@tpoisot)
• AdaBoost learning rate (#531) (@tpoisot)
• AdaBoost tweaks + rebalanced LOOCV (#532) (@tpoisot)
• Add admin areas from ESRI (#533) (@tpoisot)
• Add support for coordinates of instances in SDMs (#535) (@tpoisot)
• Update CHELSA to the new location (#537) (@tpoisot)
• Read scale and offset from geotiff layers (#541) (@tpoisot)
• Next version of SDeMo (#542) (@tpoisot)
• Add a trained label to SDMs (#543) (@tpoisot)
• Model components can be changed (just like variables) (#544) (@tpoisot)
• Conformal for SDeMo (#546) (@tpoisot)
• Documentation overhaul and SDeMo 1.8 integration (#550) (@tpoisot)
• Pretty dramatic performance improvement for sliding windows (#552) (@tpoisot)
• Allow reference value to be bassed to the summary function in sliding windows (#553) (@tpoisot)

Closed issues:

• Add global land use data projections (#31)
• Calling boundingbox on a set of occurrences should ignore the occurrences with no coordinates (#409)
• Update the SDM type with a trained flag (#513)
• Unfortunately, projections (#516)
• Add a link keyword to predict (#525)
• Re-balanced LOOCV (#529)
• Learning rate for AdaBoost (#530)
• CHELSA2 has been updated (#536)
• Conformal prediction (#538)
• Add a classifier! and transformer! method to change a model (#539)
• Check if layers have scale and offset info when reading from geotiff (#540)
• Add a Makie extension to SDeMo to plot instances (#545)
• Add integration between Conformal and layers (#547)