Implementations of Classification Algorithms
This repository demonstrates the application of different Machine Learning tools on a real world classification problem. The data being used is the Bank Marketing Data Set from the UC Irvine Machine Learning Repository.
The tools that were applied to the data are
If you want to run the algorithms with Weka or KNIME, you will need a local installation of the software. For the other tools, several example implementations are available as Jupyter Notebooks.
You can run the notebooks in your browser without any installation if you use the links below. They will access a Jupyter environment on the cloud service. If this repository has changed recently, mybinder.org will have to rebuild a Docker image for this environment which might take a while. Once the Docker image is available, the environment will be up in less than a minute.
If you insist on running Jupyter locally, you'll need to install a few things. Run Jupyter Notebooks locally has more information on that.
- Logistic Regression with TensorFlow :
- Classification by a Neural Network implemented with Keras :
- Logistic Regression on a standard benchmark, the MNIST dataset :
Binder Configuration Files
This repository uses the following Binder Configuration Files
- environment.yml : the Conda environment for the notebooks
- runtime.txt : defines the R runtime version
- install.R : specifies the required R packages
- apt.txt : specifies Linux packages for Octave (see also https://github.com/binder-examples/octave)
- postBuild : activate Jupyter TOC extension