Research On Patent Data Organization And Knowledge Discovery For Tracking And Predicting Emerging Technologies

The 27 th meeting of the Commission for Deepening Overall Reform of the CPC Central Committee presided over by the General Secretary pointed out that we should improve a new nationwide system for tackling key and core technologies,identify the main areas of attack and breakthroughs in core technologies,and focus on the research and development of key technologies with first-mover advantages and basic frontier technologies that guide future development.In April 2022,the RAND Report pointed out that emerging technologies will have a certain impact on strategic deterrence.From the perspective of breaking the strategic deterrence of the United States,emerging technologies,especially the combination of emerging technologies,will help China far more than other countries.As an important part of technological innovation,the identification,tracking,predicting and management of emerging technologies have been widely concerned.Scientific and accurate tracking and predicting of emerging technologies has become an important basis for the establishment of technological innovation system and the core position of scientific and technological self-reliance.Patent data is an important crystallization of wisdom in the process of technology development and innovation,and is of great significance to the research in the fields of technology innovation development,technology layout,emerging technology identification and so on.Therefore,patent data organization and knowledge discovery have become the realistic demand and scientific direction of technology innovation.In order to deeply explore and analyze emerging technologies,this paper takes patents,emerging technologies and their close relationship as the center,analyzes the dimensions and characteristics of patent components and emerging technology tracking prediction,and builds a mechanism model of patent data organization and knowledge discovery for tracking and predicting emerging technology.Then,using the advantages of knowledge graph in understanding of big data,data relations and knowledge organization,patent data is organized and analyzed to complete the advancement of patent data to technical knowledge.Combined with literature analysis,presentation learning,machine learning,empirical research and other methods,the identification,evolution and prediction models of emerging technologies based on patent knowledge graph are constructed respectively.Finally,it puts forward the guarantee mechanism and strategy of patent data organization and knowledge discovery oriented to the tracking and prediction of emerging technologies,so as to better empower the development of national technological innovation.First of all,Chapter 3 makes clear the components of patent data,the dimensions and characteristics of the tracking and prediction of emerging technologies,and analyzes the relationship between patent data organization and knowledge discovery,patent and emerging technology,patent data organization and knowledge graph,patent knowledge discovery and the tracking and prediction of emerging technologies in detail.The process of patent data organization and knowledge discovery oriented to tracking and forecasting of emerging technologies is analyzed,and the mechanism model of patent data organization and knowledge discovery oriented to tracking and forecasting of emerging technologies is constructed,which provides a theoretical main line for the development of the following chapters.Secondly,in chapter 4,the knowledge graph is used to realize patent data organization,complete the advancement of patent data to technical knowledge,provide data support for chapter 5,6 and 7,and carry out empirical analysis based on the patent data of the new energy automobile industry.Then,in chapter 5,6 and 7,according to the dimension of the tracking and prediction of emerging technologies,the emerging technology identification model,emerging technology evolution model and emerging technology prediction model based on patent knowledge graph are constructed respectively,and combined with the patent data of new energy automobile industry,the empirical analysis is conducted to verify the validity of the model,so as to find the characteristics and rules of emerging technology identification,evolution and prediction and lay the foundation for Chapter 8.Finally,on the basis of the previous chapter,Chapter 8 puts forward the guarantee mechanism and strategy to ensure that the patent data organization and knowledge discovery oriented to the tracking and prediction of emerging technologies can better serve the development of national technological innovation.The main research contents are as follows:Firstly,a mechanism model of patent data organization and knowledge discovery for tracking and predicting of emerging technologies is constructed.To clarify the components and characteristics of patents and emerging technologies,that is,which patent data resources can more fully represent emerging technologies,and vice versa,which dimensions can better track and predict emerging technologies.On this basis,the correlation between patents and emerging technologies,patent data organization and knowledge discovery is further sorted out.At the same time,the patent data organization oriented to the tracking and prediction of emerging technologies and the patent data resource acquisition stage in the process of knowledge discovery,the patent knowledge map-oriented data organization stage,the emerging technology tracking and prediction-oriented knowledge discovery stage as well as the interaction relationship and principle between each stage are analyzed clearly.Based on the above content,a mechanism model of patent data organization and knowledge discovery for tracking and predicting of emerging technologies is constructed.To realize the global revelation of the whole research logic of patent data organization and knowledge discovery oriented to the tracking and prediction of emerging technologies,and to provide a theoretical basis and a research framework for in-depth analysis of the subsequent empirical research of patent data organization and knowledge discovery.Secondly,the patent knowledge map should be constructed.The knowledge graph is used to realize the organization of patent data,and the original fragmented patent data is fused and correlated to form a complex