Reaserch On Heterogeneous Information Aggregation Technology From Internet Of Things Based On PML

Recently, a variety of data from Internet of things are collected by RFID readers real-timelyand different types of sensors which spread in the world. In addition, the process of acquiring thosedata couldn’t complete in only one operation rather than continue to acquire those data, and thesedata are called data stream that has real-time, continuous and order freatures.For the problems of massiveness and heterogeneity of sensor data, some researches in the pastcouldn’t be used here. In Internet of things, the XML (eXtensible Markup Language) techonologyof integrating various data from different platforms has been replaced by PML (Physical MarkupLanguage) technology, which means PML technology is more adaptive than XML in Internet ofthings. Also, some methods of XML clustering focus on static data, but couldn’t be used to massive,real-time and heteroideous data stream from Internet of things. Furthermore, there are somedifferences between XML documents and PML documents. So it is unsuitable to adapt sometechniques of XML clustering into PML documents.Hence, this thesis has proposed a method of heterogeneous information aggregationtechnology from Internet of things based on PML. Firstly, the thesis shows a model ofheterogeneous information aggregation, and describes four layers in detail, including perceptionlayer, transmisson layer, processing layer and application layer. Secondly, the thesis also designs amethod of measuring the similarity of PML documents based on Bayesian Network according to thefeatures of PML document in processing layer. Moreover, the thesis presents a solution to clusterdata stream of PML documents based on dynamic sliding window technology. Finally, someexperiments could prove that the methods of measuring the similarity of PML documents based onBayesian Network and clustering data stream of PML documents based on dynamic sliding windoware more effective and accurate than current methods.