Key Technologies And Application For Patent Big Data Analysis System

With the advent of the information age,the speed of technological development and innovation has reached an unprecedented peak,and such progress also brings unprecedented challenges to knowledge management in related technical field.Technological advances nowadays have greatly facilitated needs for organizational knowledge management,aiming at information concealed within multiple data sources,especially those hiding within patent databases.However,traditional methods for patent analysis has several drawbacks,including excessive presences of manual operation,low quality data,insufficient patent mining,poor visualization,and poor utilization of collective intelligence.A new method of patent analysis application is needed,which adapts to big data scenarios,utilizes collective intelligence,performs an accurate and efficient patent analysis result.In this paper,a patent big data collaborative analysis process is proposed based on the research in the related fields.The research focuses on three crucial technologies including plain patent analysis,patent complex network analysis and user-patent interactive analysis.Based on the theoretical framework put forward,a patent big data collaborative analysis system is developed.The main contents of this paper are summarized as follows.Chapter 1 studies the state of the art of patent activities,sorts out researches on the process of patent analysis and patent mining,complex network analysis and collective intelligence at home and abroad.Going into contemporary methods of patent analysis,the difficulties faced by the traditional patent analysis method were found,and the patent big data collaborative analysis process was proposed.Chapter 2 puts forward the plain patent literature analysis and mining technology plan,by defining the data form and content boundaries of plain patent literature,the data structure of plain patent literature is unriddled,a set of multi-profile quantitative statistical analysis methods are put forward for structural data of patent metadata,the analyzing process for unstructured patent text data is proposed including text data modeling,feature representation,feature filtering and the analysis of patent text keywords based on SAO structure.Chapter 3 proposes complex network modeling and analysis technology based on the use of concealed patent relationships.Extracting patent IPC co-occurrence relationship,citing relationship and keyword correlation based on SAO structure,patent networks are built,including citation network,patent IPC co-occurrence network and keyword network.After executing network index statistics thoroughly,critical path analysis and heterogeneous information network analysis are performed in order to extract information within patent complex network.In chapter 4,the user-patent interaction analysis technology based on collective intelligence is proposed.With the abstraction of interaction between users and patents,the user-patent interaction binary network is modeled,also the user association network.Based on these two networks,community detection and collaborative filtering of patent analysis users are completed,and the patent recommendation process is completed.Inspired by the link analysis algorithm,a bidirectional value evaluation method based on user-patent evaluation binary network is proposed.Chapter 5 develops the patent big data collaborative analysis system based on the theoretical research of this paper.It realizes the retrieval,view,evaluation and batch analysis of patent data for users,patent recommendation for users,global calculation and update of patent data,and the continuous extraction and analysis of relevant data of patent users and patents are implemented on systematical perspective.The innovation points of this paper include:(1)on the basis of combining patent literature mining,complex network analysis and collective intelligence theory,put forward the collaborative analysis method and process for patent big data;(2)put forward the patent analysis technology scheme based on collective intelligence--user-patent interaction analysis technology,with the help of collective intelligence,the optimization and intelligent expansion of patent data is completed;Based on the technical architecture of node.js+Spring+Hadoop,the patent big data collaborative analysis system is developed to realize the periodic and continuous update of global patent big data,the modeling and calculation of patent complex network and the support for users of patent collaborative analysis.