fdata-02-00038_From Big Scholarly Data to Solution-Oriented Knowledge Repository.xml

The volume of scientific articles grow rapidly, producing a scientific basis for understanding and identifying the research problems and the state-of-the-art solutions. Despite the considerable significance of the problem-solving information, existing scholarly recommending systems lack the ability to retrieve this information from the scientific articles for generating knowledge repositories and providing problem-solving recommendations. To address this issue, this paper proposes a novel framework to build solution-oriented knowledge repositories and provide recommendations to solve given research problems. The framework consists of three modules: a semantics based information extraction module mining research problems and solutions from massive academic papers; a knowledge assessment module based on the heterogeneous bibliometric graph and a ranking algorithm; and a knowledge repository generation module to produce solution-oriented maps with recommendations. Based on the framework, a prototype scholarly solution support system is implemented. A case study is carried out in the research field of intrusion detection, and the results demonstrate the effectiveness and efficiency of the proposed method.

学术论文的体量呈快速增长之势，为理解与识别研究问题、梳理前沿解决方案提供了科学依据。尽管问题求解相关信息具备重要价值，但现有学术推荐系统尚无法从学术论文中检索此类信息，以构建知识库并提供问题求解相关推荐。为解决这一问题，本文提出一种新颖的框架，用于构建面向解决方案的知识库，并针对给定研究问题提供求解推荐。该框架包含三个模块：一是基于语义的信息抽取模块，可从海量学术文献中挖掘研究问题与解决方案；二是基于异构文献计量图与排序算法的知识评估模块；三是用于生成带推荐功能的面向解决方案图谱的知识库生成模块。基于该框架，我们实现了一款原型学术求解支持系统。随后在入侵检测研究领域开展了案例研究，实验结果验证了所提方法的有效性与高效性。