Modelling Latent Travel Behaviour Characteristics with Generative Machine Learning
We implement an information- theoretic approach to travel behaviour analysis by introducing a generative modelling framework to identify informative latent characteristics in travel decision making. It involves developing a joint tri-partite Bayesian graphical network model using a Restricted Boltzmann Machine (RBM) generative modelling framework.
SP and RP survey conducted for a new train service between Montreal and New York (Train Hotel).
Dataset Tech report: Sp_TrainHotel_Draft 5_Oct 27.2016.pdf
This is a starting point if you are new to this project where you will use the python project as-is to generate the estimation model.
Python 3.5+ (with pip3), Numpy, Pandas, Theano
These are the installation instructions. Consider them work-in-progress and feel free to make suggestions for improvement.
- Clone or download the git repository and navigate to the project folder
The following system packages are required to be installed
apt-get install python3 python3-dev pip3
python3 --version
>>> Python 3.X.X
Install requirements with pip with --user option
cd project-root-folder/
pip3 install --user -r requirements.txt
The above command also installs the latest Theano from github.com/Theano/Theano
Two options:
- Install Python directly (instructions)
- By Anaconda (instructions)
verify Python is installed correctly:
Open cmd and run:
C:\>python
> Python 3.X.X. ...
Install project requirements
cd project-root-folder/
pip install -r requirements.txt
From the project folder,
To run MNL model:
python3 run_mnl.py
To run MXL model:
python3 run_mxl.py
To run ICLV model:
python3 run_iclv.py
Please read CONTRIBUTING.md for details on contributing.
0.1 Initial version
- Melvin Wong Github
See also the list of contributors who participated in this project.
This project is licensed under the MIT - see LICENSE.md for details
- BIOGEME http://biogeme.epfl.ch/