Interval-valued Streaming Data Classification Based On Time Series Layering And Its Application Research

In the era of big data,the data collected in practical application fields often show more complex characteristics.As a common form of data,interval-valued data exists widely in many fields,such as finance,meteorology,agriculture and so on.Although there have been some researches on the representation,analysis and mining of interval-valued data,most of them focus on static interval-valued data,and there are relatively few researches on the dynamic interval-valued streaming data which is updated over time in practical application.Due to the time series feature of interval-valued streaming data,the traditional analysis and mining methods of static interval-valued data can not be used directly.To solve this problem,this paper proposes an interval-valued streaming data classification method based on time series layering,which effectively fuses the time series feature and attribute feature of interval-valued streaming data to build a classification model,so as to obtain better generalization performance.The specific research contents are as follows:1.Aiming at the problem that the traditional interval-valued data processing methods can not effectively deal with interval-valued streaming data,this paper proposes an interval-valued streaming data classification method based on time series layering.On the basis of time series layering operation on interval-valued streaming data,a hierarchical interval classification model based on time series layering is constructed.Firstly,the concept of interval number and time series operation are combined to define the union operation of interval-valued streaming data.Through the union operation,the time series layering of interval-valued streaming data is realized,and the data space containing different time series concept layers is constructed;secondly,based on the vertical time series structure of interval-valued streaming data,the hierarchical interval classification model is designed;finally,by fusing the vertical time series feature and the horizontal attribute feature of the data,interval-valued streaming data is classified by cross layer superposition in time series space and attribute space.In order to improve the performance of interval-valued streaming data classification,this method fully considers the time series feature of interval-valued streaming data and measures its impact on sample similarity measurement.2.The interval-valued streaming data classification model proposed in this paper is applied to meteorological streaming data,and a meteorological streaming data analysis system based on time series layering is constructed.The system includes four modules: user management,region selection,time series layering and data analysis.The user management module mainly realizes the function of general user management;the region selection module is mainly responsible for accurately responding to the relevant national and urban meteorological information obtained by users;the time series layering module is mainly composed of method component and function component,the method component encapsulates the time series layering method of this paper to realize data layering,and the function component includes data search,update and merge function services;the data analysis module displays the elements selection and evaluation results of meteorological streaming data analysis model through visualization technology.Using the meteorological streaming data analysis system based on time series layering designed in this paper,the meteorological streaming data can be intelligently managed,analyzed and mined.