Ridiculously Simple Neural Network for nonlinear regression
This is an incredible simple feedforward neural network that I've implemented. It's written in python and it has this awsome features:
- Nonlinear regression with sigmoid functions
- A single neuron NEW! MORE THAN ONE NEURON
- No gradient descending function NEW SORT OF GRADIENT DESCENDING - Need help!
- No ability to scale input YEAH I FORGOT ABOUT THAT
- Tons of bugs LESS TONS BUGS!
This fantastic software is provided AS IS and without any warranty.
This is just for academic purpose while I'm studying nonlinear regression and model identification
RSSN is written in Python 3. I personally recommend to install Continuum Anaconda. Once you've installed it just run
python3 rssn.py
If you don't want to install the complete Anaconda package first install dependecies:
pip3 install -r requirements.txt
and then execute the script with:
python3 rssn.py