I’ve just had some of my PhD research on adapting Echo State Networks (ESNs) for identifying Parkinson’s disease published. The work describes considerations to be made when applying ESNs to classification problems, with a case study of using them to differentiate between Parkinson’s Disease patients and healthy subjects based on a longitudinal positional data source. This post will briefly summarise the work, but in case you’re interested the published version is available at the publisher’s website, while I’ve uploaded a preprint here.