Soundscape classification with convolutional neural networks reveals temporal and geographic patterns in ecoacoustic data

Agriculture & Biology

• acoustics 81%
• agriculture 12%
• biodiversity 40%
• citizen scientists 28%
• coniferous forests 17%
• diurnal variation 18%
• ecosystems 12%
• environmental health 20%
• environmental knowledge 22%
• learning 15%
• methodology 4%
• microphones 23%
• monitoring 10%
• neural networks 90%
• planning 15%
• Quercus 15%
• roads 16%
• shrublands 17%
• solar radiation 14%
• space and time 18%
• taxonomy 61%
• testing 6%
• vocalization 35%
• weather 14%
• wildlife 16%
• woodlands 16%

Earth & Environmental Sciences

• acoustic data 42%
• acoustics 40%
• agriculture 11%
• animal community 22%
• biodiversity 36%
• citizen 16%
• coniferous tree 17%
• customs 36%
• daylight 19%
• disturbance 11%
• ecosystem 9%
• ecosystem health 19%
• effect 3%
• environmental planning 20%
• human activity 12%
• indicator 9%
• learning 13%
• methodology 9%
• monitoring 8%
• oak 16%
• restoration 12%
• road 11%
• shrubland 19%
• sound 41%
• test 6%
• vocalization 44%
• weather 11%
• wildlife 14%
• woodland 15%

Business & Economics

• Agriculture 10%
• Animals 8%
• Biodiversity 48%
• Conservation 11%
• Deep Learning 19%
• Ecosystem 29%
• Environmental Planning 20%
• Gradient 15%
• Health 9%
• Interference 13%
• Methodology 7%
• Monitoring 10%
• Neural Networks 100%
• Restoration 11%
• Roads 11%
• Time to Build 20%
• United States of America 5%
• Weather 14%
• Wildlife 18%
• Woodland 17%