Identifying livestock behavior patterns based on accelerometer dataset

•A new methodology to automatic detecting animal behavior is presented.•It includes a novel algorithm, named LADA, that is able to detect activity and inactivity time periods in cattle from huge datasets.•The algorithm can be configured depending on weather greater accuracy or greater sensitivity is desired in final results.•Data is extracted from HOBO data loggers devices placed on Iberian pigs and calves and is properly preprocessed.•Performance is evaluated using behavior annotations carried out by experts as control dataset.