Design And Implementation Of Data Mining Support Subsystem Based On Big Data Of Power

Smart grid is the direction and trend of the development of electric power industry. Smart grid using advanced information and communication technology, computer technology, control technology and other advanced technology to achieve the demand of power generation,transmission, using electricity and selling electricity and coordination of functions. Due to the introduction of information technology, the grid will accumulate a large amount of data, these data is known as the big data of power. These data is very valuable, which can not only improve the network self- management and operation level to a new height, and even have a fundamental change, but also provide the conditions for electric power companies to develop value-added services. Therefore, there is an urgent need for the data mining to the big data of power. However, because of the wide range of source and diverse forms of the big data, the low quality of it has become a significant problem. At home and abroad, with the rapid development of data preprocessing technology, data cleaning began to play a role in various industries, to create a condition for improving the quality of the big data of power.In this paper, the design and implementation of data mining support subsystem based on big data of power provides users with a set of data storage, data fusion and data cleaning solutions. Comparing to the previous data processing platform in power grid, the support subsystem has the capacity of big data processing, and has a friendly user interface, can receive a variety of power network data file format, and pay more attention to improving the quality of data. The support subsystem provides a variety of data cleaning methods applied to the power grid, and has a higher accuracy and efficiency.In order to realize the support subsystem, this paper research the related technology at first, and has made clear the feasibility of the system technology. Then in the system needs analysis, according to the data pretreatment process, the paper provides the functional requirements of the system, and analyzed the non-functional requirements. In order to realize the data fusion, this paper puts forward a unified storage mode based on the requirements of power grid environment and data mining, and provides a solution for data fusion. In order to achieve the data cleaning, this paper presents a scheme for the cleaning of incomplete data, using the method of machine cleaning and machine verification, which improves the efficiency of data cleaning. And in the process of cleaning validation, this paper introduces the method of machine learning, to search for the best validation model based on the parameter optimization, to provide a guarantee for the accuracy of the validation. Then, this paper describe the design and implementation of the supporting subsystem, and the test and analysis of its results. At last, this paper has a summary, in which the deficiencies and future development of the support subsystem are proposed.