| Social network is penetrating into national security,economic development,social life and other aspects,and its value of depositing massive data gradually appears.The research and application of data mining and protection algorithm around social network are more and more valued by people.Compared with the traditional web,social networks have a wide range of data types,large proportion of unstructured data,and more complex data calculation,storage and transmission methods.It is difficult to achieve real-time,accuracy and applicability for social network data mining.In addition,social network users’ privacy data is seriously lost,which is in urgent need of protection.In view of the above problems,this thesis carried out the research and optimization of the key technologies of data mining and protection for social networks,and specifically carried out the research in the following four aspects.1.Aiming at the accuracy of user behavior attribute data mining algorithm in social network,a method of user behavior attribute data mining based on association rule data clustering is proposed.By analyzing the data storage mechanism of social network,the data distribution model of social network is constructed.The interference information and redundant information in social network are filtered by information filtering algorithm.By using the feature extraction of association rules and fuzzy data clustering technology,the associated data mining and clustering in social network are realized.The simulation results show that the algorithm has high precision and reliability,improves the community discovery ability,and has good application value in social network recommendation.2.Aiming at the real-time problem of social network hotspots and sensitive data mining algorithm,a real-time social network data mining method based on high-order spectral feature fuzzy neural network is proposed.The transmission channel model and statistical time series model of social network data are constructed,and the redundant information flow is removed and reprocessed.High order spectral feature extraction and fusion clustering analysis based on fuzzy neural network are carried out on the redundant filtered Web data.Based on the above work,the data mining model is improved to realize the optimal mining of social network data.Simulation results show that this method has the advantages of good real-time performance and high mining precision.3.Aiming at the applicability of social network multimedia data mining algorithm in cloud environment,a spatial network distributed data mining algorithm based on structure dynamic optimization BP neural network is proposed.On the basis of constructing the spatial network data and analyzing the data structure,the redundant data is compressed by extracting the time-frequency features.Combining with the adaptive matching filtering method,the feature matching of the data is carried out,the spatial spectrum feature extraction method is used to locate the characteristics of the multi-input and multi-output spatial network data,and the time.series data is reconstructed by the fourth-order cumulative slice.In order to improve the accuracy of data mining,the structural dynamic optimization BP neural network is used to classify and identify the extracted data features and realize the optimization of data mining.The simulation results show that the method has good mining precision and mining process convergence,and better practical application value.4.Aiming at the problem of user attribute data security protection and the impact of protection algorithm on the performance of data mining algorithm,a data protection algorithm based on dynamic cyclic encryption and link equilibrium configuration is proposed.Build the architecture model and routing control protocol of social network,sub-key random modulation method is adopted to realize data encryption in a social network,the dynamic cycle encryption algorithm to encrypt and transfer data.The method of link equilibrium configuration is used to conduct adaptive equilibrium scheduling for the data output of social network.Experiments show that this method has good encryption ability and improves the storage and transmission of data. |
