负载培训数据

Load the Japanese Vowels data set as described in [1] and [2]. The zip filejapaneseVowels.zipcontains sequences of varying length. The sequences are divided into two folders,火车andTest，分别包含训练序列和测试序列。在这些文件夹中的每个文件夹中，序列都分为子文件夹，这些子文件夹从1to9。这names of these subfolders are the label names. A MAT file represents each sequence. Each sequence is a matrix with 12 rows, with one row for each feature, and a varying number of columns, with one column for each time step. The number of rows is the sequence dimension and the number of columns is the sequence length.

解压缩序列数据。

filename =“ japanyvowels.zip”; outputFolder = fullfile(tempdir,"japaneseVowels"）；unzip(filename,outputFolder);

对于培训预测指标，创建文件数据存储并指定读取功能为加载功能。这加载function, loads the data from the MAT-file into a structure array. To read files from the subfolders in the training folder, set the'IncludeSubfolders'option totrue。

foldertrain = fullfile（输出折叠器，“火车”）；fdsPredictorTrain = fileDatastore(folderTrain,。。。'ReadFcn',@load,。。。'IncludeSubfolders'，真的）;

Preview the datastore. The returned struct contains a single sequence from the first file.

preview(fdsPredictorTrain)

ans =带有字段的结构：X：[12×20双]

对于标签，创建一个文件数据存储，并指定读取函数为readLabel函数，定义在示例的末尾。这readLabelfunction extracts the label from the subfolder name.

classNames =字符串（1：9）;fdslabeltrain = filedatastore（foldertrain，。。。'ReadFcn'，@（filename）readlabel（文件名，classNames），。。。'IncludeSubfolders'，真的）;

Preview the datastore. The output corresponds to the label of the first file.

预览（fdslabeltrain）

ans =分类1

转换并组合数据存储

要将序列数据从预测变量数据存储到深度学习网络中输入序列数据，序列的迷你批次必须具有相同的长度。使用padSequencefunction, defined at the end of the datastore, that pads or truncates the sequences to have length 20.

sequenceLength = 20; tdsTrain = transform(fdsPredictorTrain,@(data) padSequence(data,sequenceLength));

Preview the transformed datastore. The output corresponds to the padded sequence from the first file.

X = preview(tdsTrain)

X =1×1 cell array{12×20 double}

要将两个数据库的预测变量和标签输入到深度学习网络中，请使用combine功能。

cdstrain = combine（tdstrain，fdslabeltrain）;

Preview the combined datastore. The datastore returns a 1-by-2 cell array. The first element corresponds to the predictors. The second element corresponds to the label.

preview(cdsTrain)

ans =1×2单元格数组{12×20 double} {[1]}

Define LSTM Network Architecture

定义LSTM网络体系结构。将输入数据的功能数量指定为输入大小。指定具有100个隐藏单元的LSTM层，并输出序列的最后一个元素。最后，指定一个完全连接的层，其输出大小等于类的数量，然后是软磁层和分类层。

numFeatures = 12; numClasses = numel(classNames); numHiddenUnits = 100; layers = [。。。sequenceInputLayer(numFeatures) lstmLayer(numHiddenUnits,'outputmode','last'）完整连接的layerer（numClasses）softmaxlayer classificationlayer];

Specify the training options. Set the solver to'adam'and“梯度阈值”to 2. Set the mini-batch size to 27 and set the maximum number of epochs to 75. The datastores do not support shuffling, so set“洗牌”to'never'。

由于迷你批次很小，序列短，因此CPU更适合训练。放'ExecutionEnvironment'to'cpu'。在GPU上训练，如果有的话，请设置'ExecutionEnvironment'to'auto'(the default value).

minibatchsize = 27;选项=训练（'adam',。。。'ExecutionEnvironment','cpu',。。。'MaxEpochs'，75，。。。“ MINIBATCHSIZE”，minibatchsize，。。。“梯度阈值”，2，。。。“洗牌”,'never',。。。“冗长”，0，。。。'Plots','training-progress'）；

火车the LSTM network with the specified training options.

net = trainnetwork（CDSTRAIN，层，选项）;

Test the Network

创建一个转换包含了数据存储-out test data using the same steps as for the training data.

foldertest = fullfile（outputFolder，"Test"）；fdspredictEttest = filedatastore（foldertest，。。。'ReadFcn',@load,。。。'IncludeSubfolders'，真的）;tdsTest = transform(fdsPredictorTest,@(data) padSequence(data,sequenceLength));

Make predictions on the test data using the trained network.

YPred = classify(net,tdsTest,“ MINIBATCHSIZE”，minibatchSize）;

Calculate the classification accuracy on the test data. To get the labels of the test set, create a file datastore with the read functionreadLabel并指定包括子文件夹。指定输出是通过设置的垂直限制的“统一”option totrue。

fdsLabelTest = fileDatastore(folderTest,。。。'ReadFcn'，@（filename）readlabel（文件名，classNames），。。。'IncludeSubfolders',true,。。。“统一”，真的）;ytest = readall（fdslabeltest）;

accuracy = mean(YPred == YTest)

精度= 0.9351

Functions

这readLabel功能从指定的文件名中提取标签classNames。

functionlabel = readLabel(filename,classNames) filepath = fileparts(filename); [~,label] = fileparts(filepath); label = categorical(string(label),classNames);结尾

这padSequencefunction pads or truncates the sequence indata.xto have the specified sequence length and returns the result in a 1-by-1 cell.

function序列= padSequence(数据、sequenceLength)顺ce = data.X; [C,S] = size(sequence);如果S < sequenceLength padding = zeros(C,sequenceLength-S); sequence = [sequence padding];别的sequence = sequence(:,1:sequenceLength);结尾sequence = {sequence};结尾