DB-BiLSTM: Euclidean Distance-Based Sensor Data Prediction for IoT Applications
Made Adi Paramartha Putra, Dong‐Seong Kim, Jae‐Min Lee · 2021 International Conference on Information and Communication Technology Convergence (ICTC) · 2021
This paper proposes a novel approach to predicting future sensor data in IoT applications by utilizing node correlation, namely DB-BiLSTM. Sensor data prediction can reduce unnecessary communication in the network and mitigate the energy issue. Current data prediction only considers Spearman correlation to generate the nodes correlation in the IoT network. Euclidean distance-based could be utilized to get the nodes correlation among the network without going through the previous data of every node in the network. Moreover, the sensor node tends to produce a similar value with the others node nearby. Based on the performance evaluation, the proposed DB-BiLSTM outperforms the existing DL models in four datasets.