Research And Implementation Of Recommendation System Based On Trust In Social Networks

With the popularity of the Internet,social networks have gradually become the main platform for people to interact with each other.Therefore,it is important to study the issue of social networks.This thesis makes a deep study on social network trust in the recommended systems,focuses on the path search algorithm,graph matching algorithm and related key technologies.The social trust recommendation system is implemented by using JAVA language.To solve the technical problems such as the display of graph data,the input of pattern matching in graph matching,the fast searching path and graph matching in large scale graph data,specific contents include the following aspects:(1)Analysis of existing trust recommendation algorithms and related technologies.In this thesis,we study the concept of social network structure and the quality of trust in social networks.Based on D-MCBA algorithm,we develop the Multiple Trust Paths(MTP)selection system.The path recommendation system consists of three recommendations,the shortest path,the maximum T value path and the Top-K optimal paths,to solve the problem of quickly finding multiple trust paths in large scale graph data.(2)Based on MC-GPM algorithm,we develop the Multi-Thread Graph(MTGraph)system.In this thesis,we study a graph matching algorithm called Multi-Constrained Graph Pattern Matching(MC-GPM)based on multiple constraints and strong social relations.This thesis makes a concentrated research on the method of finding the graph data,and the corresponding characters or groups with specific relationship which can be transformed into the matching problem of the specified node and subgraph.This system displays the social network structure and participants' basic information,solving the technical difficulties of the display of social networks and the user's input to the pattern graph in graph pattern matching.In addition,the system adopts multiple threading technology to efficiently finding the graph pattern matching.Our systems realize trust paths and trust graphs matching recommendation,helping assess the trust relationship between two unknown participants.There are numerous potential applications of our system in real life.For example,the trust paths recommendation could help retailers find potential customers or help users find trustworthy works in the crowdsourcing platform.The trust graph matching recommendation could help people find experts or groups in a specific field.Therefore,recommendation service of our system has important practical significance to people's life.