Research On Knowledge Fusion For JMRH Industrial Community Based On Machine Learning

With the in-depth development of China’s socialist market economy reform,JMRH towards the direction of knowledge-driven,innovation-driven,opening-up and systematization.JMRH industrial community,as a high-level JMRH industrial development model,aims to strengthen the integration of knowledge,information,technology and other resources of JMRH industry.In the process of promoting the formation and development of JMRH industrial community,knowledge management is an important driving force.However,the knowledge of JMRH industrial community mainly comes from JMRH industry-related information,which are growing exponentially and scattering massively,hindering the knowledge acquirement of JMRH industrial community seriously.Therefore,it is urgent to adopt some certain ways and methods to process JMRH industrial information for excavating and acquiring the knowledge contained therein.And an orderly and efficient JMRH industrial community knowledge system is built through knowledge fusion,realizing knowledge management in JMRH industry and boosting the formation and development of JMRH industrial community.Knowledge fusion can extract relevant knowledge from information,and form a complete domain knowledge system through knowledge transformation,integration and merger.For solving the problems of knowledge management and service in the field of JMRH information sharing mechanism and innovation system construction,new ideas,new methods and new paths are proposed to accelerate the development of JMRH from the knowledge level,which is of great significance for hastening the formation of the further development pattern of JMRH.The main research work and contributions are as follows:(1)The framework of JMRH industrial community knowledge fusion is constructed.Based on the defined concept,connotation and characteristics of JMRH industrial community,the objectives and principles of JMRH industrial community knowledge fusion are set,the needs of JMRH industrial community knowledge fusion are analyzed,the procedure and structure of JMRH industrial community knowledge fusion are designed,and the logical framework of JMRH industrial community knowledge fusion is constructed from knowledge fusion elements and the relationship between each element,showing the overall stages of JMRH industrial community and the important activities of each stage.Based on the logical framework,the CSFA analysis model of JMRH industrial community knowledge fusion,including content dimension(C),structure dimension(S),function dimension(F)and application dimension(A)is established to describe the specific content of JMRH industrial community knowledge fusion in different dimensions.(2)The topic model algorithm is proposed to research on the acquisition and representation of knowledge in the content dimension of JMRH industrial community knowledge fusion.On the basis of big data background and machine learning advantages,the LDA model,LDA-SVM algorithm and improved Page Rank algorithm are utilized to collect and process the subject information from JMRH industrial community knowledge source;the ontology technology is applied to construct JMRH industrial community ontology from the processed information so as to obtain orderly JMRH industrial community knowledge resources;For describing and representing the knowledge content uniformly,the "Knowledge element-Ontology" technology is used to reveal the inner relationship between knowledge elements and ontology elements,the ontology semantic association method is adopted to establish the semantic link of the knowledge units,creating the JMRH industrial community knowledge content semantic network so that the hidden,inherent and orderly relationships between the semantic concepts of JMRH industrial community knowledge are presented in a structured knowledge network,which solves the problems of knowledge acquisition and representation in the JMRH industrial community knowledge fusion.(3)The domain concept taxonomy construction algorithm is proposed to study on the classification and re-organization of knowledge in the structure dimension of JMRH industrial community knowledge fusion.Based on semantic network of JMRH industrial community knowledge content,the mixed strategy domain concepts method of the domain concepts taxonomy construction algorithm is adopted to obtain JMRH industrial community domain concepts and extract the relationship between the domain concepts so as to classify the knowledge effectively;for obtaining a taxonomy with more reasonable concepts and stronger abilities of knowledge expression,BRT-Gauss algorithm is utilized to classify and organize the JMRH industrial community domain knowledge ontology,and the domain knowledge concept mapping model is constructed to merge the JMRH industrial community domain knowledge ontology and form the JMRH industrial community knowledge classification structure system,which eliminates the redundant knowledge in the JMRH industrial community domain knowledge and solves the problem of knowledge classification and re-organization in the JMRH industrial community knowledge fusion.(4)The association mining algorithm is proposed to investigate the knowledge association in the function dimension of JMRH industrial community knowledge fusion.By means of depicting the association relationship,association construction and association type of JMRH industrial community domain knowledge,the correlations between JMRH industrial community domain knowledge points and points as well as between JMRH industrial community domain knowledge point and resources are analyzed,and the correlation are measured from knowledge association content and knowledge association collaboration.On this basis,the association mining of the attributes and functions of the JMRH industrial community domain knowledge is performed by the MLFAR algorithm,and the JMRH industrial community knowledge association function group is established.The related but scattered knowledge is presented in a definite rules or patterns,which solves the problem of knowledge association in the JMRH industrial community knowledge fusion.(5)The clustering algorithm is proposed to research on knowledge clustering in the application dimension of JMRH industrial community knowledge fusion.For aggregating JMRH industrial community domain knowledge according to subject feature so as to establish a complete JMRH industrial community domain knowledge system and make the fusion path of JMRH industrial community domain knowledge clearer and more precise,CC algorithm of clustering algorithm is utilized for clustering analysis of JMRH industrial community domain knowledge,Louvain algorithm is adopted to divide knowledge application cluster community,the K-plex and Clique are used to extract knowledge application cluster community with different closeness degrees so as to represent the change and development of JMRH industrial community domain knowledge,which solves the clustering problem of the knowledge with similar categories in JMRH industrial community knowledge fusion.(6)A case analysis about China "General aviation industry" of JMRH industrial community knowledge fusion is conducted by the proposed CSFA analysis model and machine learning.Experiments of this example verify that CSFA analysis model can comprehensively fuse the knowledge of "General aviation industry" from four dimensions,and according to the passage of time,the development directions of multiple main applications of "General aviation industry" domain knowledge can be displayed by knowledge fusion,making the advancement path of "General aviation industry" more apparent,providing guidance of future development for "General aviation industry".Meanwhile,JMRH industrial community knowledge fusion methods based on above-mentioned machine learning algorithms have a good performance on accuracy,rationality,feasibility and operability in practical application,and play a prominent role in the deepened exploitation and utilization of knowledge resources in JMRH,which provides a new realization path for in-depth development of JMRH from the knowledge level.