nmf

Assignment1 Non-negative matrix factorization

Author

Chen Chen 480458339 Xiaodan Gan 440581983 Xinyue Wang 440359463

Running Environment Setup

We implemented NMF as a python package with modules of noise , file IO, evaluation metrics and visualization. Please follow the instruction to make sure the code will run properly.

1. Make a folder with name "data" in current directory. Then copy ORL and CroppedYaleB dataset inside. Please make sure you have the following file tree structure: |--- data \ ​ |--- ORL \ ​ |--- CroppedYaleB \ |--- nmf \ |--- _version_.py \ |--- algorithm.py \ |--- io.py \ |--- main.py \ |--- metric.py \ |--- noise.py \ |--- util.py \ |--- vis.py \ |--- setup.py \ |--- README.md \

2. Install nmf with following command: (Please use pip3 if the default python in your computer is python2)

$ pip install -e .

This command will run setup.py where we specify the dependencies required to run nmf. The dependencies we require are:

       "numpy>=1.14.0",
       "scipy>=1.0.0",
       "scikit-learn>=0.19.1",
       "pandas>=0.20.2",
       "matplotlib>=2.0.2",
       "Pillow>=5.2.0",
       "tqdm>=4.26.0",

Please note that if the version number of installed package in your machine is lower than the stated version number, pip will uninstall your out-of-date package and install the one with version number greater than or equal to the stated one in setup.py.

3. Run main.py in nmf with appropriate dataset name.

   To run nmf on ORL dataset, please run:

   python nmf/main.py orl
   

   To run nmf on CroppedYaleB dataset, please run:

   python nmf/main.py croppedYale
   

Progress bar will show up to indicate the rough running time for each algorithm and noise combination. All results will be auto-saved to folder results/{generated-time-dataname}. Note that we set the epoch to be 1 in main.py. This is because we have 4 (noise) x 2 (algorithm) = 8 combination in each epoch. This will cost around 4.5 minutes on a i7-6th gen laptop with ORL dataset. However, we increased the epochs to calculate average metrics and confidence interval etc.