With immense influx of historical data, quantitative inferences on history based on machine learning is becoming more prevalent, attracting many researchers. In particular, understanding the dynamics of historical factions is important as they shared academic beliefs, political views and interests, in which the interactions between the factions portray general political, social, and economic structure of a certain era. In recent years, studying such dynamics through network-based methods on human networks, constructed from genealogy data, have shown promising results. In this paper, we enhance the identification of historical factions by exploiting multi-layered network of historical figures. To understand the mechanisms of historical factions, it is pivotal to comprehend the change in relation between important historical events. The proposed method consists of constructing a multi-layered network of historical figures and applying semi-supervised learning framework to identify historical factions. The proposed method was applied to the classification of factions in the political turmoil occurred during the 15th to 16th century Korea.
Many years of huge efforts from historians to create database for historical figures and events have led to easier access to a tremendous pile of records of the old days (Manabe, 1999, 2010; Lee, 2010, 2016; Lee, Lee, Kim, & Shin, 2018). The databases for historical data now provide convenient access to genealogy records, texts from literature, images of artifacts and so on. Following from the rapid growth in influx of historical data, many studies attempt to stretch from human-labored restricted inference to machineaided comprehensive approaches. Machine learning plays a key role in the latter approach, providing historians with data-driven inference. To give few examples, Malmi, Gionis, and Solin (2018) develops automated methods, using Naïve Bayes approaches, for inferring large-scale genealogical networks. In the domain of art history, deep convolution neural networks are often employed to classify style of fine-art collections (Bar, Levy, & Wolf, 2014; Cetinic, Lipic, & Grgic, 2018; Saleh & Elgammal, 2015). For textual data, Lansdall-Welfare et al. (2017) employs text mining techniques on 150 years of British newspapers from 1800–1950 to extract macroscopic trends in history and culture. In Rochat (2015), network analysis on character network from historical writing is employed. Furthermore, many visualization efforts (Liu, Dai, Wang, Zhou, & Qu, 2017; Lee, Campbell, & Chen, 2010) for historical data are carried out to aid in understanding the structure and contents of history.
From the vast domain of using machine learning for historical big data, studies on past faction politics may convey important inferences on political, social, and economic structure of a certain era. Throughout history, people who shared academic beliefs, political views and interests joined to form factions to pursue a particular aim or purpose (Belloni & Beller, 1976). Analysis on the prevalence of factions and their rivalries is crucial for understanding political decision patterns and power mechanisms of times long past. Furthermore, the political competition of factions may be viewed as competition for goods and services (Persico, Pueblita, & Silverman, 2011), which may yield important inferences in economic perspectives