Generalized difference method for generating integrated hypotheses in social big data
Hiroshi Ishikawa, Daiju Kato, M. Endo, Masaharu Hirota · 2018
Recently there is strong demand for analytic methodology as to generation of integrated hypotheses for applications involving different sources of social big data. In this paper, first, we introduce an abstract data model for integrating data management and data mining by using mathematical concepts of families, collections of sets to facilitate reproducibility and accountability required for social big data applications. Next, we propose generalized difference methods as a methodology for integrated analysis based on different sources of data. Finally, we validate our proposal by applying it to three use cases by using our data model as their description.