Research On Feature Selection Application Of Blockchain And Deep Learning Technology In Energy Internet

Due to the continuous development of human economy and society,the total amount of global energy consumption continues to grow.However,with the environmental pollution caused by the shortage of fossil fuels and emissions,the industrial mode of burning coal resources in large quantities is coming to an end,and the application of energy Internet with the deep integration of new energy technology and information technology is emerging.A large number of smart devices are deployed in the energy Internet.With the rapid development of information technology,the multi-source heterogeneous data containing meaningful information have been significantly generated by various edge devices in Internet of Energy,which is one of essential foundations of many knowledge discovery tasks based on edge computing.For some complicated tasks,essential features are owned by different data sellers offering data by blockchains.With limited budgets,buying features are crucial steps in knowledge discovery tasks in energy Internet,especially for learning based algorithms.Features generation and repair can effectively improve the utilization and availability of their own data,thereby reducing the cost of payment.However,there are lack of proper data pricing mechanisms tailored to dynamic learning processes.Besides,existing methods cannot efficiently employ edge computing servers to obtain optimal policies for selecting features according to dynamic pricing with limited budgets.With the influx of intelligent edge devices in energy Internet,security issues such as data security and identity privacy become more important.Different devices or terminals have different identities,permissions,and data.They need to use real identities for transactions.To overcome the drawbacks above,in this paper,a framework based on blockchain for buying features with limited budgets using edge computing in electricity spot markets is proposed.Based on this framework,static and dynamic data pricing are considered.The key contributions of this paper are as follows.A knowledge discovery framework based on data dynamic pricing and distributed feature selection is proposed,which can be deployed in the environment of edge computing.The proposed framework addresses the problem of how to generate and purchase data when users lack data and have limited budgets.At the same time,in order to protect the security and privacy of data transaction,consortium blockchain is employed when data sharing and the users authenticating.For the data sellers,a pricing mechanism consisting of static and dynamic pricing is proposed.Static pricing takes into account key factors such as incomplete data sets,repeated historical queries and data set update.And dynamic pricing is adjusted according to the accuracy of the buyer's model,which can be changed by the model accuracy and user query behaviors.For the data buyers,the problem of feature selection under given budget is proposed,which takes edge computing delay and blockchain delay into consideration.Given limited budgets,a feature selection algorithm considering multiple new factors is proposed,which offers near optimal solutions for feature selection at different scenarios.