This is a programming project for Neurocognitive Methods and Programming (SoSe 2020).
This is a Neural Network Programming Project (NNPP) module for Neurocognitive Methods and Programming (SoSe 2020). This module provides classes for preprocessing and multi-, and binary classification using simple neural network based on the "Haxby dataset" http://data.pymvpa.org/datasets/haxby2001/ with .csv labels.
Those are required to run this module:
- python : 3.6.10
- nibabel : 3.1.1
- nipype : 1.6.0 dev0
- nilearn : 0.6.2
- pandas : 1.0.5
- matplotlib : 3.2.2
- tensorflow : 1.13.1
- keras : 2.2.4
- jupyter notebook : 6.0.3
And "You must install FSL to work with the nipype in this module".
This was tested on the docker "jihoonkim2100/nnpp" https://hub.docker.com/r/jihoonkim2100/nnpp environment
which leverage tensorflow and keras based on "nipype/nipype":
- docker version : 19.03.1
- docker image OS : Debian GNU/Linux 9
- host OS : Windows 10 Home
Belows are the reference for this project:
- scikit-learn https://scikit-learn.org/stable/
- nilearn https://nilearn.github.io/index.html
- nipype https://nipype.readthedocs.io/en/latest/index.html
- keras https://keras.io/
Additional information : https://osf.io/qu3y7/
Author : JiHoon Kim
Last-modified : 13th July 2020