Neural net architecture for GuanLab's winning 2017 PDDB submission

This neural net architecture was used to generate the top scoring submission to the 2017 Parkinson's Disease Digital Biomarker DREAM Challenge, Subchallenge 1 https://www.synapse.org/#!Synapse:syn8717496/wiki/422884.

guanlab_nn_architecture()

Value

keras sequential model

Details

It should be noted that GuanLab themselves wrote that: "The parameters and structures of my network does not really matter, as it is the simplest that can be seen in the deep learning field". Nevertheless, this was the top scoring submission, and can be used to generate some robust predictions to distinguish Parkinson's patients from controls.