Main Content

本页翻译不是最新的。点击此处可查看最新英文版本。

神经网络

用于二类和多类分类的神经网络

神经网络模型由一系列反映大脑处理信息方式的层构成。Statistics and Machine Learning Toolbox™ 中提供的神经网络分类器是完全连接的前馈神经网络,您可以调整全连接层的大小并更改层的激活函数。

要训练神经网络分类模型,请使用分类学习器。为了获得更大的灵活性,请在命令行界面中使用 fitcnet 来训练神经网络分类器。经过训练后,您可以通过将模型和新预测变量数据传递给 predict 来对新数据进行分类。

如果您要创建更复杂的深度学习网络并拥有 Deep Learning Toolbox™,您可以尝试深度网络设计器 (Deep Learning Toolbox)

App

分类学习器使用有监督的机器学习训练模型以对数据进行分类

模块

ClassificationNeuralNetwork PredictClassify observations using neural network classification model (自 R2021b 起)

函数

全部展开

fitcnetTrain neural network classification model (自 R2021a 起)
compactReduce size of machine learning model
limeLocal interpretable model-agnostic explanations (LIME) (自 R2020b 起)
partialDependenceCompute partial dependence (自 R2020b 起)
plotPartialDependenceCreate partial dependence plot (PDP) and individual conditional expectation (ICE) plots
shapleyShapley values (自 R2021a 起)
crossvalCross-validate machine learning model
kfoldLossClassification loss for cross-validated classification model
kfoldPredictClassify observations in cross-validated classification model
kfoldEdgeClassification edge for cross-validated classification model
kfoldMarginClassification margins for cross-validated classification model
kfoldfunCross-validate function for classification
lossClassification loss for neural network classifier (自 R2021a 起)
resubLossResubstitution classification loss
edgeClassification edge for neural network classifier (自 R2021a 起)
marginClassification margins for neural network classifier (自 R2021a 起)
resubEdgeResubstitution classification edge
resubMarginResubstitution classification margin
predictClassify observations using neural network classifier (自 R2021a 起)
resubPredictClassify training data using trained classifier

对象

ClassificationNeuralNetworkNeural network model for classification (自 R2021a 起)
CompactClassificationNeuralNetworkCompact neural network model for classification (自 R2021a 起)
ClassificationPartitionedModelCross-validated classification model

主题