Technology Opportunity Identification In The High-tech Fields Based On Data-driven Approaches

Due to its role in driving economic growth,high-tech industries have become an important focus for countries to cultivate new drivers of economic development.As an important stage to realize technological innovation in the high-tech industry,technology opportunity identification is the leading link to carry out technological innovation.However,due to the characteristics of high technology barriers,high investment in research and development,unique breakthrough mechanism,and multiple research and development processes,the data samples of technology opportunity identification are complex and diverse,and the timing is changeable.Therefore,how to accurately predict and judge technical opportunities from the relevant massive data on the basis of efficiently tracking the technological evolution path is crucial to the leapfrog development of China’s high-tech field.Up to now,there are mainly the following problems in the research of technological opportunity identification in the high-tech field: First,the tracking of technological evolution path is mostly based on single data sources such as patents,and less consideration is given to how to obtain robust multi-source data;Secondly,the prediction of technological opportunities pays more attention to the prediction of general technological opportunities and the analysis of technological hot spots,and less selects the future technological innovation opportunities based on the concrete characteristics of hightech fields.Thirdly,the research and judgment of technical opportunities pay more attention to the mining of general technical elements within a single technology,and less to the research and judgment of technical opportunities suitable for specific high-tech fields from the perspective of technical elements.In view of this,this paper focuses on the identification of technological opportunities in the high-tech field,takes "tracking of potential technological evolution path-forecasting of potential technological opportunities-research and judgment of technological opportunities" as the main line of research,combines with data-driven relevant methods,and explores the supporting methods of its core links,which mainly include the following contents:(1)Research on tracking potential technological evolution paths in high-tech fields driven by multi-source data is carried out.Firstly,on the basis of explaining the conceptual structure of the data-driven technology roadmap,this paper obtains the multi-source data in the high-tech field,and uses the BERT model to map the multi-source data layer to the technology roadmap.Then,in collaboration with the "SAO-BTM" approach,the topic content is mapped to the multi-layered framework of the technology roadmap;Further,based on the potential links identified by the link prediction method,the road map of datadriven technology is drawn,and the potential technology evolution path in the high-tech field is tracked.Finally,taking hyperuricemia drug development as an example,the above potential technology evolution path tracking method was empirically demonstrated,and the effectiveness of the technology evolution path tracking method constructed in the hightech field was verified.(2)Research has been carried out to predict potential technological opportunities in high-tech fields based on field characteristics.First,on the basis of explaining the conceptual structure of the two-layer network,this paper collects the relevant data of the evolution path,and uses SAO semantic analysis method to extract the technical components of the high-tech field.Then,the data in the high-tech field is preprocessed,and the technology component topic of the technology evolution path in the high-tech field is identified based on the BTM topic model.Furthermore,the link prediction method is used to identify the potential connections of the technical components in the high-tech field,plot the double-layer network and predict the potential technical opportunities of the technical components.Finally,taking hyperuricemia drug development as an example,the above potential technical opportunity prediction method is empirically verified,and the validity of the proposed method in the high-tech field is verified.(3)Research on the identification of technological opportunities in high-tech fields from the perspective of technological factors has been carried out.First,on the basis of explaining the conceptual structure of the multi-level network,the target technology layer nodes of the high-tech field based on the potential technology opportunities are screened.Then,the analysis dimension of candidate technology nodes in the high-tech field is developed,the technical topics in the high-tech field are identified based on the LDA topic model,the technology roadmap of process embedding is drawn,and the candidate technology layer nodes in the high-tech field are determined.Further,the nodes of the technical elements layer in the high-tech field are mined,the multi-layer network is drawn and the technical opportunities of the technical elements are evaluated.Finally,taking the drug development of hyperuricemia as an example,the technical opportunity research method of the above technical factors mining was demonstrated,and the validity of the technical opportunity research method in the high-tech field was verified.To sum up,this paper not only tracks the potential technological evolution path in the high-tech field based on the data-driven technology roadmap,but also predicts the potential technological opportunities in the high-tech field by using a two-layer network,and then builds a multi-layer network to investigate and determine the technological opportunities of the technological elements in the high-tech field.The aim is to provide reference for decision-making theory and practice for innovative subjects in high-tech field to carry out technological innovation accurately and efficiently.