Social Network Of Trusted Recommendation System Based On Cluster Analysis

Social networks as an important application scenario, the Internet as the network technology more mature and gradually in the form of social transformation. The development of social network in the new opportunities and new challenges. Among them, the effective information acquisition in huge amounts of data becomes a problem urgently to be solved. This article aims to put forward a vast and complex data under the background of social networking recommendation system solutions and ensure the credibility of the recommendation results and the safety of the whole recommendation system.First, this thesis puts forward a kind of multidimensional data processing model and combining the processing results applied to the data of recommendation system selection, recommended in order to achieve the effectiveness of data real-time and efficiency of processing efficiency. Competition- inhibiting node model is divided into four modules, bear data refactoring, data preprocessing, data processing and data precision recommendations. Data reconstruction module will have a unified quantization process different data and eventually become a membrane structure and the characteristics of gene "data nodes, which guarantees the next series of processing. Data preprocessing module in refactoring node, on the basis of protein interaction network was used to analyze the protein function matching degree between method, based on similarity measure, filtering the noise node and introducing the immune mechanism as an aid to improve the accuracy. Precision of data processing stage, the relative density of the data as the matching degree of the precision of the measure, to select the data to estimate the similarity between the target node.Secondly, in view of the credibility of the social network node and the credibility of the recommendation results. A peer feedback trust mechanism, through to the main body and object node trust two-way evaluation to enhance the credibility of the whole system. The trust of the subject is provided by the object node, at the same time, the object of trust will be the subject and the object node, other feedback and influence, so as to complete the two-way constraints. Regulates the balance of trust in the whole system and improve the effectiveness of the trust.Then, in view of the security problems of social network information transmission, this thesis proposes a control risk in the social network approach for the rapid spread of malicious information, active monitoring by honeypot technology for malicious information and find the key nodes in the network link prediction method and carries on the effective isolation measures to prevent the mass propagation of malicious information in social networks.Finally, combined with the data processing algorithm and security algorithm, design the prototype of a social network recommend system, implement a simple social network recommendations.