Multidimensional terrestrial ecoacoustics assessment

Ecologists faces challenges when analysing environmental acoustic data, such as handling big data, summarising and comparing long-duration recordings across regions and timescales. This thesis provide an innovative framework that combines time-series motif discovery and machine learning for analysing environmental sounds. The proposed methodology was successfully applied to different temporal and spatial scales, demonstrating its applicability across different contexts. The new framework provides a faster and better tool to assess environmental sounds, improving the understanding of biodiversity, uncovering patterns and better distributing resources directed at biodiversity conservation