Research On Identifying Potential Emerging Technology Topics Based On Complex Network Outlier Nodes

In the field of technological innovation,the identification and prediction of emerging technologies has become a research hotspot.It is of great significance for countries,enterprises or research institutions to adhere to innovation-driven development,grasp emerging technology,which is the driving force of technological innovation,take the lead in seizing technological opportunities,and control the direction of future technological development accurately.However,at present,the identification research of emerging technologies mostly focuses on the technology clusters that have begun to take shape and have a certain impact on the industry economy,while ignoring the technology areas that are not obvious at the present stage,deviate from most technologies,but have the potential to develop into the emerging technologies.Therefore,the main purpose of this research is to mine the potential value of the out-of-scale outlier data in related data,help relevant institutions to capture the characteristic signals of emerging technologies in the early stage of technological development,and predict potential emerging technology opportunities in the future.This paper takes the patent data from the Derwent Innovations Index as the research object,and constructs a co-occurrence network of patent technology topics,aiming to identify potential emerging technology topics from the outlier data.The specific research process is as follows:(i)Firstly,the concepts,characteristics and methods of outlier data detection and emerging technology identification are comprehensively investigated,and the advantages and disadvantages of existing researches at home and abroad are sorted out,and the contents and main methods of this research are clarified.(ii)Secondly,according to the definite research ideas,a framework of ranking indicators for the outlier of nodes in complex networks is designed.Based on the defined meaning of outlier nodes,this paper selects three dimensions of connectivity,aggregation and correlation strength from the overall and local perspectives of network topology structure to evaluate the degree of outlier of nodes in complex network,proposes the basic outlier index based on the clustering efficiency and betweeness centrality of the node itself,and the sub-outlier index based on the clustering efficiency and betweeness centrality of its neighbor nodes,and finally get the ranking results of outlier of all nodes in the whole network.(iii)Thirdly,this paper constructs a model for the identification of future emerging techonology topics based on the complex network outlier nodes,which select appropriate trend measurement indicators to conduct patent bibliometrics analysis on the outlier technology topic set corresponding to the selected outlier nodes from the two dimensions of the development characteristics of topics and the bibliometrics characteristics of patent texts,and identify which are the potential emerging technology topics.The accuracy of the judgment results will be verified by means of network research and so on.(iv)Finally,this study uses patent data in the field of nanobiomedicine and nano-robots as source data to construct a patent network based on the co-occurrence between technology topics,and then use the designed framework of ranking indicators for the outlier of nodes in complex networks to screen outlier technology topics in this field,and the model for the identification of future emerging techonology topics based on the complex network outlier nodes to identify which of these may be potential emerging technical topics.After that this paper validate the result of the empirical analysis.Through the design of the index ranking framework for outlier nodes in complex networks and the construction of the identification model for the potential emerging technology topics based on the outlier nodes in complex networks,this study draws the following conclusions.(i)Comprehensive consideration from the three perspectives of network connectivity,aggregation and correlation strength,and considering the size of the outliers of the node itself and its neighbors can better rank the outlier degree of the nodes in the complex network.(ii)It is more strategically instructive to raise the recognition result to the level of the topics than the patent texts themselves in the existing research.(iii)By designing and calculating the measurement index from the development feature of the topics and the bibliometrics characteristics of patent texts,we can detect the emerging technology topics more comprehensively than just observing the change of the bibliometrics feature of patent text.This study establishes a set of methods to identify potential emerging technology topics from outlier patent data,providing a reasonable basis for relevant personnel to mine useful information from outlier data that is easily ignored and unscaled,and predict future potential technical opportunities and development directions.At the same time,the existing method system of emerging technology identification is improved,the deficiency of emerging technology opportunity results which are predicted by only considering the large-scale and gradually emerging patent data is maked up.The main innovations of the research includes the two points.(i)A set of more perfect and flexible framework of ranking index of outliers in complex network is designed,which improves the problems of not being able to sort all the nodes,not fully considering the influence of the whole structure and local structure of the network,and the complexity of calculation.(ii)A model system is established to identify emerging technology topics from outliers in complex networks,and A two-dimensional measure index of development feature of the topics and the bibliometrics characteristics of patent texts is constructed In particular,the indicator of time-varying degree of outliers is put forward,and the result of identification is raised from patent individual to technical subject level,which is beneficial to better grasp the direction of technological development.The paper includes 40 figures,10 tables,90 references and 8 appendices.