[2A3-E-5] Human interface, education aid: chance discoveries

[2A3-E-5-03] Entropy-based Knowledge Space Visualization for Data-driven Decision Support

〇Qi Wang¹, Teruaki Hayashi¹, Yukio Ohsawa¹ (1. The University of Tokyo)

Keywords:Market of Data, Information Entropy, Decision Support, Data Utilization, Data Collaboration

This paper describes an entropy-based visualization for extracting utilization value of data in data-driven decision support. There is an increasing demand for data utilization from different domains to create new businesses or solve existing problems. Data utilization is a human-interactive field thus reusing data utilization knowledge is an important task to define data relations in the field of solving multidisciplinary problems. That is to say, knowing the application fields of data is vital in promoting data utilization for problem-solving. Thus, this paper reuses the knowledge of data utilization to re-design data relations in the visualization of knowledge space to assist data utilization recognition and decision-making processes. The Innovators’ Marketplace on Data Jackets is employed to collect data utilization knowledge, and we use data co-occurrence entropy to reflect people’s recognition process on data utilization because co-occurrence between data in utilization knowledge is an informative prior supporting for decision-making. The experimental results show the proposed entropy-based visualization method suppresses baseline methods in the data-utilization proposal numbers and qualities, which verifies proposed method assists people’s recognition and decision-making processes in data utilization.