Skip to main content
School of Mathematical Sciences

PSD - Dr Linqi Wang

Statistical Methods for High-Frequency and High-Dimensional Financial Data

Supervisor: Dr Linqi Wang

Project description:

Modern financial markets generate vast amounts of data on individual transactions, price quotations and activity in limit order books. These data offer opportunities to understand how financial risk, dependence and liquidity evolve within the trading day. This project will develop statistical methods that use ultra-high-frequency data to estimate key market quantities accurately and model their dynamics across many assets. Extracting reliable information at such fine time scales presents substantial challenges. Observed prices are affected by market microstructure noise arising from the trading process. Transactions occur at irregular times, observations are not synchronised across assets, and trading intensity varies throughout the day and around important events. Consequently, using more observations does not automatically improve estimation. When many assets are analysed together, these measurement problems are compounded by the growing complexity of modelling their dependence.

Research directions:

A focused PhD programme will be developed around the following two closely connected aspects. Accurate and robust measurement. The first aspect focuses on developing robust estimators for quantities such as volatility, covariance, correlation and market liquidity from transaction, quotation and order book data. Depending on the research focus, the project could estimate these quantities over a trading period or track how they change within the day. Statistical questions include how to separate underlying market variation from trading noise, accommodate irregular and asynchronous sampling, and distinguish abrupt changes from measurement effects. For liquidity, an additional challenge is to construct measures that reflect trading costs or price impact while accounting for changing trading activity. Multivariate and high-dimensional modelling. The second aspect focuses on capturing changing relationships among many assets while keeping estimation reliable and computation manageable. Possible approaches include factor models that summarise common market and sectoral movements, regularisation methods that stabilise estimation in large models, and flexible

dynamic models that allow dependence and liquidity conditions to evolve over time. The methods will be designed to ensure that estimated covariance and correlation matrices are mathematically valid and to distinguish genuine cross-asset dependence from correlated measurement noise. A key question connecting these aspects is how uncertainty in estimated market measures affects subsequent modelling, statistical inference and forecasting. Potential applications include assessing portfolio risk, identifying common liquidity shocks, and examining changes in dependence around news announcements or periods of market stress. These applications would provide a setting for evaluating whether improved measurement leads to more reliable risk assessment and better forecasting performance.

Project approach and applicant profile:

The research will combine statistical theory, simulation studies and empirical analysis. It will focus on developing new estimators and models, investigating their theoretical properties, designing computationally efficient algorithms and evaluating their performance on financial market data. The project would suit applicants interested in statistics, econometrics, probability or financial mathematics who enjoy combining mathematical modelling with computation and data analysis. It offers scope for methodological research motivated by practical questions about how financial markets function.

If you wish to inquire about the project, please contact the supervisor Dr Linqi Wang linqi.wang@qmul.ac.uk.

Funding Notes:

This project is open to applicants applying through the New Talent Research Enabling Scheme only.

New Talent Research Enabling Scheme => 3 years stipend and Home tuition fees. Students with International fee status are required to cover the difference between Overseas and Home fees. Stipend at QMUL rate (e.g. 2026/27 QMUL annual stipend rate £22,618 full-time; 2027/28 rate tbc).

Further information: 
How to apply 
Entry requirements 
Fees and funding

 

 

As one of the UK’s most diverse universities, QMUL fosters an inclusive and supportive academic community.

The School of Mathematical Sciences is committed to the equality of opportunities and to advancing women’s careers. As holders of a Bronze Athena SWAN award, we offer family-friendly benefits and support part-time study.  

 

Back to top