The Research Of Trust Node Mining Method For Social Media Platform

Nowadays more and more fields and applications require the support of key node mining technology.Especially as the development of Mobile terminal technology we greeted new eraWeibo 2.0 times.The new challenges of computer field in large complex social networks are investigated based on date.Meanwhile complex networks node influence power of mining algorithm as the core of the large social media have attracted a large number of academicians research and innovation.As the researchers continue to make up for the faultiness of traditional algorithms,the accuracy and serviceability of existing model needs to be improved.This thesis points out that prolongation user node structural and behavior property at the same time add to the trusting relationships sign on the network that results improve the accuracy of the model.In order to contribute our share to influence of social network node mining field.During the actual social media users not only have their own static nature but interact in behaviour and substance with other users moreover emotion is produce.However most of the traditional algorithms consider the importance of node the discrepancy between the truth and it bring about the accuracy of the influence power of node mining results are not high level in view of the above this thesis proposes a method of using trust element to construct a symbolic network that to perform influence power of node mining way.First of all the interference of “water army” and influence power of node with higher behavior quality are often ignored.The user's concordance between defining the user and behavior worth concept by means of prolongation node in structural and behavior property.Based on the H-index algorithm,a new computational method named Hmining is proposed.The iterative analysis method of H-factor is adopted to reduce the interference and improve the mining quality.Secondly,the concept of special trust is constructed in sociology on the basis of the above model,the HT-mining calculation method is proposed to improve the accuracy of the method under the premise of ensuring the time complexity of the algorithm.Finally,the Weibo data set is used to test the effectiveness of the proposed method.And feasibility.The experimental results show that the trust node mining method proposed for social media platform has higher accuracy and lower time complexity for mining large-scale complex social network influence nodes,and is active in the face of users in the network.High accuracy is also maintained when the time period or hot topic of high user engagement is high.Compared with the traditional PageRank influence node mining algorithm and H-index based HRANK calculation method,the trust node mining method proposed for social media platform in this paper guarantees the accuracy of the calculation result when the time complexity is O(n).There is a big improvement.In order to verify the feasibility of this algorithm,two different face databases are selected to test and evaluate the proposed algorithm.Compared with the PCA+KNN,the SVM and the based convolutional neural network algorithm,the proposed algorithm in a short training time,the classification accuracy is greatly improved,which verifies the rationality of the algorithm.