abalone.m

使用高斯模型对威斯康辛州大学医学院长期乳腺癌数据进行了贝叶斯模式识别。识别率为95以上

clear;
load data.mat

Total  = length(data(:,1));
VaribleNumber = length(data(1,:));
ClassNumber = 12;
SexNumber = 3;
Index = [1:Total];


[TestMatrix,TrainMatrix,TestIndex,TrainIndex] = SplitMatrixRandom(data);

for swap = 1:2
    if swap == 2
       tmpIndex = TrainIndex;
       TrainIndex = TestIndex;
       TestIndex = tmpIndex;
       tmpIndex = TrainMatrix;
       TrainMatrix = TestMatrix;
       TestMatrix = tmpIndex;
    end
    TrainNumber = length(TrainIndex);
    TestNumber = Total - TrainNumber;
    TrainAnswer = TrainMatrix(:,VaribleNumber);
    TrainClassIndex_1 = find((TrainAnswer-5).*(TrainAnswer-1) <= 0);
    TrainClassIndex_2 = find(TrainAnswer == 6);
    TrainClassIndex_3 = find(TrainAnswer == 7);
    TrainClassIndex_4 = find(TrainAnswer == 8);
    TrainClassIndex_5 = find(TrainAnswer == 9);
    TrainClassIndex_6 = find(TrainAnswer == 10);
    TrainClassIndex_7 = find(TrainAnswer == 11);
    TrainClassIndex_8 = find(TrainAnswer == 12);
    TrainClassIndex_9 = find(TrainAnswer == 13);
    TrainClassIndex_10 = find(TrainAnswer == 14);
    TrainClassIndex_11 = find((TrainAnswer-19).*(TrainAnswer-15) <= 0);
    TrainClassIndex_12 = find(TrainAnswer >= 20);

