Jong Hee Park

  • studies political methodology and international political economy.
  • currently employed as Associate Professor in the Dept. of Political Science and International Relations at Seoul National University.
  • received Ph.D. from Washington University in St. Louis.
  • maintains MCMCpack and Bayesian taskview in CRAN.
  • recently working on bilateral treaty networks, text analysis of North Korean document, multilayer network data, official development assistance, and dynamic estimation of media slants.

Detecting Structural Changes in Network Data: An Application to Changes in Military Alliance Networks, 1816-2012

In an empirical endeavor of network change-point analysis, the substantive importance lies in (1) correct identification of break numbers and (2) the estimation of changes in networkgenerating parameters. However, most existing network change-point detection methods model changes in descriptive statistics and often fail to provide a principled criterion to identify break numbers. To overcome these limitations, we develop a degree-corrected hidden Markov multilinear tensor model (HMTM) that combines the multilinear tensor regression model (Hoff, 2011) with a hidden Markov model using Bayesian inference. Then, we discuss two fully Bayesian methods to identify break numbers: the approximate log marginal likelihood (Chib, 1995) and the Watanabe-Akaike Information Criterion (WAIC) (Watanabe, 2010). Our simulation results demonstrate that the proposed method correctly detects the number, location, and type of changes in subgroup structures that are often the most important units for understanding the organizations and dynamics of networks. By applying the proposed method to military alliance networks from 1816 to 2012, we identify structural changes in the coalition structure of military alliance networks.

Learn More

Text Analysis of Korean Document

Understanding the logic, stretegy, and intention of North Korean government's rhetoric is such an important issue both for academics and for policymakers. Text analysis is one effective way to this goal.

Learn More (Korean)

Changepoint Analysis

Big and abrupt changes in history make important junctures in social history. How do we address these historical changes in our study of history? Our empirical methods are ill-suited to address dramatic changes in social history.

Learn More

Multilayer Network Analysis

We encounter networks almost everywhere in social science data. Most of them are not a single-shot observation, but have multiple dimensions such as time, topic, and group. How do we analyze these multilayer structures of network?

Learn More

세월호 참사 언론보도의 정치적 경도

본 논문은 세월호 참사와 관련된 1년 동안의 언론보도를 이용하여 여론의 흐름과 언론매체의 정치적 경도(partisan slant)를 측정하는 것을 목적으로 한다. 본 논문은 자동화된 텍스트 분석기법과 베이지안 추정을 결합하여 28개 언론매체(19개 신문사와 9개의 방송사)의 보도내용을 분석하였다. 이를 통해 세월호 참사 1년 동안 언론보도를 통해 드러난 여론의 흐름이 급격한 변화를 보였으며 참사 초기의 정부비판적 여론이 대통령의 대국민 담화와 두 번의 선거(6.4 지방선거와 7.30 재보궐선거)를 거치면서 급격하게 약화되었음을 확인할 수 있었다.

세월호 참사 언론보도 연구 페이지로 이동

Studying Historical Family Records in Korea

Based on the new historical data sets of family records in Korea, stretching over seven centuries, we study historical changes in determinants of social positions.

Coming Soon