Output details

Classification of newborn EEG maturity with Bayesian averaging over decision trees

<13> This paper represents a new set of features discovered in newborn electroencephalograms as part of a Leverhulme-funded project. Being tested on a set of 1,000 clinical recordings, these features provide most accurate and reliable estimates of risks, that clinicians urgently need. Such estimates are achieved with Bayesian methods which however has a limited scale of applications. This paper reports the first result of successful application of the method to a large-scale clinical problem. The paper initiated a collaboration with a leading research group of Prof Witte in Germany. A PhD thesis was successfully completed and available via http://uobrep.openrepository.com/uobrep/handle/10547/293806.