The geological guided adversarial learning model for geochemical anomaly detection

by Chunjie Zhang

Installation

Install Anaconda and Pycharm in computer

Requirements

time

numpy

tensorflow

logging

importlib

sys

os

sklearn

Training and Testing

Download all python files from this Github platform

Once finished this step, you can open the Pycharm in your computer and install all files in Pycharm.

Firstly, open arrhythmia.py, checked the code and set the data storage path

code: path_ = 'D:/1_data science/1_code/ALAD_myDATA/data/arrhythmia.mat'

The dimension of data is up to you, pleasr don't forget to check the dimension in code to keep consistent with your data.

Secondly, run the file 'run.py'

Thridly, after running finished, the two folds "train-logs" and "results" will be automatic produced.

The results files include evaluation results of the mode with the training step.

Some paratermers need to be adjusted for achieving better performance of models, such as

learning_rate

batch_size

latent_dim

Reference

Adversarially-Learned-Anomaly-Detection

ALAD (Proceedings of IEEE ICDM 2018) official code

The code for the paper "Adversarially Learned Anomaly Detection" (authors: Houssam Zenati*, Manon Romain*, Chuan Sheng Foo*, Bruno Lecouat, Vijay Ramaseshan Chandrasekhar) is now open source!

geological constrained ALAD model was developed by Chunjie Zhang

The code for the paper ["A geologically constrained deep learning algorithm for recognizing geochemical anomalies " (authors: Chunjie Zhang, Renguang Zuo; Yihui Xiong)] if you have any question, you can contact me via email [email protected]