The viewer is disabled because this dataset repo requires arbitrary Python code execution. Please consider
removing the
loading script
and relying on
automated data support
(you can use
convert_to_parquet
from the datasets library). If this is not possible, please
open a discussion
for direct help.
Part of MONSTER: https://arxiv.org/abs/2502.15122.
CornellWhaleChalleng consists of hydrophone recordings [1]. The processed dataset consists of 30,000 (univariate) time series, each of length 4,000. The task is to distinguish right whale calls from other noises. (An abridged version of this dataset is included in the broader UCR archive.) This version of the dataset has been divided into stratified random cross-validation folds.
[1] André Karpištšenko, Eric Spalding, and Will Cukierski. (2013). The Marinexplore and Cornell University whale detection challenge. https://kaggle.com/competitions/whale-detection-challenge. Copyright 2011 Cornell University and the Cornell Research Foundation.
- Downloads last month
- 22