Commentary to: a cross-validation-based approach for delimiting reliable home range estimates.

Area of Science:

• Ecology
• Wildlife Biology
• Spatial Ecology

Background:

• The Time Local Convex Hull (T-LoCoH) method is used for home range delimitation.
• The original cross-validation approach for T-LoCoH has demonstrated performance issues.

Purpose of the Study:

• To address limitations in the T-LoCoH cross-validation method.
• To introduce a statistically rigorous approach for optimizing home range parameters.

Main Methods:

• Replaced the original cross-validation score with a new formulation based on total log probability of out-of-sample predictions.
• Interpreted normalized LoCoH hulls as probability densities to derive these probabilities.

Main Results:

• Optimal parameter sets derived using the new formulation differ significantly from the original.
• Home range size, revisitation rates, and visit durations are altered by the corrected formulation.

Conclusions:

• The revised cross-validation approach offers a statistically grounded framework for home range analysis.
• This method moves beyond subjective user selections, relying instead on inherent statistical properties of animal movement data.