Caprice 3-4
Text Classification Framework for Assisting Systematic Literature Review (304894)
*Xiaoxiao Li, Penn State UniversityLe Bao, Penn State University
Keywords: automate systematic review, text mining, classification
Systematic review provides a complete, exhaustive summary of literature relevant to a research question. Meanwhile, it also suffers from high cost in production due to the time-demanding nature. In this presentation we propose a framework using techniques in text mining and information retrieval to assist the process. We first introduce a feature-clustering classification model, then improve its performance by adding key features obtained from variable screening. By analyzing the precision and recall, our procedure is able to relieve the burden of readers and maintain a satisfactory recall rate at the same time.