/hsmm4acc

Behaviour detection in wearable movement sensor data

Primary LanguagePythonApache License 2.0Apache-2.0

DOI

HSMMs for accelerometer data

This repository is for the project by NLeSC in collaboration with UCL, of analyzing movement sensor data of adolescents. It contains code for processing accelerometer data using Hidden Semi Markov models (using the pyhsmm package). This software is meant for accelerometer data processed with the R-package GGIR.

Please cite the DOI when using this software.

Installation

Prerequisites:

  • Python 2.7
  • pip

Instalation with conda

Navigate to the root of this repository. Create a conda environment with the environment.yml file:

conda env create -f environment.yml

Activate the environment:

source activate ucl

Then install the package:

pip install .

Installation with pip

Navigate to the root of this repository. To install, try:

pip install .

Installation troubleshooting

ImportError: No module named hmm_messages_interface

The pyhsmm package needs the right gcc compiler (it seems to work with gcc 4.7). You can clone the pyhsmm package and compile it:

python setup.py build_ext

Which should solve the issue. See also mattjj/pyhsmm#55.

Intel MKL FATAL ERROR: Cannot load libmkl_avx.so or libmkl_def.so.

You can disable the use of mkl with: conda install nomkl

Running tests

To run tests:

nosetests test/