分类:Hotness Tracing

Hotness tracing means that when facing options, one chooses the one with the largest hotness. Such hotness can be a fashionable cloth which has been chosen by many, or a research topic which has already many publications, or a research papers which has been cited by many other papers.

We have investigated the second^[1][2], when choosing research topics to work on and publish papers in, how often a hot topic, which is measured by the accumulated (starting from certain point of time) number of papers on the topic, is chosen. Roughly speaking [math]\displaystyle{ \frac{p\left(k\right)}{n\left(k\right)} }[/math], where [math]\displaystyle{ p\left(k\right) }[/math] is the empirical distribution function the size of the topics which new papers are published on during a short period of time and [math]\displaystyle{ n\left(k\right) }[/math] is the number of topics with size [math]\displaystyle{ k }[/math]. We found that overall, all scientists trace hot topics, no matter which discipline they are in or which country they are from. However, Chinese scholars has the largest degree of hotness tracing.

There is a tiny extension can be done to this work: looking at time evolution of size of topics in various countries. There might be some countries where size of certain topics started to grow much earlier/later than other countries. To identify what are such topics for given countries might be an interesting problem.

Furthermore, we are thinking to generalize the study of hotness tracing from publications to citations. That is to ask roughly, whether or not when a paper is heavily cited already the chance for it to get new citations is higher. This can be done simply via looking at [math]\displaystyle{ \frac{p\left(k\right)}{n\left(k\right)} }[/math], where [math]\displaystyle{ k }[/math] now becomes the number of received citations.

However, there are several subtleties when combining countries with hotness tracing in citations. Let us define a few notations. We denote all the papers from a country C as [math]\displaystyle{ P_{C} }[/math] and group papers that are cited by [math]\displaystyle{ P_{C} }[/math] as [math]\displaystyle{ P_{Cc} }[/math]. One can also separate [math]\displaystyle{ P_{Cc} }[/math] into papers from various countries, ie. [math]\displaystyle{ P^{D}_{Cc} }[/math] and [math]\displaystyle{ P_{Cc}=\sum_{D}P^{D}_{Cc} }[/math].

Then for the first, one may ask how citations from those papers trace hotness. That is when [math]\displaystyle{ P_{C} }[/math] are published (in the given targeted short period of time) how they choose their references, which will be in [math]\displaystyle{ P_{Cc} }[/math]. Furthermore, when looking into this issue, and study distribution function and hotness tracing in the citation from [math]\displaystyle{ P_{C} }[/math] to [math]\displaystyle{ P_{Cc} }[/math] or [math]\displaystyle{ P^{D}_{Cc} }[/math]. By the way, the distribution can be quantifies for example by the Gini coefficient of the distribution.

For the second, 这个引文分为被引的热点追踪和施引的热点追踪。施引的热点追踪就是考虑中国的学者引用所有文章（以及分开来看的中国和其他国家的文章）的时候，是不是比较集中在少数被引次数比较高的论文上（这是分布函数的研究），以及是不是已经高被引的文章得到新的引用的几率比较高。被引的热点追踪，是指对于中国的所有文章来说，被中国或者其他国家引用的时候，追热点程度有没有区别。当然，为了做比较，中间这个“中国”需要替换成任意一个主要国家，都算一遍。

具体计算上和文章发表追热点领域的计算是一样的。还可以做这个引文热点追踪现象的文章的其他特征的对比（例如文章的年龄、作者的h指数等）。另外，在时间序列分析上，还可以讨论，是否存在主题被引用和引用次数时间序列和国家领域贡献量的相关性（超前、滞后、等时）。

和这个工作相联系的，我们还可以讨论基金、收入、论文数量、被引次数等分布不均匀性的研究，也就是大鱼吃小鱼现象，以及不公平性对科学技术进步的影响。

整个工作合起来，并且结合国家-学科的投入产出分析以及基金机构影响力度量（尤其是中间中国发表的论文的论文数量或者被引次数占被引次数或者其他指标top百分之几的文章的比例，中国施引的论文的论文数量或者被引次数占全球top百分之几的比例。中国发表的论文表示中国的科研产出。中国施引表示中国科学的思想源头和眼界），我们还有一个隐含的主题：中国科学研究现状和国际地位怎样。我们已经发现，中国科学家的在文章数量上的追热点非常高，而中国在小领域的相对贡献率非常小（也就是中国在小领域里面没有做出来相配的贡献量）。如果我们进一步检验中国科研经费的分配情况、发表论文的施引和被引情况，我们就可能可以更深入地了解中国科学研究的现状和可能的原因。

References