一种文本挖掘和文献计量的科技论文评估方法
|
摘要
【目的/意义】随着我国科技水平的不断提升,越来越多的科研人员致力于研究如何客观有效地评估科技论文的质量。【方法/过程】本文提出了基于文本挖掘和文献计量的科技论文评估方法,该方法首先利用论文的引用频率计算得到训练论文的质量,其次分别使用LSI和LDA方法来计算新论文和训练论文之间的相似度,设计论文质量评估算法,根据训练论文的质量和论文间的相似度加权计算得到新论文的质量评价值。【结果/结论】在涉及多主题和单一主题的数据集上的实验结果表明,本文提出的方法能够适用于包含多个主题的论文集,其计算出来的论文质量值真实有效。
【Purpose/significance】With the development of the technological level of our country, more and more researchers are concerned about how to evaluate scientific papers objectively and effectively.【Method/process】To combine the bibliometrics with content analysis for the quality evaluation, we propose a scientific paper evaluation method based on text mining and reference frequency. At first, the quality of training papers is calculated by the use of the papers' reference frequency. And then it measures the similarity between the training papers and the new paper based on LSI model and LDA model respectively. Finally, according to the quality of training papers and the similarity between papers, this method calculates the weighted quality of new papers.【Result/conclusion】The experiments are implemented on two datasets involving multiple topics and single topic respectively, and the results show that our method is effective and it can be applied to the datasets involving a variety of topics.
【Purpose/significance】With the development of the technological level of our country, more and more researchers are concerned about how to evaluate scientific papers objectively and effectively.【Method/process】To combine the bibliometrics with content analysis for the quality evaluation, we propose a scientific paper evaluation method based on text mining and reference frequency. At first, the quality of training papers is calculated by the use of the papers' reference frequency. And then it measures the similarity between the training papers and the new paper based on LSI model and LDA model respectively. Finally, according to the quality of training papers and the similarity between papers, this method calculates the weighted quality of new papers.【Result/conclusion】The experiments are implemented on two datasets involving multiple topics and single topic respectively, and the results show that our method is effective and it can be applied to the datasets involving a variety of topics.
引文
1邱均平.文献计量内容分析法[M].北京:国家图书馆出版社, 2008.
2 王曰芬.文献计量法与内容分析法的综合研究[D].南京:南京理工大学, 2007.
3 邱均平,邹菲.关于内容分析法的研究[J].中国图书馆学报, 2004, 30(2):12-17.
4 张玉华,潘云涛,马峥.科技论文评估方法研究[J].编辑学报, 2004, 16(4):243-244.
5 龙莎,葛新权.科技论文学术水平评估[J].科技与管理,2007, 9(1):133-135.
6 管进.核心期刊的综合评价法[J].情报杂志, 2003, 22(2):63-64.
7 邓雪,李家铭,曾浩健,等.层次分析法权重计算方法分析及其应用研究[J].数学的实践与认识, 2012, 42(7):93-100.
8 唐有文.模糊层次分析法[J].青海师范大学学报(自然科学版), 2002,(3):19-23.
9 阮剑,杨海霞,黄瑱.基于主题模型的大学学报文献挖掘研究——以计算机科学领域为例[J].情报工程,2017,3(4):39-47.
10 Deerwester B S, Dumais S F, Landauer G, et al.Indexing by latent semaintic analysis[C]//JASIS,1990.
11 李琳娜,江雪琴.推荐系统中的隐语义模型研究[J].情报工程,2016,2(4):30-39.
12 Zelikovitz S, Hirsh H.Using LSI for text classification in the presence of background text[C]//Tenth International Conference on Information and Knowledge Management.ACM, 2001:113-118.
13 黎楠,杜永萍,何明.基于主题发现的专利发明人推荐方法[J].情报工程,2015,1(3):90-97.
14 Blei D M, Ng A Y, Jordan M I.Latent dirichlet allocation[J].Journal of Machine Learning Research, 2003,(3):993-1022.
2 王曰芬.文献计量法与内容分析法的综合研究[D].南京:南京理工大学, 2007.
3 邱均平,邹菲.关于内容分析法的研究[J].中国图书馆学报, 2004, 30(2):12-17.
4 张玉华,潘云涛,马峥.科技论文评估方法研究[J].编辑学报, 2004, 16(4):243-244.
5 龙莎,葛新权.科技论文学术水平评估[J].科技与管理,2007, 9(1):133-135.
6 管进.核心期刊的综合评价法[J].情报杂志, 2003, 22(2):63-64.
7 邓雪,李家铭,曾浩健,等.层次分析法权重计算方法分析及其应用研究[J].数学的实践与认识, 2012, 42(7):93-100.
8 唐有文.模糊层次分析法[J].青海师范大学学报(自然科学版), 2002,(3):19-23.
9 阮剑,杨海霞,黄瑱.基于主题模型的大学学报文献挖掘研究——以计算机科学领域为例[J].情报工程,2017,3(4):39-47.
10 Deerwester B S, Dumais S F, Landauer G, et al.Indexing by latent semaintic analysis[C]//JASIS,1990.
11 李琳娜,江雪琴.推荐系统中的隐语义模型研究[J].情报工程,2016,2(4):30-39.
12 Zelikovitz S, Hirsh H.Using LSI for text classification in the presence of background text[C]//Tenth International Conference on Information and Knowledge Management.ACM, 2001:113-118.
13 黎楠,杜永萍,何明.基于主题发现的专利发明人推荐方法[J].情报工程,2015,1(3):90-97.
14 Blei D M, Ng A Y, Jordan M I.Latent dirichlet allocation[J].Journal of Machine Learning Research, 2003,(3):993-1022.