Machine Learning Time-Series Platform Documentation

Installation

The latest version of cesium can be installed via pip:

pip install cesium

Note: We depend on NetCDF4, which currently has no Linux wheel, so tries to compile itself. You may therefore need to install the netcdf4 headers and library separately.

The cesium library has the following dependencies:

• numpy
• scipy
• pandas
• scikit-learn
• cython
• dask
• xarray
• NetCDF4

For parallel processing, you will also need:

• Install RabbitMQ
  • Ensure server is running with rabbitmq-server -detached

The cesium library runs on Python 2 and 3.

Testing

• pip install nose nose-exclude mock
• make test

Examples

Examples of how to use the cesium library to perform machine learning for time series data.

EEG Example

examples/EEG_Example_output

API Reference