Research On Identification Of Emerging Technologies Based On Evolution Analysis Of Patent Novelty

Emerging technologies play a vital role in promoting the development of industrial economy,and identifying potential emerging technologies is of great significance to the development of innovation strategy.Under the background of the rapid development of innovative technologies,it is of great significance for the government and enterprises to seize strategic opportunities in the early stage to find and identify emerging technologies with industry subversive potential.The development trend,evolution direction and influence of emerging technologies are uncertain and complex,which leads to the difficulty and low accuracy of their identification.The traditional identification methods of emerging technologies mainly include subjective judgment method,bibliometric method and citation network analysis method,which have the problems of subjective one-sidedness,time lag and insufficient research on technical knowledge and information system contained in text content.The emerging technology identification method based on text mining can make up for the shortcomings of traditional research methods.This method combines terms,keywords and other features with technical functions,effects and social utility,which is one of the increasingly popular research directions of emerging technology identification research.Aiming at the problem of insufficient attention to the future development and change trend of patent novelty in the current emerging technology identification process,this study analyzes the evolution process of patent novelty based on patent data by using the methods of text mining and bibliometrics,and proposes an emerging technology identification method based on the evolution analysis of patent novelty combined with the emerging technology identification index system.Firstly,patent text data sets of related technical fields are obtained from massive patent documents;Secondly,Python is used to preprocess the data,generate patent data set with patent publication date as time series,obtain the keywords of patent text again,and calculate TF-IDF value to construct word co-occurrence matrix;Thirdly,the text similarity between every two patents is calculated by cosine similarity,and the patent similarity matrix based on time series is constructed,then the patent novelty matrix is obtained according to the patent novelty calculation formula;Fourthly,taking the patent publication year as a time slice,the evolution matrix of patent novelty is constructed;Fifthly,according to the characteristics of emerging technologies,the identification index system of emerging technologies is constructed,and the index values of patents are compared,analyzed and judged,so as to complete the identification of emerging technologies;Finally,the identified emerging technologies are discriminated,analyzed and interpreted.In order to verify the effectiveness of the identification method of emerging technologies proposed in this study,the patent data in the field of bio-based fiber were obtained from Derwent Innovation Index for empirical research,and two emerging technologies,bio-based polyamide 5X and bio-based graphene composite fiber,were successfully identified.The research results show that the identified emerging technologies are consistent with the industry analysis,which shows that this method can effectively identify emerging technologies in patent documents,and provides new perspectives,new ideas and new methods for the identification research of emerging technologies.