Author(s)

ZhiYi Tan

Affiliation(s)

School of Communication and Artificial Intelligence, School of Integrated Circuits, Nanjing Institute of Technology, Nanjing 211167, Jiangsu, China.

Corresponding Author

ZhiYi Tan

ABSTRACT

As a core infrastructure of modern society, the stability of communication networks directly affects the operation of key areas. Traditional passive maintenance methods suffer from drawbacks such as delayed fault detection and difficulty in real-time monitoring. Machine learning's strong data processing and pattern recognition capabilities offer a new path to solve this problem. This study aims to construct a machine learning-based communication network link quality evaluation and fault prediction system to achieve accurate link quality assessment and early fault warning. The research designs a complete research framework including data preprocessing, feature engineering, model building, and prediction evaluation. Key performance indicators such as link bandwidth utilization and packet loss rate are collected, and multi-dimensional data quality control is implemented. A deep learning model combining traditional machine learning algorithms such as support vector machines and random forests with CNN-LSTM is used to conduct link quality evaluation, and a fault prediction model is constructed based on continuous-time Markov chains. Experimental results show that the fault detection accuracy of the proposed machine learning method reaches 92.3%, while the deep learning method further improves it to 95.8%. Compared with traditional methods, this significantly shortens the average fault location time, significantly increases the fault prediction lead time, and reduces the false alarm rate. The link quality assessment and fault prediction framework constructed in this study provides an efficient technical solution for communication network fault prevention and control, promoting the intelligent development of communication network operation and maintenance.

KEYWORDS

Machine learning; Communication network; Link quality evaluation; Fault prediction; Data preprocessing; Deep learning

CITE THIS PAPER

ZhiYi Tan. Link quality evaluation and fault prediction of communication networks based on machine learning. Journal of Computer Science and Electrical Engineering. 2026, 8(1): 64-68. DOI: https://doi.org/10.61784/jcsee3119.

REFERENCES

[1] Zhuang Zhiyong. A Brief Analysis of Fault Detection and Prediction of Communication Lines Based on Artificial Intelligence. China News Communications, 2024.

[2] Zhou Mingzhe, Feng Bailong, Mo Mingfei, et al. Application of Intelligent Communication Technology in Energy Internet. Electronic Technology, 2024.

[3] Li Ming, Guan Wei. Application of Digital Technology in Power Communication Optical Cable Resource Management. Integrated Circuit Applications, 2024.

[4]  Zhu Ruochong. User Behavior Analysis and Optimization Based on Machine Learning in Mobile Communication Core Network. Digital Communication World, 2024.

[5] Xie Xufeng, Chen Shangshang, Pan Pan, et al. Prediction and Diagnosis of Electrical Equipment Faults Based on Machine Learning Algorithms. Technology & Markets, 2024.

[6] Geng Kelei. Application Analysis of Intelligent Technology in Electrical Automation Control. Electrical Technology and Economics, 2024.

[7] Liu Jiangwei. Research on Fault Diagnosis and Prediction of Automated Mechanical System Based on Machine Learning. Home Appliance Maintenance, 2024.

[8] Song Shuang, Xing Jianping, Cai Huizhong. Research on Maintenance Prediction and Optimization of Electrical Equipment Based on Artificial Intelligence. Home Appliance Repair, 2024.

[9] Ni Wenqin, Ling Yizhan, Zhao Mengqi, et al. Discussion on Online Monitoring and Diagnosis Technology of Intelligent Electrical Equipment. Power Equipment Management, 2024.

[10] Han Congwei. Research on the Combination Strategy of Boiler Operation Status Monitoring and Automated Control Technology. Instrument and Meter User, 2024.

[11] Shi Yuwei, Ma Lifan. Application of Intelligent Technology in Power Equipment Operation and Maintenance. Electronic Technology, 2024.