Siyang Li

• Bayesian framework and astronomical statistics
• Stochastic processes and probability theory
• Statistical theory
• Biostatistics

University of Auckland

Master of Science, Statistics • July 2022 — June 2023

• Master of Science with Distinction

Master thesis = Comparison of Marginal Likelihood Computation Methods with an example of Gravitational Wave Data

University of Auckland

Bachelor of Advanced Science, Statistics • March 2019 — June 2022

• First Class Honours, Senior Scholar

Honours dissertation = Inhomogeneous-Branching Brownian Motion

Macleans College

High School Diploma • January 2015 — November 2018

Diligence Award, Hillary student of the year

X-ray source intensity estimation with presence of dark and overlapping sources

supervised by David van Dyk • 2023 — Present

Using hierarchical models to estimate dource intensities.

Fourier Integral Evidence Estimation

supervised by Renate Meyer, Patricio Maturana-Russel and Avi Vajpeyi • 2023 — Present

An alternative way to estimate the marginal likelihood for posterior samples.

Inhomogeneous-branching Brownian Motion

supervised by Simon C. Harris • July 2021 — June 2022

This project explains and extends Bocharov and Harris (2014) by obtaining some corresponding mathematical results for a model where two points allow binary fission to occur.

The inflated density ratio approach for Bayesian model selection

supervised by Renate Meyer and Patricio Maturana-Russel • July 2022 — June 2023

Examples of marginal likelihood computation using the inflated density ratio approach and comparisons with other computation methods.

Risk factors for endometrial cancer and endometrial intraepithelial neoplasia in premenopausal women with abnormal uterine bleeding and a model for clinical assessment, A cohort study

with Leyi Wang, Tonghao Lei and Siqi Liang • February 2022 — Present

Bayesian Model Selection via Reversible Jump Markov Chain Monte Carlo

supervised by Matt Edwards • July 2021 — Present

By applying rjmcmc in the computation of the number of principle components to be included in Bayesian models, the methodology of algorithm construction for Bayesian model selection based on rjmcmc for core-collapse supernovae gravitational wave data is investigated.

Magnifying the differences between frequentist and Bayesian frameworks

supervised by Brendon Brewer • November 2021 — January 2022

Numerical examples where differences between frequentist and Bayesian frameworks exist solidly, followed by a tool to magnify these differences.

Auckland bus reliability study

• March 2023 — June 2023

Modeling cancellation rates and first-stop punctuality of Auckland buses using weather and route attributes.

Control and Stability of Partially Observable Markovian Queueing Systems

supervised by Azam Asanjarani • November 2020 — January 2021

Methodology of applying Quasi-Birth-and-Death Markov Chains on policy evaluation in Partially Observable Markov Decision Processes.