Using non-continuous accelerometry to identify cryptic nesting events of Galapagos giant tortoises

Table 2 Classifier output comparison for the tortoise nest detection algorithm

Time period	Algorithm	F-1 Score			Overall average
Time period	Algorithm	Temporal	Individual	Random	Overall average
15:00–21:00	BRT	0.78 (0.77–0.78)	0.83 (0.78–0.87)	0.70 (0.55–0.78)	0.770
15:00–21:00	RF	0.84 (0.79–0.88)	0.85 (0.81–0.89)	0.75 (0.65–0.79)	0.813
16:00–21:00	BRT	0.85 (0.81–0.89)	0.84 (0.81–0.86)	0.73 (0.67–0.82)	0.810
16:00–21:00	RF	0.87 (0.80–0.94)	0.90 (0.84–0.94)	0.78 (0.74–0.83)	0.837
16:00–22:00	BRT	0.86 (0.84–0.87)	0.88 (0.84–0.93)	0.78 (0.75–0.82)	0.827
16:00–22:00	RF	0.86 (0.81–0.91)	0.90 (0.88–0.93)	0.82 (0.78–0.86)	0.860
16:00–23:00	BRT	0.88 (0.87–0.88)	0.88 (0.81–0.94)	0.77 (0.72–0.83)	0.843
16:00–23:00	RF	0.89 (0.86–0.91)	0.91 (0.85–0.95)	0.81 (0.77–0.86)	0.867
16:00–00:00	BRT	0.89 (0.88–0.89)	0.90 (0.85–0.93)	0.81 (0.78–0.86)	0.860
16:00–00:00	RF	0.89 (0.88–0.91)	0.90 (0.86–0.93)	0.85 (0.79–0.90)	0.883
16:00–01:00	BRT	0.89 (0.88–0.91)	0.85 (0.75–0.89)	0.82 (0.79–0.86)	0.870
16:00–01:00	RF	0.90 (0.88–0.93)	0.86 (0.83–0.90)	0.84 (0.81–0.86)	0.880

Boosted Regression Tree (BRT) and Random Forest (RF) classifier output comparison for the tortoise nest detection algorithm from triaxial accelerometer data. Training data include 112 confirmed nesting events by 21 female Galapagos giant tortoises. We report the F-1 scores (harmonic mean of Precision and Sensitivity) for data summarized over six biologically relevant time periods and three validation methods: withholding a year of data (“Temporal”), withholding data from random individuals (“Individual”), and withholding a random subset of data (“Random”). The value reported is the average across iterations with the range in brackets. The highest performing model for each validation method is shown in bold

ISSN: 2050-3385