论文部分内容阅读
借助主成分分析,以PLoS ONE的Article-Level Metrics为数据源对物理学、化学、社会学、免疫学四学科的科学评价主要维度进行解析。分析表明一维空间的科学评价在覆盖50%信息的同时将损失其余的50%,3个维度才能以80%的精度描述论文的学术影响力,进而将科学评价的3个维度分别命名为引用维、共享维与利用维,从而说明以引用为基础的传统科学评价的片面性,并揭示科学评价的多维构成。
Using principal component analysis and PLoS ONE’s Article-Level Metrics as a data source, the major dimensions of scientific evaluation in physics, chemistry, sociology and immunology are analyzed. The analysis shows that the scientific evaluation of one-dimensional space will lose the remaining 50% while covering 50% of the information, and the three dimensions can describe the academic influence of the paper with 80% accuracy, and then the three dimensions of scientific evaluation are respectively referred to as the reference Dimension, dimension of sharing and dimension of use, so as to illustrate the one-sidedness of traditional scientific evaluation based on quotation and reveal the multi-dimensional constitution of scientific evaluation.