Dear all,

We are pleased to announce a special joint event presented by the Turing
Data Science for sports, activity and well-being interest group and the RSS
Statistics in Sport Section.

The event is free and takes place *Friday 22nd June 14:00-16:00* at The
Alan Turing Institute, the event will feature two excellent speakers, Prof
Kerrie Mengersen and Dr Paul Wu (see below for titles and abstracts). As
well as joining the event locally, it will be possible to stream the talks
via the Turing's system, see https://turing-uk.zoom.us/j/232301158 to join
the stream (anybody can connect on the day with no password).

If you are wishing to join the event locally we ask that you register via
eventbrite at https://www.eventbrite.co.uk/e/hidden-markov-models-and-dyna
mic-bayesian-networks-for-sports-and-resilience-modelling-de
veloping-a-tickets-46987204995
to aid registration. Those physically joining will need to register.

many thanks,
The Turing and RSS sports interest groups

*Hidden Markov Models and Dynamic Bayesian Networks for Sports and
Resilience Modelling (Dr Paul Wu)*
Recent advances in technology have made possible the collection,
management, analysis and application of data and models to support
individualised management of complex organisations and systems. In many
domains, however, current models and/or approaches have not yet adapted to
the deluge of new data and challenges associated with this data given the
complexity of the underlying system. As a result, they are limited in the
insight that they can provide.

We discuss the development and application of Bayesian methods and state
space models including Hidden Markov Models (HMMs) and Dynamic Bayesian
Networks (DBNs) as motivated by case studies in ecosystem management and
sports. HMMs sampled using the Bayesian framework can provide a way to
better understand and predict changes in the physiological state of
individual athletes. With a focus on individualised performance and
management, we used the Bayesian approach to infer model parameters and
predictions from cycling datasets focused on individual athletes.

On the other hand, DBNs provide an approach to modelling complex system
processes and their interactions over time. We use them to incorporate
diverse data sources and models at different scales including quantitative
data and expert knowledge. A non-homogeneous approach was developed to find
timing-based management strategies to minimise the impact of stresses on
ecosystem resilience.

*Developing a Global Short Course on Bayesian Statistics in Sport (Prof
Kerrie Mengersen)*
Bayesian statistics has held a niche role in sports analytics for over a
decade, but interest is growing exponentially. This interest is fuelled in
part by the capabilities of Bayesian statistics in terms of flexible
modelling, probabilistic prediction and ranking, accommodation of
uncertainty and so on. There is now a wide range of journal articles,
online discussions, books and sports-tech analytics companies that focus on
Bayesian approaches, but there is a commensurate lack of formal training
for those who want to understand, develop and utilise Bayesian statistics
in sport.

In this presentation, I will discuss our efforts in this direction and
canvas ideas about how we can help to spread the word in a more effective
manner.

You may leave the list at any time by sending the command

SIGNOFF allstat

to [log in to unmask], leaving the subject line blank.