Theoretical Research And Software Implementation About The 3D Evolution Map Of Scientific & Research

The development of weapon and equipment technology can not be separated from the science and technology.With the development of information technology,the era of big data has come to be.Big data has its wide influence and has penetrated into all aspects of social life.Following the experimental science,theoretical and computational science,there comes the fourth research paradigm,namely,"data intensive science",which has become a new model in the era of big data.Big data has brought a great opportunity for scientific research,which brings about the effect of "vertigo" and "data redundancy".In the face of the great challenge to the scientific research work,it has become an urgent need of scientific research workers and scientific research management to grasp and predict the development trend of scientific research.On the basis of the research of the traditional scientific knowledge map,this paper proposes a three-dimensional scientific research situation evolution model based on the distributed representation of network and word.This method has great reference value for mastering the development trend of science and technology,mining technology frontier,grasping the technological opportunity,and promoting the development of equipment technology.And the work of this paper mainly includes the following aspects:Based on the keywords cocition network,the network embedding algorithm is used to generate the structural vector which reflects the characteristics of the network.Then original keyword node are mapped into a real number vector of low dimensional space by using complex network embedding method,so the problem of data sparsity is solved.The obtained keywords vectors can be used to evaluate the structural similarity between nodes.Keywords semantic vector generation algorithm based on Labeled LDA and word2 vec.This method can be used to calculate the mapping relationship between keywords and abstract words in Labeled LDA model.Then,the mapping relationship can be reflected by the semantic vector,which can reflect the semantic similarity between keywords.A new method for calculating the important degree of the keywords considering the author’s weight information is presented.In this method,the influence of the neighbor nodes and neighbors in the network is considered,and the important degree of nodes is also integrated into the calculation process.This paper constructs a framework of three-dimensional scientific research situation evolution map and proposes four kinds of generating algorithms for threedimensional scientific research situation evolution map.From the point of view of node type,node vector and time factor,the three-dimensional scientific research situation map is classified,and the combination frame of three-dimensional scientific research map is constructed.A software platform named TVIZ based on the above theoretical is implemented.The platform based on the theory of this paper,it can do a comprehensive analysis for the records form SCI,CSSCI,EI,CNKI and patent literatures,and also supports network analysis,technical maturity analysis,natural language analysis and visualization display and other functions.In this paper,the research of three-dimensional scientific research situation evolution map,which is integrated into the deep learning theory,complex networks theory and natural language processing technology,has deepened the research on the scientific measurement and scientific knowledge map.