摘要
正则化弹性网络是一种强大的深度学习模型,结合线性回归和逻辑回归的特点,可以同时进行特征选择和参数控制,避免了传统正则化的部分局限性。离散傅里叶变换特征提取是一种常用的信号处理方法,可以提取信号中的特定频率的特征,在众多领域都有广泛的应用。通过弹性网络正则化和加窗离散傅里叶变换的信号分析技术结合,进行了相应的研究和应用。以凯斯西储大学故障轴承振动数据为例,进行信号分析处理,再经过神经网络模型的学习和预测,从而得到了一个准确率较高的弹性网络模型。其方法对于众多复杂的问题都有着重要的研究价值。
A regularized elastic net is a powerful deep learning model,which can combine the characteristics of linear regression and logistic regression to perform both feature selection and parameter control,avoiding some of the limitations of traditional regularization.Discrete Fourier transform feature extraction is a commonly-used signal processing method that can extract specific frequency features from signals,and it has a wide range of applications in many fields.This article combines the signal analysis technology of elastic net regularization and windowed discrete Fourier transform to conduct corresponding research and applications.This article takes the vibration data of faulty bearings from Case Western Reserve University as an example,performs signal analysis and processing,and then obtains an elastic net model with high accuracy through the learning and predicting of the neural network model.This method is of great research value for many complex problems.
作者
王尉旭
WANG Weixu(Chongqing Jiaotong University,Chongqing,400074 China)
出处
《科技资讯》
2024年第7期32-35,共4页
Science & Technology Information