| In scientific research activities,it is a difficult problem that every researcher will encounter in his academic career whether to dig deep into a certain field or seek a breakthrough in a new direction.In view of this problem,the academic community has put forward two completely different views.One thinks that a certain field should be deepened to achieve greater achievements,while the other holds the opposite view that the way to seek a breakthrough is to change the research direction in time and constantly change and innovate.So should researchers shift the research topic? Is there any correlation between the degree of transfer of research topics and scientific research performance? There have been many studies on the factors that affect the performance of scientific research,such as scientific research pressure,scientific research environment and knowledge sharing,but there are few studies on whether the transfer of scientific research topics will affect it.Therefore,this paper focuses on the following three questions: 1)Has the research topic shifted in the career of researchers and how to measure the degree of shift? 2)what are the characteristics of the distribution of research topics in researchers’ careers? 3)Is there a relationship between research topic transfer and scientific research performance and what kind of quantitative relationship exists?In order to solve these three problems,this paper takes the researchers of computer,artificial intelligence and business and economy from 1975 to 2017 as the research objects,and probes into them from three aspects.1)Using the method of semantic similarity calculation to measure the transfer of research topics in two disciplines,and adopting the method of case analysis,eight researchers from two disciplines were selected to analyze the transfer of research topics with time in their careers by co-word clustering.2)Using co-word clustering to analyze the overall distribution of research topics of researchers in the two disciplines,and observe the changes of their research hotspots with time.3)Through correlation analysis and negative binomial regression model,the relationship between research topic transfer and scientific research performance is explored.The research results show that: according to the calculation results of the similarity of academic papers,if the similarity is less than 0.4,the research topic has shifted,then45.6% of scholars in computer and artificial intelligence disciplines have shifted their research topics,and 39.5% of scholars in business and economics disciplines have shifted their research topics.From the perspective of case analysis,it can be seen that six of the eight scholars selected by the two disciplines have shifted their research topics to varying degrees.From the results of cluster analysis of the overall co-words of scientific researchers,it can be concluded that the research topics of computer and artificial intelligence have changed greatly compared with those of business and economics,and new theories and methods are emerging one after another.The research hotspots of business and economics are relatively fixed,but the proportion of hotspots will change to varying degrees over time.The results of correlation analysis show that the Similarity of academic papers fails the significance test at the total cited frequency(TCC)when the research object is the researchers of computer and artificial intelligence.When the research object is researchers in business and economics,the Similarity of academic papers only passes the significance test in TSRO,and there is no correlation between research topic transfer and scientific research performance.The results of negative binomial regression show that the relationship between research topic transfer and scientific research performance shows disciplinary differences.When the research object is researchers in computer artificial intelligence,there is a correlation between research topic transfer and scientific research performance.When the research object is researchers in business and economics,there is no correlation between research topic transfer and scientific research performance. |