Decision tree classification of behaviors in the nesting process of green turtles (Chelonia mydas) from tri-axial acceleration data
- 421 Downloads
Sea turtles are endangered marine-adapted reptiles that are obligate terrestrial nesters. They have a nesting process that involves a series of behavioral phases: (1) crawling, (2) digging a body pit, (3) digging an egg chamber, (4) depositing eggs, (5) covering the egg chamber, and (6) camouflaging the site. Discrimination among these behavioral phases is necessary for estimating which parts of the nesting process are affected by human activities or characteristics of beaches. We calculated multivariate features from tri-axial accelerations obtained by a data logger to extract key features for discriminating between these behavioral phases of green turtles (Chelonia mydas). The developed decision tree classifier discriminated these behaviors with high accuracy, resulting in 5 behaviors after combining digging a body pit and camouflaging the site (86.207 %). The structure of the tree showed key features for discriminating the 5 behaviors on the basis of differences in the movements of turtles. The decision rules in this study will enable us to discriminate behavioral phases quantitatively without observer interference, thereby providing a basis for estimating turtle energy budget and the influence of human activities on nesting behavior.
KeywordsAcceleration Decision tree Ethogram Nesting Sea turtle
We acknowledge the following people for their help in attaching and removing data loggers: the members of the Ishigaki Island Sea Turtle Research Group; K. Okuzawa and the staff of the Research Center for Subtropical Fisheries; Seikai National Fisheries Research Institute; D. Imakita, Faculty of Agriculture, Kinki University; and Y. Kawabata, K. Ichikawa, H. Watanabe, T. Hashiguchi, T. Koizumi, and A. Nakabayashi, Graduate School of Informatics, Kyoto University. We thank the two anonymous reviewers for valuable comments on this manuscript. This study was partly supported by the Global COE Program, Informatics Education and Research for a Knowledge-Circulating Society.
- Bao L, Intille SS (2004) Activity recognition from user-annotated acceleration data. In: Ferscha A, Mattern F (eds) Pervasive. Springer, Berlin, pp 1–17Google Scholar
- Bjorndal KA, Jackson JBC (2003) Roles of sea turtles in marine ecosystems: reconstructing the past. In: Lutz PL, Musick JA, Wyneken J (eds) The biology of sea turtles, vol II. CRC Press, Boca Raton, pp 259–273Google Scholar
- Hailman JP, Elowson AM (1992) Ethogram of the nesting female loggerhead (Caretta caretta). Herpetologica 48:1–30Google Scholar
- Hall MA (1998) Correlation-based feature subset selection for machine learning. Doctoral thesis of the University of Waikato. Hamilton, New ZealandGoogle Scholar
- Kooyman GL (2004) Genesis and evolution of bio-logging devices: 1963–2002. Mem Natl Inst Polar Res, Spec Issue 58:15–22Google Scholar
- Lutcavage ME, Plotkin P, Witherington BE, Lutz PL (1997) Human impacts on sea turtle survival. In: Lutz PL, Musick JA (eds) The biology of sea turtles. CRC Press, Boca Raton, pp 387–409Google Scholar
- Miller JD (1997) Reproduction in sea turtles. In: Lutz PL, Musick JA (eds) The biology of sea turtles. CRC Press, Boca Raton, pp 51–81Google Scholar
- Naito Y (2004) New steps in bio-logging science. Mem Natl Inst Polar Res, Spec Issue 58:50–57Google Scholar
- Quinlan JR (1993) C4.5: programs for machine learning. Morgan Kaufmann Publishers, San MateoGoogle Scholar
- Ravi N, Dandekar N, Mysore P, Littman ML (2005) Activity recognition from acceleration data. IAAI-05, American Association for Artificial Intelligence, pp 1541–1546Google Scholar
- Wyneken J (1997) Sea turtle locomotion: mechanics, behavior, and energetics. In: Lutz PL, Musick JA (eds) The biology of sea turtles. CRC Press, Boca Raton, pp 165–198Google Scholar