AN APPLICATION OF ARTIFICIAL NEURAL NETWORK ON WEB OF THINGS
Real world with embedded and globally networked sensors has led to the development of Internet of Things (IoT), which contains much information of things in the world and becomes an important part of Internet. As an important extension of IoT, Web of Things (WoT) standing for an intelligent web based information platform built on IoT, has also been studied to be a bridge between IoT and traditional Internet. Currently there is little study on data mining on WoT, since data from WoT is multi-source, unstructured, temporal and noisy, which brings great challenges to traditional learning methods. We propose an Artificial Neural Network (ANN) based learning framework to build an intelligent data analysis application upon WoT with sliding windows, aiming at discovering the underlying knowledge of information gathered from an RFID based sensor network. The ANN definition is extended to process temporal data gathered from the sensor network. Experiment result on a 6-month-long dataset from WoT application shows the effectiveness and superiority of the proposed method.
Web of things Internet of things Artificial neural network Temporal data analysis Knowledge discovery
Q.Pan Z.H.Liang Z.P.Li G.Zhang X.Y.Zhang Z.M.Huang
Faculty of Economics and Management, Zhuhai City Polytechnic College, Zhuhai, 519090, China These au The 2nd Affiliated Hospital, Guangzhou University of Chinese Medicine, Guangzhou, 510120, China Thes The 2nd Affiliated Hospital, Guangzhou University of Chinese Medicine, Guangzhou, 510120, China Faculty of Automation, GuangDong University of Technology, Guangzhou, 510006, China
国际会议
2012 International Conference on System Simulation(2012年国际系统仿真学术会议)
上海
英文
392-396
2012-04-06（万方平台首次上网日期，不代表论文的发表时间）