Convolution Neural Network-Based Expert Recommendation for Alert Processing

Haidong Huang, Liming Wang, Rao Fu, Jing Yu, Ding Yuan, Danqi Li

2023

Abstract

Alerts and associated trouble tickets provide extremely useful information for IT system maintenance. However, the continuously occurrence of thousands of tickets also leads to a big challenge for accurately and effectively dispatching them to skilled experts for quick problem-solving. To cope with such a challenge, this paper develops a convolution neural network-based expert recommendation approach for intime alert processing. First, expert profile is built by extracting domain words from historical tickets, and is encoded into a sentence like problem description. Second, an attention-based convolution neural network (CNN) is developed not only to learn a unified representation for both problem description and expert profile but also measure the semantic similarity between them. Finally, an ordered expert list is outputted. We evaluate our approach on a real-world data set. Experimental results show that compared to the best baseline approach, our approach can not only improve 2.8% in terms of p@1 but also shorten 11.7% in terms of the mean number of steps to resolve (MSTR).

Paper Citation

in Harvard Style

Huang H., Wang L., Fu R., Yu J., Yuan D. and Li D. (2023). Convolution Neural Network-Based Expert Recommendation for Alert Processing. In Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT; ISBN 978-989-758-677-4, SciTePress, pages 401-408. DOI: 10.5220/0012284900003807

in Bibtex Style

@conference{anit23,
author={Haidong Huang and Liming Wang and Rao Fu and Jing Yu and Ding Yuan and Danqi Li},
title={Convolution Neural Network-Based Expert Recommendation for Alert Processing},
booktitle={Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT},
year={2023},
pages={401-408},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0012284900003807},
isbn={978-989-758-677-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Seminar on Artificial Intelligence, Networking and Information Technology - Volume 1: ANIT
TI - Convolution Neural Network-Based Expert Recommendation for Alert Processing
SN - 978-989-758-677-4
AU - Huang H.
AU - Wang L.
AU - Fu R.
AU - Yu J.
AU - Yuan D.
AU - Li D.
PY - 2023
SP - 401
EP - 408
DO - 10.5220/0012284900003807
PB - SciTePress