Font Size: a A A

The Discovery And Analysis Of Scientific Success Reason Based On Data Mining

Posted on:2016-10-04Degree:MasterType:Thesis
Country:ChinaCandidate:Q XiaFull Text:PDF
GTID:2308330479995563Subject:Software engineering
Abstract/Summary:PDF Full Text Request
Papers are the most direct representations of the achievements of scientific researches and total cites of the papers reflect the impact of their researches. Studying on these frequently cited papers(Citation Classic) will provide valid conclusions on hotspots and directions, which can lead the development of researches. Being highly cited also means the success of a scientific. Therefore, analysis on the reasons for success of these scientific researches will be of great importance to carrying out the scientific researches.This paper is based on the authors’ self-commentaries in Citation Classic on how it leads to success. The study is carried out in following three aspects:(1) Dig the reason for scientific success. Establish the self-commentary document sets for Citation Classic authors. Get the information from the disciplinary fields, and extract the self-commentary contents and relevant information. By analyzing the topic words accounting for the reason for scientific success, we find out that the common themes we concerned are often reflected by low-frequency topic word instead of high-frequency ones. The existing methods of topic discovery are inclined to obtain high-frequency words, so they won’t work in extracting the low-frequency topic words. As a result of the properties of low independency and homogeneous distribution, in this paper we propose a new method of topic discovery in interdisciplinary fields to evaluate whether some words accord with common themes. With the extraction of low-frequency topic words, the method can not only exactly extract the topic words which manifest the reason for scientific success, but also classify them and provide the information about contribution of different reasons.(2) Analysis on reasons for the success of scientific researches. This paper focuses on statistical analysis about the potential relationship between the reason for successful digging from the content and the year of publication of the self-comments and original papers, the discipline fields, the geographical information. Then it show us some interesting speculation.(3) Establish the platform for digging, analyzing and visualization. In order to embody the value of the achievement of this paper and make itself more convenient to understand and use, a Web platform is established and used to display the achievements of this paper via visualization.
Keywords/Search Tags:Discovery of topic words, Low-frequency topic words, Citation Classic, Reason for success
PDF Full Text Request
Related items