    Ppostsex = zeros(ClassNumber,3);
    Ppostsex(1,1) = sum(TrainMatrix(TrainClassIndex_1,1) == 0)/length(TrainClassIndex_1);
    Ppostsex(1,2) = sum(TrainMatrix(TrainClassIndex_1,1) == 1)/length(TrainClassIndex_1);
    Ppostsex(1,3) = sum(TrainMatrix(TrainClassIndex_1,1) == 2)/length(TrainClassIndex_1);
    Ppostsex(2,1) = sum(TrainMatrix(TrainClassIndex_2,1) == 0)/length(TrainClassIndex_2);
    Ppostsex(2,2) = sum(TrainMatrix(TrainClassIndex_2,1) == 1)/length(TrainClassIndex_2);
    Ppostsex(2,3) = sum(TrainMatrix(TrainClassIndex_2,1) == 2)/length(TrainClassIndex_2);
    Ppostsex(3,1) = sum(TrainMatrix(TrainClassIndex_3,1) == 0)/length(TrainClassIndex_3);
    Ppostsex(3,2) = sum(TrainMatrix(TrainClassIndex_3,1) == 1)/length(TrainClassIndex_3);
    Ppostsex(3,3) = sum(TrainMatrix(TrainClassIndex_3,1) == 2)/length(TrainClassIndex_3);
    Ppostsex(4,1) = sum(TrainMatrix(TrainClassIndex_4,1) == 0)/length(TrainClassIndex_4);
    Ppostsex(4,2) = sum(TrainMatrix(TrainClassIndex_4,1) == 1)/length(TrainClassIndex_4);
    Ppostsex(4,3) = sum(TrainMatrix(TrainClassIndex_4,1) == 2)/length(TrainClassIndex_4);
    Ppostsex(5,1) = sum(TrainMatrix(TrainClassIndex_5,1) == 0)/length(TrainClassIndex_5);
    Ppostsex(5,2) = sum(TrainMatrix(TrainClassIndex_5,1) == 1)/length(TrainClassIndex_5);
    Ppostsex(5,3) = sum(TrainMatrix(TrainClassIndex_5,1) == 2)/length(TrainClassIndex_5);
    Ppostsex(6,1) = sum(TrainMatrix(TrainClassIndex_6,1) == 0)/length(TrainClassIndex_6);
    Ppostsex(6,2) = sum(TrainMatrix(TrainClassIndex_6,1) == 1)/length(TrainClassIndex_6);
    Ppostsex(6,3) = sum(TrainMatrix(TrainClassIndex_6,1) == 2)/length(TrainClassIndex_6);
    Ppostsex(7,1) = sum(TrainMatrix(TrainClassIndex_7,1) == 0)/length(TrainClassIndex_7);
    Ppostsex(7,2) = sum(TrainMatrix(TrainClassIndex_7,1) == 1)/length(TrainClassIndex_7);
    Ppostsex(7,3) = sum(TrainMatrix(TrainClassIndex_7,1) == 2)/length(TrainClassIndex_7);
    Ppostsex(8,1) = sum(TrainMatrix(TrainClassIndex_8,1) == 0)/length(TrainClassIndex_8);
    Ppostsex(8,2) = sum(TrainMatrix(TrainClassIndex_8,1) == 1)/length(TrainClassIndex_8);
    Ppostsex(8,3) = sum(TrainMatrix(TrainClassIndex_8,1) == 2)/length(TrainClassIndex_8);
    Ppostsex(9,1) = sum(TrainMatrix(TrainClassIndex_9,1) == 0)/length(TrainClassIndex_9);
    Ppostsex(9,2) = sum(TrainMatrix(TrainClassIndex_9,1) == 1)/length(TrainClassIndex_9);
    Ppostsex(9,3) = sum(TrainMatrix(TrainClassIndex_9,1) == 2)/length(TrainClassIndex_9);
    Ppostsex(10,1) = sum(TrainMatrix(TrainClassIndex_10,1) == 0)/length(TrainClassIndex_10);
    Ppostsex(10,2) = sum(TrainMatrix(TrainClassIndex_10,1) == 1)/length(TrainClassIndex_10);
    Ppostsex(10,3) = sum(TrainMatrix(TrainClassIndex_10,1) == 2)/length(TrainClassIndex_10);
    Ppostsex(11,1) = sum(TrainMatrix(TrainClassIndex_11,1) == 0)/length(TrainClassIndex_11);
    Ppostsex(11,2) = sum(TrainMatrix(TrainClassIndex_11,1) == 1)/length(TrainClassIndex_11);
    Ppostsex(11,3) = sum(TrainMatrix(TrainClassIndex_11,1) == 2)/length(TrainClassIndex_11);
    Ppostsex(12,1) = sum(TrainMatrix(TrainClassIndex_12,1) == 0)/length(TrainClassIndex_12);
    Ppostsex(12,2) = sum(TrainMatrix(TrainClassIndex_12,1) == 1)/length(TrainClassIndex_12);
    Ppostsex(12,3) = sum(TrainMatrix(TrainClassIndex_12,1) == 2)/length(TrainClassIndex_12);

    Pw = zeros(ClassNumber,1);
    Pw(1,1) = length(TrainClassIndex_1)/length(TrainIndex);
    Pw(2,1) = length(TrainClassIndex_2)/length(TrainIndex);
    Pw(12,1) = length(TrainClassIndex_12)/length(TrainIndex);
    Pw(3,1) = length(TrainClassIndex_3)/length(TrainIndex);
    Pw(4,1) = length(TrainClassIndex_4)/length(TrainIndex);
    Pw(5,1) = length(TrainClassIndex_5)/length(TrainIndex);
    Pw(6,1) = length(TrainClassIndex_6)/length(TrainIndex);
    Pw(7,1) = length(TrainClassIndex_7)/length(TrainIndex);
    Pw(8,1) = length(TrainClassIndex_8)/length(TrainIndex);
    Pw(9,1) = length(TrainClassIndex_9)/length(TrainIndex);
    Pw(10,1) = length(TrainClassIndex_10)/length(TrainIndex);
    Pw(11,1) = length(TrainClassIndex_11)/length(TrainIndex);

    d = 7;
    section = [2 3 4 5 6 7 8];
    MeanMatrix_M = zeros(ClassNumber,d);
    MeanMatrix_F = zeros(ClassNumber,d);
    MeanMatrix_I = zeros(ClassNumber,d);
    ConvarianceMatrix_M = zeros(d,d,ClassNumber);
    ConvarianceMatrix_F = zeros(d,d,ClassNumber);
    ConvarianceMatrix_I = zeros(d,d,ClassNumber);

