JIACONG Yuan

• “Estimate of value of VaR of the portfolio using Monte Carlo simulation based on Copula theory” From 02/2024 to 05/2024

Constructed an equally weighted portfolio of the SSE Composite Index and Nikkei 225, using weekly closing prices from 2000 to 2023. We employed an AR(p) model and GARCH(m, n) model to capture time-varying means and volatility. Using the Box-Jenkins methodology to identify lagging order. Estimated parameters, and performed residual diagnostics with the Ljung-Box test. Residuals were fitted to two-dimensional Copula model was fitted. Using Monte Carlo simulations generating 10,000 simulated pairs of data to compute the VaR.

• “Estimate of the Proportion of 'Leave' Votes in Each UK Borough”. From 02/2023 to 05/2023

 Revised and simplified variables related to voters' demographics and careers into indicator variables. Used hierarchical clustering to categorize and simplify 45 covariates, then applied Principal Component Analysis to reduce the model to 23 variables. Constructed three statistical models: a linear regression model,  a generalized linear model with a binomial family and logit link, a Generalized Additive Model with a quasibinomial family. Adjusted these model based on p-values,  residual plots, and back testing.

• FRM ( Passed Exam Part 1 )