学术论文的定性评价定量化研究

Quantitative Research on Qualitative Evaluation of Academic Papers

[目的/意义]随着教育部科技部提出的"审慎选用量化指标",如何有效地使用定量方法进行学术论文评价成为科学计量学领域的一个重要研究方向。[方法/过程]文章基于融合论文信息和开放获取数据进行定性评价定量化的研究思路,提出了融合定性评价的论文质量评价模型。以论文标题、摘要、Twitter评论、同行评议作为模型输入,以论文审稿平均分作为实际分数,采用情感分析模型挖掘开放获取文本情感极性,根据评论文本情感极性对论文进行评价。[结果/结论]通过对比情感模型的预测分数与实际分数、预测排名与实际排名,证明了该定性评价定量化模型的有效性。完成训练的论文质量评价模型,可以直接使用于大量没有审稿分数的定性评价文本进行定量化研究,由此完成论文质量评价任务。

[Purpose/significance]With the“prudent selection of quantitative indicators”proposed by the Ministry of educa-tion and science and technology,It has become an important research direction to effectively use quantitative methods to evaluate academic papers in the field of scientometrics.[Method/process]Based on the idea of integrating paper information and open ac-cess data for quantifying the qualitative evaluation,we propose a quality evaluation model for paper based on qualitative evaluation model,taking the title,abstract,twitter review and peer review and as the model input,and the average score of peer review as the real score,the sentiment polarity of open access text is mined by using sentiment analysis model,and different predictive scores are given according to the sentiment polarity of the review text.[Result/conclusion]By comparing the predicted score with the ac-tual score,the predicted ranking with the actual ranking,the effectiveness of the qualitative evaluation and quantitative model is proved.The trained quality evaluation model for paper can be directly applied to a large number of qualitative evaluation texts with-out peer review scores for quantitative research,so as to complete the task of quality evaluation for paper.