Youngmin Ju

Data Scientist / Economist

PhD in Economics @ USC

MA in Economics @ Korea Univ.

BS in Mathmatical Sciences @ KAIST

Publications

PUBLISHED PAPERS (SCI/SSCI-LISTED):

Ju, Y.-M. and Lee, M.-J. (2017), Control Function Approach for Partly Ordered Endogenous Treatments: Military Rank Premium in Wage. Oxf Bull Econ Stat, 79: 1176-1194. doi: 10.1111/obes.12199

Abstract

In treatment effect analysis, there are many cases where the treatment of interest is ordered (e.g. general-education years or medicine doses) and the control treatment is not zero, but a different type of treatment (a vocational training or a surgery). We develop an approach to find effects of partly ordered treatments, while correcting for possible treatment endogeneity with nearly parametric control functions. We use this control function approach, along with its supplementary version, to estimate effects of military ranks (ordered treatments) on wage relative to non-veteran status (control treatment) with the Wisconsin Longitudinal Study data. In our empirical analysis, the military rank effects differ much: officer has large positive effects, but enlisted ranks have small or no effects.

  • Developed a method to find the effects of partly ordered treatments while correcting for possible treatment endogeneity with nearly parametric control functions
  • Estimated effects of military ranks (ordered treatments) on wage relative to non-veteran status (control treatment) and discovered that the military rank effects differ much: officer has large positive effects (17.9%) but enlisted ranks have near-zero effects

PUBLISHED PAPERS (Non-SCI/SSCI-LISTED):

Lee W.H., K.H. Park, Y.M. Ju, S.J. Kim and D.S. Lee. (2013), A Study on Scale of Defense Expenditure for Security Menace: A Panel Regression Analysis Approach, Journal of Korea Army Academy at Yeong-cheon, 76(1), 207-242 (in Korean)

  • Conducted a government research project, which is a part of a project with a total value of over $40,000, to design robust economic models to estimate optimal national defense R&D expenditure and efficient management

RESEARCH PROJECTS:

(2021) Affirmative Action in Korea - Regression Discontinuity with Multiple Assignment Variables

  • Developed an identification of a fuzzy regression discontinuity design (RDD) with multiple assignment variables to analyze the effect of Affirmative Action in Korea on female employment rate in the private sector
  • Discovered that while the overall policy has no effect, but a partial effect (company size) increases female employment rate by 5% points

(2021) Store Item Demand Forecasting Project, kaggle store item demand data

  • Implemented Keras (TensorFlow) to deploy a Recurrent Neural Net (RNN) with Long Short-Term Memory (LSTM) to predict 3 months of item sales at different stores to build baseline sales predictions to help with cash flow management, business planning and strategy
  • Reduced error rate attained by LSTM to 86% of ARIMA’s error rate

(2020) Customer Churn Prediction Project, kaggle customer churn data

  • Identified the customers most likely to churn and the features with the greatest impact on churn by building a multi-classification model with XGBoost and investigating features with logistic regression
  • Confirmed that XGBoost outperformed the rest of the tested algorithms with an Area Under Curve (AUC) value of 93.3% (GBM 90.89%, Random Forest 87.76%, Decision Trees 83%)

(2020) Online Retail Project, kaggle online retail data

  • Segmented and cleaned business performance metrics such as monthly revenue, activation rate, monthly retention rate, and churn rate and conducted Lifetime Value (LTV) methods, increasing accuracy of a multi-classification model from 76.5% to 84%

(Jan. 2012 – Dec. 2012) A Study on the Estimation of Optimal Defense R&D Expenditure and Efficient Management, Korea Army Academy at Yeongcheon (KAAY)

(Jan. 2011 – Nov. 2011) The Study on Scale of Defense Expenditure, KAAY

(Oct. 2009 – Mar. 2010) Economic Effects of Alleged Anti-Competitive Behavior of Top Automobile Parts Company, Research Assistant, Hyundai MOBIS

  • Designed causal inference models to investigate the economic effects of alleged anti-competitive behaviors of Hyundai Mobis on retail agencies, mediating companies,repair shops, and consumers.
  • Conducted research and crafted economic evidence to reduce the fine from $150 million to $30 million