Skip to content

A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data

Notifications You must be signed in to change notification settings

victormvy/sigma-convkernels

Repository files navigation

A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data

This is the companion code for the paper "A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data".

Authors are:

  • Víctor Manuel Vargas (@victormvy)
  • Riccardo Rosati (@rosati1392)
  • César Hervás-Martínez (chervas@uco.es)
  • Adriano Mancini (a.mancini@staff.univpm.it)
  • Luca Romeo (@whylearning22)
  • Pedro Antonio Gutiérrez (@pagutierrez)

Instructions

The following has been tested to run on an up-to-date Linux installation (Debian 10 buster).

Preparing the environment

You can use anaconda or miniconda with an environment that has at least Python 3.9 installed.

Then, you can install the requirements:

pip install -r requirements.txt

Preparing the data

The dataset presented in this work is contained in this repository in a compressed zip file. To run the experiments, you should unpack it first:

cd data
unzip sigma_pdm.npy.zip

The file obtained after unpacking the zip file is a numpy binary format. The function to load this dataset is enclosed in functions.py. However, the dataset in time series .ts format is also included in data/sigma_pdm.ts.zip.

Running the experiments

All the experiments can be run using the run.py script:

python run.py

Citation

BibTex

@article{vargas2023hybrid,
	title = {A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data},
	journal = {Engineering Applications of Artificial Intelligence},
	year = {2023},
	author = {Víctor Manuel Vargas and Riccardo Rosati and César Hervás-Martínez and Adriano Mancini and Luca Romeo and Pedro Antonio Gutiérrez}
}

About

A hybrid feature learning approach based on convolutional kernels for ATM fault prediction using event-log data

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages