COMPSTAT 2022: Start Registration
View Submission - COMPSTAT2022
A0460
Title: Development of visualizing system based on research networked data for open science age Authors:  Hiroka Hamada - The Institute of Statistical Mathematics (Japan) [presenting]
Keisuke Honda - The Institute of Statistical Mathematics (Japan)
Abstract: To see how papers are influencing different fields, an index has been developed that defines the degree of heterogeneity based on only citation relations and scores the scattering of citations. In general, clustering of network structures, like stochastic block models, requires computing the entire data. Non-negative matrix factorization is applied to estimate the global citation structure of the bibliography from a small subset by sampling. This allows the system to be offered as a license-free product using various open science services. We will introduce the system is provided with a 3D plot which is based on kernel PCA to easy understand the multidimensional assignments of the scores of the diversity indexes. Furthermore, a use case for the administrator of an institution to analyze research resources by linking this system with funding information is discussed in our demo.