Research On Detecting Of Emerging Trends With Article Downloads

The recognition of the hot topics and emerging trends is an important field of scientometrics. In this paper, we propose a new method to find the emerging trend quickly and timely. The thought is:if the scientists are researching a particular field, he will retrieval and download the relational papers in the scientific literature database. Therefore, we can find the hot topics and emerging trends by studying the download numbers of papers. Based on computational neuroscience as an example, we analyze the hot topics and emerging trend.First of all, we collect the meta data, content data and downloads of all papers in the four journals through the springer.realtime.com platform. The four journals are Neural Computing & Applications, Journal of Computational Neuroscience, Neural Processing Letters and Neuroinformatics, a total of 4476 articles. Everyday downloads of all literatures were recorded from February 2014 to January 2015. We deal with the abstracts by topic model which is based on the traditional topic model, and add the article downloads as weights.In the chapter 4. the monthly hot topics are found and analyzed. We divide these hot topics into 4 categories. They are the biological prototype research, the algorithm of neural network and technology, the simulation model as brain, and the applied research. We found that the research in the biological prototypes is mainly concentrated in Neurons Form and Operation Mechanism, as well as the Visualization, Simulation of brain neurons. In the field of algorithm and technology, scientists focus on Machine Learning. There is no completely dominant algorithm, that is, hot topics in algorithms are lack of continuity.In the chapter 5, we analyze the dynamic development of every class of hot topics. With the time changing, the downloads of Machine Learning topic and Neurons Form and Mechanism topic are growing, Neural Network named on its algorithm and its attribute topics etc. are developing stably. In the end, we analyze the evolution trend of single topics and find Neural Network topic is stable, SPM is extending, RBF and Color Printing are merging.