Discovering The Main Path Of A Domain Based On The Importance Of Nodes

The trajectory formed by today's field development(including knowledge discipline research and patented technology innovation)can be seen as a connected whole,with sources to be traced,current laws to be explored,and future trends to be expected.Subsequent research and development are based on the inheritance and innovation of previous achievements,in which knowledge information is transferred and grown.The next well-defined strategic plan towards domain's future development requires a global perspective(a quick,clear and comprehensive interpretation of the context of the field)to understand how the field has evolved step by step so far and which key figures and achievements have influenced the follow-up promotion.This paper reviews the domestic and foreign research status of domain development as well as main path discovery mainstream methods,redefines the domain main path and then sets up reasonable and effective discovery models separately from the perspectives of the patents(literature)and the inventors(authors)involved.Before this,the necessity of starting from the importance of nodes was pointed out firstly,considering the important nodes as those with three parallel relations: high influence,strong control and inclusive latest results.Then,combined with the existing achievements,the paper discusses the involved key issues in the main path discovering process under two perspectives and provides with corresponding solutions.First under the patents(literature)perspective,we define the domain development main path as a global-connected path including important nodes to a upper limits involved in the development of the field,reflecting the key inheritance relationships and carrying the maximum amount of information.There are four key issues involved in the discovering process: measurement of key nodes,inclusion of latest results,measurement of the strength of relationship between nodes and simplification of the network to achieve maximum information volume.Corresponding to key issues,the paper proposes solutions respectively.Among them,importance nodes measurement is solved by way of combing revised PageRank and traversal weights;with regard to network reduction,a network is reduced with Pathfinder at first,and then nodes that disturb the explanation of main paths are ruled out,and finally the nodes,referred to as ?middle end nodes?,are bridged to the existing main paths with the method of shortest path,which is beneficial for the connectivity of the final main path.Then under the inventor(author)perspective,path which is able to relatively clearly reflects the key and ordered communication between important inventors in the field refers to the domain exchange main path.It mainly includes the selection of important nodes,the simplification of network,the classification of inventors‘attributes and the orderly visualization.In this regard,the improved PageRank for the author's influence evaluation will be migrated to this point for important nodes selection,extracting the largest connected subgraph based on that;and then with the application of Louvain and Pathfinder algorithm for further subdivision and reduction,nodes are categorised according to their in-and-out degree as: knowledge creators,knowledge transferors and knowledge receivers.Finally,classified nodes are placed with rules in order to understand how knowledge is conveyed and valued more effectively.Finally,in order to consider the rationality and practical effect of model,main path discovery in OLED field is taken as an example for empirical and applied research,comparing with other mainstream methods.It is proved that the results obtained in the paper are more interpretative,comprehensive,scientific,and orderly,showing certain comparative advantages and effectiveness for domain‘s main path discovering.The mentioned methods also provide a new perspective and reference for further research in this area.