Python implementation of the unsupervised Deep Learning Algorithm SOM
A self-organizing map (SOM) is a type of artificial neural network (ANN) that is trained using unsupervised learning to produce a low-dimensional (typically two-dimensional), discretized representation of the input space of the training samples, called a map, and is therefore a method to do dimensionality reduction.
To read more about SOMs, check out this link
#Dataset The dataset Credit_card_applications.csv is contains several unnamed columns which are used to determine if the transaction conducted was fraudulent or not. It can be downloaded from the repository.
-
Download the dataset and set the folder containing the downloaded data as the working directory.
-
Make sure you have all the libraries used in the som.py file. In case you need to download any of the libraries, use this command on your Command Prompt:
pip install 'your library name'
- Once you have all the libraries imported, copy the code from som.py and run it.