Research On An Application Of Data Stream Query And Data Stream Mining In Oil Field

The research of data stream mining and analysis is new and hot in the fields of data miningand analysis. There are many applications of data stream in many fields , such as Performancemeasurements in network monitoring and traffic management;Call detail records in tele-communications;Transactions in retail chains, ATM operations in banks and Sensor networkdata. Because of the particularity of data stream :the trend of varying with time,rapid arrival ofmassive data in short time,it is very important, at the same time a chance and challenge for datamining how to deal with these massive data to capture the useful information using the limitedstorage over these data stream.With the constant development of the oil field informationization,A large number of data ofoil field have realized progressively that the network upload. It is urgent thing to develop thenetwork monitoring system with muti-server and mastive data,which can provide the method ofmonitoring the information quality and management and can know Comprehensivly the user'susage of database through replacing pattern of manual information management of programme.The following is the main research work of this paper:(1) This paper conclude the basic character,model and key technology et about data stream,data stream query,data stream mining and DSMS through the research of data streamtechnology.(2) This paper explain the Design and realization of network monitoring system of NO.2factory of oil field, which save the alarming condition into XML Document;monitoring andalarming over the unsual condition of database with ORACLE audit ,trigger,job function . Atthe same time,this paper adopt the different level alarming combining with real time andtime-delayed because of the contradictions of time and rate.(3) This paper takes emphasis on extracing unsual pattern., in which the condition of thresholdis experiential of oil field and the expansibility of system is increased with XML combining withdata stream system. This paper introduced a Learning Environment-YaLE and concept driftplugin which extends YALE by Operators for handling real and simulated concept drift intime-varying data streams. A concept drift experiments is structured simulating a timely orderand concept drift on a data set in this paper.(4) This paper explain the Design and realization of network monitoring system of NO.2factory of oil field, which save the alarming condition into XML Document;monitoring andalarming over the unsual condition of database with ORACLE audit ,trigger,job function . Atthe same time, this paper adopt the different level alarming combining with real time andtime-delayed because of the contradictions of time and rate. This paper provided the experienceand thinking over DSMS and data mining through the real project.