System requirement

This software implements the anomaly detection algorithms proposed in paper https://ieeexplore.ieee.org/abstract/document/9006046. To reproduce the results, you need to go to ANL ALCF website to download MIRA data first https://reports.alcf.anl.gov/data/mira.html.

I suggest users to install these packages with anaconda.

Compile

Use make file to compile code and test cases with command "make".

Unzip all RAS_EVENT data you have downloaded. Unzip the dataset and put in the same folder.

Use the R code in https://github.com/QiaoK/AMASE/blob/master/developer_note.txt to clean up the data.
Compile the C++ code. Use the "./streaming_features" to generate the training and testing features.
Usage: ./streaming_feature warn_table_name fatal_table_name fatal_t_threshold window_size lead_time_count
Example: ./streaming_feature sys_events.csv sys_fatal_spatial.csv 1 300 4
Run the python training code for neural network training.
Usage: python training_lead_time.py feature_file_name training_epochs retrain(0 or 1) reload_data(0 or 1) lead_time_interval(3600) lead_time_stride(1)
feature_file_name is the output file generated by ./streaming_feature

Contact me at qiao.kang@eecs.northwestern.edu if you have questions.

Name		Name	Last commit message	Last commit date
Latest commit History 26 Commits
.gitattributes		.gitattributes
.gitignore		.gitignore
COPYRIGHT		COPYRIGHT
Makefile		Makefile
README.md		README.md
amase.h		amase.h
attributes.txt		attributes.txt
batch.R		batch.R
data_functions.cpp		data_functions.cpp
data_functions.h		data_functions.h
developer_note.txt		developer_note.txt
fatal_attributes.txt		fatal_attributes.txt
fatal_lags.txt		fatal_lags.txt
feature_save_main.cpp		feature_save_main.cpp
filter_attributes.txt		filter_attributes.txt
filter_coolant.txt		filter_coolant.txt
hclust.R		hclust.R
job_train.py		job_train.py
lead_times.txt		lead_times.txt
random.c		random.c
random.h		random.h
script.R		script.R
script.sh		script.sh
st_clustering.cpp		st_clustering.cpp
st_clustering_main.cpp		st_clustering_main.cpp
streaming_features_main.cpp		streaming_features_main.cpp
sys_attributes.txt		sys_attributes.txt
train_window_features.py		train_window_features.py
training_anomalies.py		training_anomalies.py
training_lead_time.py		training_lead_time.py
training_lib.py		training_lib.py
training_nersc_data.py		training_nersc_data.py
training_window.py		training_window.py
window_prediction.cpp		window_prediction.cpp
window_prediction_main.cpp		window_prediction_main.cpp