Assessing the Performance of ARBIMON for Nocturnal Bird Monitoring: A
Sensitivity and Specificity Approach.
1: Passive Acoustic Monitoring (PAM) revolutionises ecological research,
utilizing sounds for species-specific inferences. However, PAM generates
large volumes of data, posing challenges in annotation, classification,
and review complexity, necessitating efficient data management
strategies. 2: Given this particular need, this research aimed to
improve the performance of a pattern-matching algorithm for detecting
signals of interest in two nocturnal bird species. The study pursued two
main objectives: first, to evaluate various similarity scores and
determine the optimum one through a sensitivity-specificity analysis.
Second, we investigate potential relationships between species-specific
spectral features, such as high, low, and peak frequencies, and the
algorithm’s performance by reviewing and comparing their dispersion with
a Levene test. 3: The outcomes demonstrated a generally favorable
algorithm performance, achieving up to 80% sensitivity and specificity.
This underscores its effectiveness in identifying target signals. Our
investigation indicated that factors like individuality, which could be
reflect on the spectral features, could potentially impact the
algorithm’s efficacy. 4: ARBIMON provides transformative collaborative
solutions in the field of bioacoustics. However, additional research is
imperative to fully grasp the performance and potential applications of
such tools. This exploration extends beyond ARBIMON to encompass the
burgeoning technologies within the discipline.