Intelligent Prediction for Device Status Based on IoT Temporal Knowledge Graph

Shujuan You, Xiaotao Li, Wai Chen

Abstract: With the development of Internet of Things (IoT) technology, unattended operation of devices has become an important feature of IoT, which requires devices to perform proper actions to provide services without human intervention. To achieve the unattended operation of IoT, a major challenge is how to accurately predict the actions that the device will perform and meet the personalized requirements of users. To address this challenge, we propose a novel method to predict device status based on the IoT temporal knowledge graph (TKG) and the long short term memory (LSTM) model. We firstly build a TKG for the IoT to provide rich semantic information for the objects and the continuously changing time series data in the IoT. Then, leveraging the advantages of LSTM in sequence learning, the timing characteristics of semantic information in TKG are learned to realize intelligent prediction of the equipment status. To verify the effectiveness of our method, we conducted an experimental verification of device status prediction in a smart home use case, and the experimental results show that our method achieves the state-of-the-art performance.