Scholarlink

<p>Intelligent optimization techniques are needed in modern optical communication networks to improve signal security and efficiency in the face of rapid technological advancement. In order to forecast traffic flow in optical networks, this paper proposes a quantum machine learning framework that combines quantum support vector machine (QSVM) with term frequency-inverse document frequency (TF-IDF) feature extraction. For thorough traffic analysis, real-time network packets are recorded using the Wireshark network analyzer. With a 2-qubit configuration, the framework outperforms the 4-qubit and 8-qubit configurations, achieving root mean square error (RMSE) of 0.144, mean absolute error (MAE) of 0.380, and mean absolute percentage error (MAPE) of 0.256. In optical network traffic forecasting applications, the suggested QSVM framework greatly outperforms current classical methods with a prediction accuracy of 97.36%.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Quantum machine learning framework for traffic flow forecasting in optical networks

  • 期刊论文

  • 发表日期：

    2025年8月31日

查看全文
  • Jagdish Jangid,
  • Sachin Dixit,
  • Shubham Malhotra

摘要

Intelligent optimization techniques are needed in modern optical communication networks to improve signal security and efficiency in the face of rapid technological advancement. In order to forecast traffic flow in optical networks, this paper proposes a quantum machine learning framework that combines quantum support vector machine (QSVM) with term frequency-inverse document frequency (TF-IDF) feature extraction. For thorough traffic analysis, real-time network packets are recorded using the Wireshark network analyzer. With a 2-qubit configuration, the framework outperforms the 4-qubit and 8-qubit configurations, achieving root mean square error (RMSE) of 0.144, mean absolute error (MAE) of 0.380, and mean absolute percentage error (MAPE) of 0.256. In optical network traffic forecasting applications, the suggested QSVM framework greatly outperforms current classical methods with a prediction accuracy of 97.36%.