Modelling and Prediction of Human Behaviour (Map-HB)
Our goal is to devise new methods and techniques for modelling, understanding and predicting human behaviour from large datasets.
Our methods come from theoretical computer science (approximation algorithms, combinatorial optimisation), network science, applied mathematics and statistics.
Theoretical highlights
- new measure-adjusted p-error- to compare time-series that work on ‘peaky’ data, allowing for local time permutations
- new mathematical model for innovation diffusion (combining Ordinary Differential Equations globally with agent-based modelling locally)
Applications highlights
- Smart energy solutions
- Energy maps for small and medium enterprises (SMEs) – 3 papers, RAE project, myaems.com
- Modelling of Low voltage networks – 7 papers, video
- short term forecasting (SpringerBriefs book)
- Measuring effects of storm naming in press, traffic and on Twitter (MetApps journal paper, the research was mentioned on ITV Meridian News on 31/05/2019)
- Rumours on Twitter
- Modelling spread of mood on Twitter
- Wellbeing interventions in social networks
- Modelling conversations on Twitter
- Political debate data analytics, clustering and visualisation
