| Technical evolution analysis is to analyze the evolution of technology generation,innovation,and development through bibliometrics and text mining.It can provide an important reference for the rational allocation of scientific and technological resources and the formulation of national and corporate strategic decisions,and thus attracts widespread attention from government,enterprises and academia.Patent is an important carrier of independent innovation achievements,and has become one of the most reliable data sources for technological evolution analysis.How to deeply reveal patent technology information is the focus of the current technology evolution analysis.However,the existing analysis is limited to technical topics,and insufficient attention has been paid to technology splitting and reorganization,which makes it difficult for researchers to reveal the features of technology evolution at a more specific and detailed level.In order to settle above issues,this study takes the constituent elements of patents as data sources,extracts multiple types of technical elements contained in patents as research objects,explores the evolution of patent technology based on the technical elements and their association in the view of multidimensional,semantic,fine-grained and more.In particular,the main work of this study is as follows:(1)Starting with the constituent elements of the patent,we collect the text information of the patent,and construct a patent text representation model based on the technical elements and their associated relationships.Through the construction of typical patent analysis,we provide a recognition framework for technical element by technical element feature analysis and feature recognition mark,etc.This framework can be used for SAO structure extraction,SAO structure classification based on verb lexicon,technical keyword feature labelling,and the construction of rule base,which help to identify and extract the technical element entities and their relationships.(2)Based on the technical element entities and their associations,single-element and multi-element technology element networks are constructed according to different analysis purposes.From statistical characteristics,social network index analysis,etc.,we analyze the links between time series networks in different time slices and other characteristics,and propose a framework for evolutionary state analysis based on technical element networks.(3)We propose an association relationship mining method based on attribute information of technical element,and provide a formula for calculating the intensity of association relationship based on attribute similarity for technical element analysis,and further give an association relationship description model for different attribute information dimensions.Combining time attributes,we propose a method for analyzing the evolution path of technology based on technical elements in specific technical fields.(4)Taking the field of slow and controlled fertilizer release as an empirical field,we identify the key words of technical elements in different time slices and their associations.Based on the time-series single-element and multi-element network,we analyze the evolution of the technology elements and their associations.Based on product technology elements,the technology evolution path of products based on material attributes was identified.The feasibility and effectiveness of the proposed method were verified through the comparative analysis of expert consultation and the key results of evolutionary analysis.The innovation of this paper is to propose a technology evolution analysis model based on the feature mining of patent elements.Based on the analysis of typical patents,we provide a recognition framework for technical element by technical element feature analysis and feature recognition mark.Based on the technology elements and their relationship,we propose a method of technology elements network construction and evolution analysis,which can reveal the technology evolution path in a fine-grained way. |