Haoyuan Zhang

---

I am a Computer Science Ph.D. Candidate and Research Assisant in Risk & Information Management Group, Queen Mary University of London (QMUL).

My supervisors are Dr. William Marsh, Prof. Norman Fenton, and Prof. Martin Neil.

---

Research Interests

I am working on Bayesian modelling and machine learning techniques in the fields of decision support for asset maintenance and data mining from medical health records.

I work on:

• Hierarchical Bayesian statistical models and Bayesian Networks
• Probabilistic Relational Models
• Machine learning techniques for classification and feature engineering

With application on:

• Medical knowledge discovery
• System safety and reliability
• Inspection and maintenance decisions support

 

---

Education

Queen Mary University of LondonSep 2015 – Present
Ph.D. student in Computer Science (supervised by Dr. William Marsh, Prof. Norman Fenton and Prof. Martin Neil): Using Bayesian Networks and Complex Data to Optimize Infrastructure Maintenance in Railways

 

 

The University of Hong Kong Sep 2013 - Dec 2014
M.Sc. in Industrial Engineering and Logistic Management (supervised by Dr. Tak Nam, Wong): Colour Petri Net – based Modelling for Integrated Process Planning and Scheduling. My master thesis obtained the highest grade among the department

 

Jinan University Sep 2009 - Jun 2013
B.Sc. in Electronic Commerce (supervised by Dr. Hua, Bai): Tourism Supply Chain Collaborative Demand Forecasting Model based on Colour Petri Net. Awarded for the best undergraduate thesis of Jinan University

 

---

Other Roles

Research Assistant on Project: Knowledge Discovery from Health Use Data (KNIFE)Apr 2019 - Present

• Data Processing - Use PostgreSQL to select cohorts of patients from clinical databases and preprocesses selected cohorts (e.g. missing data imputation)
• Data Analysis - Perform Exploratory Data Analysis (EDA), including feature projection (e.g. Principal Component Analysis (PCA)), feature selection (e.g. random forest), Conditional Independence(CI) tests (e.g. mutual information)
• Model Building and Validation - Build Bayesian Networks based on elicited medical knowledge and CI tests, fit their parameters with Maximum Likelihood estimates or Bayesian Parameter estimates, and compare the fitted models’ performance with other machine learning techniques
• Application - Generate synthesis clinical data from the learned models using forward/rejection sampling to bypass the restriction of health data usage

Ph.D Research Committee RepresentativeJune 2016 - Mar 2018
I am the Ph.D. research committee representative for Risk and Information Management Group (RIM), EECS, QMUL

Liguo Steel Group (HK) LimitedJun 2014 - Apr 2015
I was a manager trainee, worked on commodity shipping, trading and financing

 

---

Contact

Email: haoyuan.zhang@qmul.ac.uk

Office: CS437, Computer Science Building, Queen Mary University of London, Mile End Road, London E1 4NS