A B S T R A C T Knowledge base construction (KBC) aims to populate knowledge bases with high-quality information from unstructured data but how to effectively conduct KBC from scientific documents with limited preknowledge is still elusive. This paper proposes a KBC framework by applying computational intelligent techniques through the integration of intelligent bibliometrics-e.g., co-occurrence analysis is used for profiling research topics/domains and identifying key players, and recommending

more »

... potential collaborators based on the incorporation of a link prediction approach; an approach of scientific evolutionary pathways is exploited to trace the evolution of research topics; and a search engine incorporating with fuzzy logics, word embedding, and genetic algorithm is developed for knowledge searching and ranking. Aiming to examine and demonstrate the reliability of the proposed framework, a case of gene-related cardiovascular diseases is selected, and a knowledge base is constructed, with the validation of domain experts. data such as scalability, uncertainty, and robustness, and domains for which interests on intelligent bibliometrics are raised, emphasizes the refinement required of traditional bibliometric approaches by incorporating intelligent techniques, e.g., topic models, network analytics, neural networks, and other machine learning approaches. Such endeavors include dynamic topic detection and tracking [8], word embedding-incorporated topic extraction [9], streaming data analytics for identifying complicated semantic relationships among research topics over time [10] , etc. However, gaps between analytic results of established bibliometric approaches and KBC still exist such as how to design a KBC framework to systematically integrate bibliometric models and effectively manage knowledge, and how to implement the proposed framework for decision support in realworld cases.