Research On Technique And Application Of Mining Data Streams

Data stream technique is becoming hot in the field of data processing. Mining data streams is one of the important branches in data stream researches. The focus on mining data stream is to design one-pass scan algorithm over data set, and maintain a few synopsis data in memory. The synopsis data could express the main characteristic of data stream, and it is far smaller than the size of whole data stream.The synopsis data is used for the real-time and online analysis of data streams. The clustering and classification of data streams are widely used in many fields. This paper presents a reasonable model of mining data streams. The synopsis data structure and its incremental algorithm are designed. Traditional technique of data mining is improved to realize the real-time and online processing. Frameworks of the application in power load analysis and online power quality disturbance classification are proposed on the basis of mining data streams, and methods related to them are given. Experiment results show the effectiveness of the presented methods.