Research On Bank Information Interactive Service Composition Algorithm Based On Knowledge Graph

With the rapid development of the information age,new collaboration needs continue to appear in various industries all over the world.Traditional software applications are no longer applicable to the business needs across multiple collaboration industries because of their single function.Therefore,integrated network collaboration plays an increasingly important role in various industries.Web service composition has become the development trend of information industry.How to solve the inefficient and high-cost collaborative system design caused by business differences among different industries,communication protocol differences among multiple network systems,and language and cultural differences has become a difficult problem.This treatise studies the combination of service composition and bank financial business,The knowledge graph which combined with unstructured data sets and constructed by a graph relationship based on semantics,business elements and functional services is used to describe the point-to-edge relationship of business services.It also introduce a graph search algorithm to improve the efficiency of service adaptation and combination and reduce the cost of scene service support.Firstly,a collaborative filtering recommendation model with knowledge graph representation learning is proposed to alleviate the problem of data sparsity.Secondly,a service composition algorithm based on knowledge graph is proposed to improve the success rate and execution efficiency of service composition algorithm.Finally,the service combination algorithm based on knowledge graph and recommendation model are verified by the actual bank transaction data set and bank service example.The experimental results show that the introduction of financial transaction text and the recommendation algorithm in knowledge graph can effectively alleviate the problem of data sparsity.The proposed combinatorial optimization algorithm based on knowledge graph has the potential to explore the relationship between knowledge graph and service topology graph.Compared with traditional methods,its accuracy and execution efficiency are improved by about 17.7%and 6.3%respectively.In the research process,two optimization designs are proposed.The first one is the optimization of the collaborative filtering recommendation model,which uses the tool of knowledge map and representation learning technology,projects the knowledge map coding vector of business information and the coding vector of transaction information into the collaborative filtering algorithm,and finally decomposes the original scoring matrix.The optimized algorithm is used to reduce the matrix dimension and solve the problem of data sparseness.The second one is the introduce of the knowledge graph into the service composition algorithm,which reflects the user behavior pattern,embeds the service entity into the high-dimensional matrix,and performs matrix operation to obtain the service composition pattern.This algorithm can perform reinforcement learning tasks under the knowledge graph,and use topological relationship modeling to rearrange and send continuous input space instead of discrete nodes,so that it can meet the effective reinforcement learning path based on path search paradigm under multiple constraints,and improve the power and execution efficiency of service composition algorithm.