摘要
为了提高光纤通信系统中的光功率预测精度,提出了一种基于机器学习算法的光功率预测模型。首先采用小波分析对光功率数据进行预处理得到不同分量,然后采用自回归移动平均与相关向量机对各分量分别进行建模,最后采用小波分析对预测结果进行组合,并采用具体光功率数据对模型性能进行测试,结果表明,相对于经典光功率预测模型,本文模型可以对光功率预测实现准确预测,提高了光功率预测精度,预测结果可以满足光功率预测的实际应用要求。
In order to improve the prediction accuracy of optical power in optical fiber communication systems,this paper puts forward an optical power forecasting model based on machine learning algorithm. Firstly,wavelet analysis is used to process the optical power data to get different components,and then the autoregressive moving average and the relevance vector machine are used to model of each component,finally wavelet analysis is used to combination the prediction result,and the performance of model is test by specific optical power data. The results showed that,relative to the current classic forecasting model,the proposed model can achieve accurate p forecast the optical power,improve the prediction accuracy and the forecasting results can satisfy the practical application requirement of optical power forecasting.
出处
《激光杂志》
北大核心
2015年第7期131-134,共4页
Laser Journal
基金
山东省高等学校科技计划项目(J13LN77)
关键词
光功率预测
小波分析
相关向量机
回归移动平均
optical power forecasting
wavelet analysis
relevance vector machine regression moving average
