The Path Query Technology Research Of Complex Multiple Attribute Graph Data

With the continuous growth of graph data size and the increasing of attribute factors,path query technology faces the problems of complexity of attribute itself,complex relationship be-tween attributes and large-scale graph data.These problems lead to the conditions that many existing path query algorithms can not meet the actual needs,so the multi-attribute path queries attract more and more attention.In this paper,the comprehensive evaluation of complex multi-attribute path,especially the uncertain fuzzy attribute path and the query optimization of large-scale graph data are important innovation fulcrum.Based on the existing theory and technology,this paper is focusing on path reachability,optimal path and TOP-K path to carry out the work.The graph data reachability query is a basic path query problem,but existing algorithms is unable to meet the requirement of path reachability query in large-scale and complex multi-attribute.In order to solve this problem,the paper proposes a TCRQDG algorithm.First of all,because some algorithms of graph data seldom simultaneously concern nodes and edges,the article put forward a new method which expands the graph by using the technology of virtual nodes so that it can realize comprehensive consideration and evaluation to the information of node and edge.Secondly,in view of large scale data,we propose a method to eliminate the unqualified path by filtering,and then reduce the number of edges.Thirdly,because the results of single attribute obviously can not meet the multi-attribute requirements of users,this paper analyzes the relationship among attributes,and evaluates the influence of each attribute of path by principal component analysis.The fourth,due to the different types and characteristics of the path attribute value,each path is comprehensively evaluated by the technology of complex multi-attribute decision making.The fifth,aiming at the integrity of the path information,the circle constriction technique is designed to ensure that the edge connection information is not lost.The sixth,on the basis of comparison and analysis of large-scale reachability query technology,the filtered multiple interval label technique is used to realize the query between source node and destination node.Finally,the performance analysis and experiment prove that the method proposed in this paper can solve the reachability queries of complex multi-attribute.The optimal path query of graph data is another basic problem of path query.However,due to the increasing size of data,the path has more and more uncertain attributes.How to deal with these attributes and data size become a new problem of optimal path query.In this paper,we design two algorithms to solve the query of optimal path for the attributes of mixed data of deterministic and uncertainty,and pure language value as well.The first algorithm aims at the different types and characteristics of deterministic and uncertainty in complex multiple attribute situation.This paper uses the method of information entropy and subjective evaluation to work out the objective weight and subjective weight of each attribute.Then it comprehensively analyzes the two kinds of weight to calculate the comprehensive score of each path.Secondly,based on the comprehensive score of the path,the article aims at time and space requirements of large scale graph data.It uses the graph decomposition and hierarchy shrinkage technology to reduce the search space of the optimized path query,and uses the bidirectional search technology to accelerate the query process.Because the uncertain attributes of language values like the inclination of psychology could hinder the path query,on the premise of diversification,the second path query algorithm uses bias function and multi-objective optimization techniques to obtain the weight of language attributes.Besides,the comprehensive evaluation of each path of both single-decision and multi-decision users is calculated by LWAA technology.Secondly,in the case of large-scale data,this paper uses landmark and community technology to optimize the path query,and improves the efficiency of query.Finally,by analyzing the experimental results,the algorithm in this paper can effectively achieve the optimal path query of the pure language value attributes.The TOP-K path query of graph data is an extension of the optimal path query.It provides multiple and alternative paths for people.As the demand for decision-making users continues to grow,TOP-K path query faces the effects of complex multi-attribute.This paper analyzes the advantages and disadvantages of TOP-K algorithm commonly-used.And considering the different types and characters of graph data attributes,it use three different methods to solve the TOP-K path for three different types of attribute respectively.The first algorithm aims at diffi-culty of comprehensive evaluation of uncertain value and precise values,based on the analysis of the characteristics of interval numbers,a comprehensive evaluation of each path by extreme and TOPSIS technology.Secondly,by means of analysis,we find that the core of TOP-K path algorithm of deviation path is the Dijkstra algorithm which be called several times.To opti-mize algorithm of deviation path,this paper uses graph decomposition and bidirectional search technology.By using experimental analysis,the algorithm of this paper can better finish the TOP-K path query with mixed attributes.The second algorithm analyzes each attribute of af-fecting actual traffic.It aims at fuzzy character of many attributes.This article analyzes the characteristics of each fuzzy attribute,and evaluates the path with fuzzy decision technique of subordinating degree function,composition operator and so on.Secondly,according to the situ?ation of large scale graph data,the genetic algorithm and deviation path algorithm are integrated in the paper to optimize TOP-K path query.By analyzing the experiments and performance,the fuzzy multi-attribute genetic algorithm of this paper can solve the TOP-K path query of fuzzy attributes better.The third algorithm analyzes the hesitant fuzzy language set of decision users at the beginning,because the language set reflects the complexity of path and fuzziness of user thinking.So this algorithm aims at these features.And it obtain objective weight and subjec-tive weight by using information entropy technology and subjective evaluation method.It also solved the path comprehensive evaluation for hesitant fuzzy language set by path matrix conver-sion and TOPSIS technology.Secondly,for large scale graph data,we use landmark,community and priority queuing techniques to implement TOP-K path queries.Finally,by experiments and performance analysis,the algorithm can solve the TOP-K path query of hesitant fuzzy language attributes better.