Neural Computation Research Group
We are research & consultancy providers in the Department of Mathematics & Statistics.
The focus of the research is on computational data analysis using methodologies from statistics & machine learning. This is underpinned by methodological research and driven by real-world applications. The research is structured along the following themes-
1. Computer-Based Decision Support for Clinical Medicine
Funding: FP6 Network of Excellence Biopattern.
We support clinical research at St Helen’s and Knowsley Hospitals NHS Trust and Alder Hey Children’s Healthcare Hospitals NHS Trust.
In breast oncology we have developed & applied flexible survival models to study the progression of the disease by making smooth inference of event rates over time. This is a generic methodology for longitudinal data analysis that does not require assumptions about proportionality of hazards.
In brain oncology we are at the forefront of non-negative matrix factorisation as a methodology for tissue identification and grading from Magnetic Resonance Spectra (MRS).
Risk Modelling for Commissioning in Health & Social Care has involved the development of bespoke conditional independence maps to estimate the level of individual interventions required in a multi-sectoral plan to achieve desired changes in outcome variables. This is one of few tools available for the design of multi-sectoral interventions.
Commissioned work included National Reporting of Local Alcohol Profiles for England:
BBC News - Problem drinking shows up north-south England divisions
BBC News - Liverpool and Blackpool top alcohol problem lists
3. Sports Analytics
Funding: Premier League Football contracts through the Football Exchange
High performance sport increasingly relies on measurement data to monitor and profile players not only by performance but also by risk of injury. We develop risk models & performance models make use of structure finding algorithms combined with rigorous failure time models and advanced visualisation methods, in order to systematise the analysis of complex databases and to make this available to Sports specialists in a way that is informative and readily understood.
4. Computational Marketing
Funding: Commercial contracts
We develop scalable graphical models to map consumers and products using retail data. The results generate insights into consumer behaviour and inferences for loyalty analysis and product development.
Conditional independence map – example from a wellbeing survey
On-line implementations of personalised recommender systems: