tags:
- time series
- time series classification
- monster
- audio
license: other
pretty_name: MosquitoSound
size_categories:
- 100K<n<1M
Part of MONSTER: https://arxiv.org/abs/2502.15122.
| MosquitoSound | |
|---|---|
| Category | Audio |
| Num. Examples | 279,566 |
| Num. Channels | 1 |
| Length | 3,750 |
| Sampling Freq. | 6 kHz |
| Num. Classes | 1 |
| License | Public Domain |
| Citations | [1] [2] |
MosquitoSound, taken from the broader UCR archive, consists of 279,566 (univariate) time series, each of length 3,750, representing recordings of wingbeats for six different species of mosquito [1, 2]. The recordings are single channel with a sampling rate of 6 kHz (i.e., the time series represent just over half a second of data). As for the FruitFlies dataset, and using similar hardware, the recordings were made using an infrared sensor detecting the vibration of the wings of the mosquitoes. The task is to identify the species of mosquito based on the recordings. This version of the dataset has been split into stratified random cross-validation folds.
[1] Eleftherios Fanioudakis, Matthias Geismar, and Ilyas Potamitis. (2018). Mosquito wingbeat analysis and classification using deep learning. In 26th European Signal Processing Conference, pages 2410–2414.
[2] Ilyas Potamitis. (2018). Wingbeats. https://www.kaggle.com/datasets/potamitis/wingbeats; https://timeseriesclassification.com/description.php?Dataset=MosquitoSound. Public Domain.