    FailGroup = ones(ClassNumber,3);
    SampleNumber = 8;

    tmp = TrainMatrix(TrainClassIndex_1,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(1,1) = 0;
    end
    MeanMatrix_M(1,:) = mean(sextmp);
    ConvarianceMatrix_M(:,:,1) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 1),section);
    if length(sextmp) < SampleNumber
        FailGroup(1,2) = 0;
    end
    MeanMatrix_F(1,:) = mean(sextmp);
    ConvarianceMatrix_F(:,:,1) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 2),section);
    if length(sextmp) < SampleNumber
        FailGroup(1,3) = 0;
    end
    MeanMatrix_I(1,:) = mean(sextmp);
    ConvarianceMatrix_I(:,:,1) = cov(sextmp);

    tmp = TrainMatrix(TrainClassIndex_2,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(2,1) = 0;
    end
    MeanMatrix_M(2,:) = mean(sextmp);
    ConvarianceMatrix_M(:,:,2) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 1),section);
    if length(sextmp) < SampleNumber
        FailGroup(2,2) = 0;
    end
    MeanMatrix_F(2,:) = mean(sextmp);
    ConvarianceMatrix_F(:,:,2) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 2),section);
    if length(sextmp) < SampleNumber
        FailGroup(2,3) = 0;
    end
    MeanMatrix_I(2,:) = mean(sextmp);
    ConvarianceMatrix_I(:,:,2) = cov(sextmp); 

    tmp = TrainMatrix(TrainClassIndex_3,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(3,1) = 0;
    end
    MeanMatrix_M(3,:) = mean(sextmp);
    ConvarianceMatrix_M(:,:,3) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 1),section);
    if length(sextmp) < SampleNumber
        FailGroup(3,2) = 0;
    end
    MeanMatrix_F(3,:) = mean(sextmp);
    ConvarianceMatrix_F(:,:,3) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 2),section);
    if length(sextmp) < SampleNumber
        FailGroup(3,3) = 0;
    end
    MeanMatrix_I(3,:) = mean(sextmp);
    ConvarianceMatrix_I(:,:,3) = cov(sextmp);

    tmp = TrainMatrix(TrainClassIndex_4,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(4,1) = 0;
    end
    MeanMatrix_M(4,:) = mean(sextmp);
    ConvarianceMatrix_M(:,:,4) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 1),section);
    if length(sextmp) < SampleNumber
        FailGroup(4,2) = 0;
    end
    MeanMatrix_F(4,:) = mean(sextmp);
    ConvarianceMatrix_F(:,:,4) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 2),section);
    if length(sextmp) < SampleNumber
        FailGroup(4,3) = 0;
    end
    MeanMatrix_I(4,:) = mean(sextmp);
    ConvarianceMatrix_I(:,:,4) = cov(sextmp);

    tmp = TrainMatrix(TrainClassIndex_5,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(5,1) = 0;
    end
    MeanMatrix_M(5,:) = mean(sextmp);
    ConvarianceMatrix_M(:,:,5) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 1),section);
    if length(sextmp) < SampleNumber
        FailGroup(5,2) = 0;
    end
    MeanMatrix_F(5,:) = mean(sextmp);
    ConvarianceMatrix_F(:,:,5) = cov(sextmp);
    sextmp = tmp(find(tmp(:,1) == 2),section);
    if length(sextmp) < SampleNumber
        FailGroup(5,3) = 0;
    end
    MeanMatrix_I(5,:) = mean(sextmp);
    ConvarianceMatrix_I(:,:,5) = cov(sextmp);

    tmp = TrainMatrix(TrainClassIndex_6,1:end-1);
    sextmp = tmp(find(tmp(:,1) == 0),section);
    if length(sextmp) < SampleNumber
        FailGroup(6,1) = 0;
    end