Research On Pattern Recognition And Pattern Matching In Internet Of Things

The Internet of Things has become a basic technology, which is closelylinked in people’s life and work, with the rapid development of science andtechnology. The Internet of Things is no longer a theoretical hypothesis which isfar away from us. It has been a project which is good for people. The Internet ofThings has played an important role in the practical application, and it will bemore important. Internet of Things has been greatly expanded in the rapidDevelopment of science and technology in some field. The Internet of Things hasbeen widely used in the field of city management, smart home, navigationpositioning, logistics management, food safety, retail, medical digital andnetwork security etc.The system, which the Internet of Things is in, is complex. Datatransmission is susceptible to be distracted. To keep data reliability, stability andaccuracy is difficult in transmission. There is no method which has provided apowerful guarantee at present. A method is proposed in my paper aiming at thisproblem. A research on two cases is carried out for the problem of missing data.The Time correlation of the same node is reduced because of a longer cycle timein collecting data. Data, which is collected by adjacent nodes, is regarded asparameters, which are used to estimate missing data by multiple linear regressionmodel. Dynamic changes are used to estimate missing data under the timesynchronization. In the condition of less nodes,spatial correlation of adjacentnodes is reduced because of less nodes. The node, which has missing data isregarded as the only data source, is proposed in my paper, aiming at this problem.The data, which is collected by the node, is regarded as parameter to estimatemissing data by multiple linear regression model. Different types of data is usedto estimate the missing data dynamically.Missing data, which is caused by theenergy consumption and stabilization of the network, is solved. The resolving of different types of missing data is improved. And pattern recognition and patternmatching technology are used to make classification and prediction. Accuracyand scope of application of estimating missing data is developed. Greenhouse isthe physical environment in this paper. We use my method to estimate missingdata. And good results are archived.