mainak-ghosh / AutoEncoder

Comparison of dimensionality reduction ability of different autoencoders on different datasets.

Autoencoder

Comparison of dimensionality reduction ability of different autoencoders on different datasets. Used Keras deep learning API for building deep network models for this study. Also used other state of the art dimension reduction techniques to compare the result.

Types of Autoencoders used for comparison:

1. Simple Autoencoder
2. Sparse Autoencoder
3. Deep Autoencoder
4. Stacked Autoencoder
5. Denoising Autoencoder
6. Convolutional Autoencoder

Datasets:

1. Forrest Cover Type Dataset: https://www.kaggle.com/uciml/forest-cover-type-dataset
2. MNIST Digits Dataset : https://www.kaggle.com/oddrationale/mnist-in-csv
3. Wisconsin Breast Cancer Dataset : https://www.kaggle.com/uciml/breast-cancer-wisconsin-data

Documentation of the project is given in "AutoEncoder_Project.pdf" file.