Abstract
Recently, agricultural damage caused by monkeys has become a critical problem in Japan. This paper proposes a system that predicts the dates and times at which monkeys approach farmland. In order to make predictions, monkeys were made to wear collars with transmitters, and the trend data of the monkeys' distances were collected by receivers in the system. Two years after this system was installed around a mountain, an average radio communication ability of 97.22% was obtained, and 25 million monkey signals were received. The time of appearance of the monkeys according to those signals as well as the environmental conditions were taken for estimation. Bayesian estimation and support vector machine (SVM) were adopted as linear and nonlinear methods, respectively. The use of Bayesian estimation resulted in little predictive effect. SVM, on the other hand, achieved 31% accuracy by combining two learning pattern methods, outperforming the Bayesian estimation and confirming the effectiveness of environmental conditions in a system of predicting the appearance of monkeys. © 2016 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have