Non-Negative Oja’s Subspace Learning Rule (NN-OSLR)

This repository contains an implementation of manuscript "Non-Negative Sparse PCA: An Intelligible Exact Approach".

Usage

1. Requirements

The requirements are in the requirements.txt file.

2. Download Dataset

You can download the dataset from here and extract them to the folder dataset (see directory structure below).

3. Train and Test

To train the algorithm, you can run

run_nnsOSLR.py --task_name mnist

To test the algorithm, you can run

run_classification.py --task_name mnist

Directory structure

.
├── code
│   ├── nnsOSLR.py
│   ├── run_classification.py
│   ├── run_nnsOSLR.py
│   └── utils.py
├── dataset
│   └── mnist
│       ├── t10k-images-idx3-ubyte
│       ├── t10k-images-idx3-ubyte.gz
│       ├── t10k-labels-idx1-ubyte
│       ├── t10k-labels-idx1-ubyte.gz
│       ├── train-images-idx3-ubyte
│       ├── train-images-idx3-ubyte.gz
│       ├── train-labels-idx1-ubyte
│       └── train-labels-idx1-ubyte.gz
├── README.md
└── requirements.txt

Reference

If you use this code in your experiments please cite this work by using the following bibtex entry:

@ARTICLE{9305265,
  author={I. {Tsingalis} and C. {Kotropoulos} and A. {Drosou} and D. {Tzovaras}},
  journal={IEEE Transactions on Emerging Topics in Computational Intelligence}, 
  title={Non-Negative Sparse PCA: An Intelligible Exact Approach}, 
  year={2020},
  pages={1-13},
  doi={10.1109/TETCI.2020.3042268}}