Data Preprocessing And Feature Analysis For Real-time Data Stream

With the rapid development of network technology and the popularization of its applications in daily life, all of the colleges and universities have established campus networks and put them into use, which is considered as an important infrastructure. The campus network security influence much about education activities, thus most colleges have strengthened the network management among which the network data analysis is one of the most effective means.Meanwhile, the intensive study of real-time data streams has spread into many fields, so it can be used for campus network management. The technology is helpful to pre-process the data of campus network access and carry out feature analysis, so that the running conditions and service conditions of campus networks can be grasped comprehensively. The application is very significant and will help to adopt an effective measure to enhance the network management reform for higher efficiency.This thesis first introduces the interrelated concept and technology of data pre-processing and feature analysis, especially data scrubbing, data integration, data transformation and data aggregation. Then, the features of pre-processed data of campus network access are analyzed to obtain the information about the running conditions, variation tendency and general characters. Finally, the evaluated information is as a guide to carry out the corresponding features and manage the campus network with higher efficiency.