Irawan Dwi Wahyono, Khoirudin Asfani, Aripriharta, Irham Fadlika, Gwo Jia Jong
In recent years, floods have often occurred in Indonesia. The impact of floods must be mapped and informed to reduce the impact of flood victims. The technology used to obtain flood information is a wireless sensor. Uneven wireless sensor distribution and wireless sensor resource conditions make the information obtained sometimes inaccurate. Several studies have conducted research on wireless sensors in uneven distribution. This study develops information on potential flooding in flood-prone areas. This study uses wireless sensors that are scattered along potential flood rivers as data senders. Data sent in the form of rainfall data, river water discharge, and river water level. These data are processed using the hybrid Artificial Intelligence algorithm, namely the k-NN algorithm and the Naive Bayes algorithm. The results of this data processing are in the form of three prediction conditions, namely floods, early warnings, and not floods. The results of the implementation of the system with the hybrid algorithm obtained prediction accuracy rate of 93,4%. This system has an error of 6,6% when the system cannot predict wireless sensor data. © 2019 IEEE.
Electrical Engineering, Universitas Negeri Malang, Malang, Indonesia; National Kaohsiung University of Science and Technology, Kaohsiung city, Taiwan