semantic network.Based on the interconnection of patent technical knowledge,the patent data resources can be stored and represented visually,the complex and diverse relationships among patents can be revealed,the patent data can be advanced to technical knowledge,and the integrity and relevance of technical information can be understood and analyzed easily.Firstly,the principles and objectives of patent knowledge map-oriented data organization are clarified.Secondly,the domain ontology model is constructed,attributes are created and examples are imported.Taking the new energy automobile industry as an example,the patent knowledge map is constructed based on Neo4 j.Finally,by using the retrieval and reasoning functions of the patent knowledge graph,the patent data is deeply mined,and the technical knowledge discovery scenarios such as patent holographic file visualization,association technology analysis and co-occurrence relationship tracing are implemented,laying a foundation for index feature extraction,evolutionary feature extraction and prediction feature extraction of emerging technology identification.Thirdly,an emerging technology recognition model is constructed which integrates patent knowledge graph and deep neural network.On the basis of in-depth analysis of the characteristics of emerging technology,further understand the attributes of emerging technology,so as to clarify the principle of emerging technology recognition.At the same time,centering on the three main characteristics of emerging technologies,namely novelty,social impact and fundamental innovation,a complete and quantifiable index system of emerging technologies is constructed from the three dimensions of "novelty-application range-development ability".With the patent knowledge map constructed in Chapter 4 as the data support,the complex semantic information of the patent knowledge map is used to extract the characteristic values of each index,and the recognition model of emerging technologies is trained by the deep neural network algorithm and other comparison algorithms,so as to realize the accurate identification of emerging technologies.Fourthly,an emerging technology evolution model integrating patent knowledge map and presentation learning is constructed.Firstly,the patent citation network,IPC co-occurrence network and patent technology semantic information are extracted on the basis of patent knowledge map,and transformed into emerging technology evolution feature vector by presentation learning correlation algorithm.Secondly,the k-means clustering algorithm is used to identify technical groups,and then the topic feature vector of technical groups is extracted,and the technical groups in adjacent time intervals are associated by cosine similarity.Finally,the evolution path and heat change of the identified emerging technology groups in the new energy automobile industry are analyzed respectively from two aspects of the evolution of the technology group topic and the trend of the technology group topic heat change.Fifthly,a prediction model of emerging technologies based on patent knowledge map and machine learning is constructed.Based on the internal mechanism of technology fusion,the future development situation analysis of emerging technologies can be realized by predicting the possible future convergence trend among the emerging technology fields that have not been integrated before.First of all,from the two perspectives of similarity characteristics and fusion index characteristics,the characteristics of emerging technology prediction are defined,including network structure similarity,node relationship similarity,node attribute similarity and technology fusion characteristic index.Then,on the basis of patent IPC co-occurrence network,the similarity features and fusion index features are extracted by link prediction,association rules and DEMATEL,and technical semantic association.On this basis,IPC pair feature vectors are constructed.Relying on KNN,SVM and NB algorithms,the emerging technology prediction model is trained,and the voting classifier is used to determine the final prediction result.Sixth,the guarantee mechanism and strategy of patent data organization and knowledge discovery for tracking and predicting emerging technologies are proposed.Starting from the service scenarios of enabling national technological innovation,facilitating key technology research,and supporting competitive advantage grabbing,this paper provides a "base" for patent data organization and knowledge discovery oriented to tracking and forecasting of emerging technologies to play their value.Through "two-way enabling" as the core of the patent data resource construction mechanism,"digital intelligence education" as the core of the cross-disciplinary talent support mechanism,"technology innovation" as the core of the intelligent technology iteration embedding mechanism and "platform building" as the core of the science and technology innovation platform ramming mechanism,finally formed "resource construction-talent support-technology embedding-platform support" four-in-one specialization Benefit data organization and knowledge discovery service enabling guarantee mechanism system,in order to better realize the national technology innovation and development of the intelligence and empowerment.On the basis of the safeguard mechanism,from the perspective of synergistic linkage,technology development and multi-subject perspective,further put forward the corresponding safeguard strategy,more specific and comprehensive implementation of the safeguard mechanism,to achieve better patent data organization and knowledge discovery effect,deepen the in-depth understanding of emerging technologies,and provide more sufficient auxiliary reference for technological innovation and development.This study tries to enrich the research system of patent data organization and knowledge discovery from the theoretical level,and deepen the integration and application of patent data organization and knowledge discovery in the field of emerging technologies.At the practical level,this study starts from the patent data to accurately identify emerging technologies,clarify the evolution path of emerging technologies,and predict the future integration trend of emerging technologies,in order to provide scientific basis and practical reference for the technological innovation development and management of the country,industry and enterprises.