Analysis Of Literature Evaluation Index Based On Citation Network Dataset

Reading scientific literature is the basis of scientific research.It is a common problem for every researcher to select the most worthy articles from the vast literature.The existing literature evaluation system is based on the number of citations per document.Intuitively,the more people cite the literature,the greater the probability is of high value.However,scientists often do not just want to read the literature that everyone has read and cited,but hope to find the most original and innovative literature and learn from it.At this time,the existing literature evaluation system can not provide accurate reference information fo r researchers.Main work of this paper is to put forward a kind of both based on the literature cited situation,and based on its own the evaluation methods of information,references to evaluate the literature contribution of originality.This paper verifies the effectiveness of the evaluation method through theoretical analysis and real world data set calculation simulation.Further,this paper also uses the evaluation method,the existing journals to emphasis the contribution of original paper analyzed and ranked,can provide reference for scientific research workers contribute choose target.The main contributions of this paper are as follows:1.This paper constructs a computer science citation network data set.The citation network data set contains more than 500 ten thousand references and 200 nodes,in which each node has 61 attribute information.It is currently known that that data set for open source contain a maximum number of attribute,the most recent computer science citation network datasets that are the most recent in connection with the paper data.2.This paper proposes a method of measurement of originality contribution.This method measures the originality contribution of the paper from two aspects: the influence of the thesis and the strength and weakness of the work foundation.The experimental results show that this method can effectively make up the deficiency of traditional evaluation method in the original contribution of the paper,and can be used to screen original papers.At the same time,it is found that the method has the characteristics of dynamic and research independence.3.This paper proposes a periodical evaluation method based on original contribution of paper.At present,periodical evaluation index is the influence of evaluation periodical.In this paper,a periodical evaluation method is proposed to measure the acceptance of original papers.The experimental results show that the index and influence index are different in the citation network data sets.