XINYI XU

An empirical analysis of interactions among Asian stock markets, 2008-2023

• Select weekly closing stock prices covering the period of January 2008 to November 2023
• Test the stationarity and cointegration of the stock indices and return series
• Use Vector Error Correction Model (VECM) to capture the long-term and short-term relationships among cointegrated pairs
• Apply ARIMA model to the return series and make future prediction

A Statistical Study Of Credit Analysis For Lending Club Loans

• Selected 14 significantly influential variables and 50,000 observations out of 53 variables and 260,000 observations, by data cleaning methods (missing values and outliers), from public data pool of Lending Club, a peer-to-peer lending platform, through the time period between 2013 and 2014
• Divided the variables into category variables and numeric variables according to the features of outcomes, then conducted exploratory data analysis to figure out distributions and relationships between the variables
• Applied statistical classification models (Logistic Regression, Random Forest) to identify high-risk individuals, since the response variable (Loan Status) is binary (Default or Not Default)
• Evaluated the performance of each model, and concluded that Random Forest model achieved a relatively low error rate while keeping a reasonable false negative rate; As for Logistic Regression model, to choose an optimal threshold is rather a trade-off problem and further studies are needed

Portfolio Theory On Stock Price Analysis

• Selected stocks for the portfolio and collected closing price data from Yahoo finance
• Implemented Principal Component Analysis and Portfolio Theory to identify the highly correlated stocks and solve the optimization problem in R (quadratic programming)
• Computed minimum variance portfolio and tangency portfolio for selected stocks, estimated value-at-risk and expected shortfall, then constructed their confident intervals using bootstrap method

A Study On The Relationship Among Oil Prices, Exchange Rates, Stock Prices, and Interest Rates Using an SVAR approach

• Applied SVAR model to identify the multi-directional, contemporaneous causality between oil prices and some other macroeconomic variables

Can additional shocks help to explain the US business cycles?

• Parameterized Expectations Algorithm

A Case Study in Option Valuation in Computational Finance

• Multilevel Monte Carlo Method

Member of American Economic Association;

Society of Actuaries/Canadian Institute of Actuaries (EXAM P);