Hot Topic Extraction And Prediction From Management Science Publications Based On Text Mining And Latent Dirichlet Allocation

Management Science is a broad and interdisciplinary field.Conduct a research in that field may be very challenging when one's have no idea of emerging topics.This research' first aim is to lighten the burden of scholars in choosing their research direction path by avoiding multiple preliminary readings just to identify current issues in management science field.By analyzing publications in management science field,this study will help to identify hot topics and draw a map of core area of interests in management science.Once,scholars start with a research activity,it is really important for their research to be meaningful and significant in the scientific sphere.Indeed,a prestigious research may be published in distinguished journals and may guarantee hiring,promotion,or even scholarships.Therefore,scholars need to ensure that their research topic is hot enough to get the attention of the scientific college and the reviewers.The second aim of the present research is to meet the need of scholars by offering a hotness predicting tool based on the identified trends to better orientate scholars research activity.Thereby,scholars may get a clear focus for research driving in management science area according to their personal interest and academic future.This study presents an analysis performed on 6,749 publications metadata(title+ abstract+ keywords)from 12 journals in management science.Based on a topic model,Latent Dirichlet Allocation,the investigation conducted has revealed a set of twelve topics in management science domain with three hot topics related to financial risk,transportation planning & traffic management,digital marketing.The results provided by the Latent Dirichlet Allocation model also helped to discover the dynamics of the 12 journals in regard to their content(uniqueness;similarity).A predicting tool based on the LDA model was also developed to assign topic;probable publishing journal as well as indicate the hotness on new documents related to management science and not yet published.The results of the present research could benefit scholars,researchers and even lecturers evolving in the field of management science,by helping them to better direct their research in term of topic choice,and may be journal search for paper submission.The methodology use in this study could be applied in other disciplines and may lead to the development of an overall collaborative academic tool to lighten the burden of scholars in regard to research activities.