首页 文章

使用预先计算的chi2内核与libsvm（matlab）时的错误结果

提问于
浏览 次
6

我正在尝试使用libsvm，并按照示例在软件附带的heart_scale数据上训练svm . 我想使用我自己预先计算的chi2内核 . 培训数据的分类率降至24％ . 我确信我正确地计算了内核，但我想我一定是做错了 . 代码如下 . 你能看到任何错误吗？非常感谢帮助 .

%read in the data:
[heart_scale_label, heart_scale_inst] = libsvmread('heart_scale');
train_data = heart_scale_inst(1:150,:);
train_label = heart_scale_label(1:150,:);

%read somewhere that the kernel should not be sparse
ttrain = full(train_data)';
ttest = full(test_data)';

precKernel = chi2_custom(ttrain', ttrain');
model_precomputed = svmtrain2(train_label, [(1:150)', precKernel], '-t 4');

这是内核预先计算的方式：

function res=chi2_custom(x,y)
a=size(x);
b=size(y);
res = zeros(a(1,1), b(1,1));
for i=1:a(1,1)
    for j=1:b(1,1)
        resHelper = chi2_ireneHelper(x(i,:), y(j,:));
        res(i,j) = resHelper;
    end
end
function resHelper = chi2_ireneHelper(x,y)
a=(x-y).^2;
b=(x+y);
resHelper = sum(a./(b + eps));

使用不同的svm实现（vlfeat），我获得了训练数据的分类率（是的，我在训练数据上测试，只是为了看看发生了什么）大约90％ . 所以我很确定libsvm结果是错误的 .

matlab machine-learning svm libsvm

2 回答

  • 15

    使用支持向量机时，将数据集规范化为预处理步骤非常重要 . Normalization 将属性放在相同的比例上，并防止具有大值的属性偏向结果 . 它还提高了数值稳定性（最小化浮点表示引起的溢出和下溢的可能性） .

    同样准确地说，你对卡方内核的计算略有偏差 . 而是采用下面的定义，并使用这个更快的实现：

    chi_squared_kernel

    function D = chi2Kernel(X,Y)
    D = zeros(size(X,1),size(Y,1));
    for i=1:size(Y,1)
        d = bsxfun(@minus, X, Y(i,:));
        s = bsxfun(@plus, X, Y(i,:));
        D(:,i) = sum(d.^2 ./ (s/2+eps), 2);
    end
    D = 1 - D;
end

    现在考虑使用与您相同的数据集的以下示例（代码改编自我的previous answer）：

    %# read dataset
[label,data] = libsvmread('./heart_scale');
data = full(data);      %# sparse to full

%# normalize data to [0,1] range
mn = min(data,[],1); mx = max(data,[],1);
data = bsxfun(@rdivide, bsxfun(@minus, data, mn), mx-mn);

%# split into train/test datasets
trainData = data(1:150,:);    testData = data(151:270,:);
trainLabel = label(1:150,:);  testLabel = label(151:270,:);
numTrain = size(trainData,1); numTest = size(testData,1);

%# compute kernel matrices between every pairs of (train,train) and
%# (test,train) instances and include sample serial number as first column
K =  [ (1:numTrain)' , chi2Kernel(trainData,trainData) ];
KK = [ (1:numTest)'  , chi2Kernel(testData,trainData)  ];

%# view 'train vs. train' kernel matrix
figure, imagesc(K(:,2:end))
colormap(pink), colorbar

%# train model
model = svmtrain(trainLabel, K, '-t 4');

%# test on testing data
[predTestLabel, acc, decVals] = svmpredict(testLabel, KK, model);
cmTest = confusionmat(testLabel,predTestLabel)

%# test on training data
[predTrainLabel, acc, decVals] = svmpredict(trainLabel, K, model);
cmTrain = confusionmat(trainLabel,predTrainLabel)

    测试数据的结果：

    Accuracy = 84.1667% (101/120) (classification)
cmTest =
    62     8
    11    39

    根据您的预期，我们可以获得大约90％的准确度：

    Accuracy = 92.6667% (139/150) (classification)
cmTrain =
    77     3
     8    62

    train_train_kernel_matrix

    回复于
  • 0

    问题是以下几行：

    resHelper = sum(a./(b + eps));

    它应该是：

    resHelper = 1-sum(2*a./(b + eps));
    回复于

相关问题