MammalNet视频数据集围绕涵盖 17 个目、69 个科和 173 个哺乳动物类别的生物哺乳动物分类学构建，并包括 12 种常见的高级哺乳动物行为（例如狩猎、梳理行为）。 MammalNet 能够单独或联合研究动物和行为识别。它还有助于研究具有挑战性的构图场景，测试模型的零样本和低样本迁移能力。发表于CVPR2023，https://mammal-net.github.io/

About

Real-world animal interaction network data sets. Animal interaction data from published studies of wild, captive, and domesticated animals.

A machine learning algorithm was applied to identify clusters of detections on feeders. Next, the network was generated based on patterns of co-occurrence by individuals in the same feeding events. Asociations between birds were defined using the simple ratio index.

Category Animal Social Networks Collection Animal Networks Source https://bansallab.github.io/asnr/data.html Short Animal Networks Vertex type Animal, Bird, songbird Edge type Interaction Format Undirected Edge weights Weighted Species Haemorhous mexicanus Taxon. class Aves Population free-ranging Geo. location Virginia Tech, Virginia, USA Data collection RFID Interaction type social projection bipartite Edge weight type simple_ratio_index Data collection duration 137 days Time resolution (within a day) 1 sec Time span (within a day) 24 hours Citation Adelman, James S., et al. "Feeder use predicts both acquisition and transmission of a contagious pathogen in a North American songbird." Proc. R. Soc. B. Vol. 282. No. 1815. The Royal Society, 2015."

This data package is formatted as an ecocomDP (Ecological Community Data Pattern). For more information on ecocomDP see https://github.com/EDIorg/ecocomDP. This Level 1 data package was derived from the Level 0 data package found here: https://pasta.lternet.edu/package/metadata/eml/knb-lter-sgs/137/17. The abstract below was extracted from the Level 0 data package and is included for context:

 This data package was produced by researchers
                       working on the Shortgrass Steppe Long Term
                       Ecological Research (SGS-LTER) Project,
                       administered at Colorado State University.
                       Long-term datasets and background information
                       (proposals, reports, photographs, etc.) on the
                       SGS-LTER project are contained in a comprehensive
                       project collection within the Digital Collections
                       of Colorado
                       (http://digitool.library.colostate.edu/R/?func=collections&collection_id=3429).
                       The data table and associated metadata document,
                       which is generated in Ecological Metadata
                       Language, may be available through other
                       repositories serving the ecological research
                       community and represent components of the larger
                       SGS-LTER project collection. Additional information and referenced materials can be found: 
                       http://hdl.handle.net/10217/83452.
 Small mammals (rabbits, rodents) are integral
                       components of semiarid ecosystems because of their
                       roles as consumers of plants, seeds and
                       arthropods, as soil disturbance agents, and as
                       food for raptors, snakes and mammalian carnivores.
                       Because of their vagility and intermediate trophic
                       position, populations of small mammals may track
                       changes in vegetation and the abiotic environment
                       that may result from shifts in land-use and other
                       anthropogenic disturbances. However, these
                       populations are variable over space and time, and
                       their response to environmental changes may not be
                       immediately apparent given their behavioral
                       flexibility and relatively long life-spans and
                       generation times. Patterns in the distribution and
                       abundance of small mammals thus may simultaneously
                       reflect and affect the stability of the
                       shortgrass-steppe ecosystem. Long-term studies of
                       population and community dynamics therefore are
                       needed to fully understand the role of small
                       mammals in grassland ecosystems. In 1994, we
                       implemented a sampling scheme to monitor long-term
                       changes in relative abundance of small mammals in
                       representative habitats of shortgrass steppe. We
                       live-trapped nocturnal rodents twice each year
                       (spring, late summer) on trapping webs in upland
                       prairie (GRASS) and saltbush-dominated (SHRUB)
                       habitats. Three 3.14-ha webs were established in
                       each habitat. Each web had 124 Sherman traps,
                       which were spaced 10-m apart on 12 100-m spokes,
                       with 30 degrees betwe

This data package is formatted as an ecocomDP (Ecological Community Data Pattern). For more information on ecocomDP see https://github.com/EDIorg/ecocomDP. This Level 1 data package was derived from the Level 0 data package found here: https://pasta.lternet.edu/package/metadata/eml/knb-lter-sgs/136/17. The abstract below was extracted from the Level 0 data package and is included for context: This data package was produced by researchers working on the Shortgrass Steppe Long Term Ecological Research (SGS-LTER) Project, administered at Colorado State University. Long-term datasets and background information (proposals, reports, photographs, etc.) on the SGS-LTER project are contained in a comprehensive project collection within the Digital Collections of Colorado (http://digitool.library.colostate.edu/R/?func=collections&collection_id=3429). The data table and associated metadata document, which is generated in Ecological Metadata Language, may be available through other repositories serving the ecological research community and represent components of the larger SGS-LTER project collection. Additional information and referenced materials can be found: http://hdl.handle.net/10217/83448. Rabbits are the most important small-mammal herbivores in shortgrass steppe, and may significant influence the physiognomy and population dynamics of herbaceous plants and woody shrubs. Rabbits also are the most important prey of mammalian carnivores such as coyotes and large raptors such as golden eagles and great horned owls. Two hares (Lepus californicus, L. townsendii) and one cottontail rabbit (Sylvilagus audubonii) occur in shortgrass steppe. In 1994, we initiated long-term studies to track changes in relative abundance of rabbits on the Central Plains Experimental Range (CPER). On four nights each year (one night each season, usually on new moon nights in January, April, July, October), we drove a 32-km route consisting of pasture two-track and gravel roads on the CPER. This was the same route as that driven for carnivore scat counts. Surveys began at twilight. Observers with two spotlights sat in the back of a 4WD pick-up driven at <15 mph. We recorded the perpendicular distance (in m) to each rabbit using a meter tape or range finder; the direction of the rabbit relative to the road; and noted if individuals were solitary or in clusters. We also recorded any carnivores or other animals seen during spotlight surveys. Beginning in Spring 1998, we recorded the vegetation type (habitat) and topographic position of each rabbit seen, as well ... Visit https://dataone.org/datasets/https%3A%2F%2Fpasta.lternet.edu%2Fpackage%2Fmetadata%2Feml%2Fedi%2F327%2F2 for complete metadata about this dataset.