NAC

Neuron Automatic Classification

Tool for automatic classification of biological neural signals.

Table of contents

• Descritption
• User installation
  • Dependencies
  • Windows
  • MacOS and UNIX
• Source code
• Customer
• Contact
• Related documents
  • Specifications
  • Report
  • Defense
• Usefull links

Description

Electrophysiological data for two distinct populations of neurons were collected by several research teams with common experimental measurements and parameters. The goal of the project is to find a way to automatically classify neurons from these data with the best possible prediction rate. This alternative would have several advantages such as avoiding the use of transgenic animals.

A dataset is made available to students with the category (1 or 2) of each neuron. Ideally a supervised learning method is envisaged using the Caret library and the R software but also using the Python language and the Scikit-Learn library. Any other solution can be considered and this can be discussed with the group of students.

User installation

Dependencies :

• Python2.7
• Sckiti-leran: sklearn
• Matlplotlib
• Tkinter

Windows :

Just unpack the file : Classification-Neuronale.zip and run LaClassificationNeuronale.exe

MacOS and UNIX :

Before use the script you need to install some packages:
-python2.7 (if you are on MacOS)
-pip
-numpy
-scipy
-scikit-learn
-matplotlib
-Tkinter

For the first time, open a terminal and use these command :

     sudo apt-get install python2.7
     python get-pip.py
     pip install -U numpy,scipy,scikit-learn
     python -m pip install matplotlib
     apt-get install python-tk

If you already have a working installation of any of these packages :

     pip -review --auto
     sudo apt-get update && apt-get upgrade

Source code

You can check the latest source swith the command :

     git clone https://github.com/0AlphaZero0/Projet-Neuro.git

Customer

André Garenne (andre.garenne@u-bordeaux.fr), Team MNEMOSYNE, laboratory IMN

Contact

Author :
Thouvenin Arthur : athouvenin@outlook.fr

Contributors :
Blais Benjamin : ben.blais@laposte.net
Cottais Déborah : d.cottais1@gmail.com
De Oliveira Lila : lila.de-oliveira@etu.u-bordeaux.fr
Jouan Clément : jouan.clement@hotmail.fr

Related documents

Order :

Link to order (in french)

Specifications :

Link to specifications (in french)

Report :

Link to report (in french)

Final Report :

Link to final report (in french)

Defense :

Link to defense (in french)

Usefull links

• Scikit-Learn : Sklearn

  • SVM from Scikit-Learn : SVM
  • Neural Network from Scikit-Learn : Neural Network

• Playlist of youtube videos which explain how scikit-learn work (by Sentdex) : Scikit-learn Machine Learning with Python and SKlearn