datanormalitation.m

决策树

%归一化 

datarow=xlsread('1.xls','血常规','i2:n994');
save datarow;
%min-max normalization

data_min_max=zeros(993,6);
for i=1:6
    minA=min(datarow(:,i));
    maxA=max(datarow(:,i));
    for j=1:993
        data_min_max(j,i)=(datarow(j,i)-minA)/(maxA-minA);
    end
end
save data_min_max;

%z-score normalization
data_z_score=zeros(993,6);
for i=i:6
    meanA=mean(datarow(:,i));
    sum=0;
    for j=1:993
        sum=sum+(datarow(j,i)-meanA)^2;
        meanB=sum/(993-1);
        xigmad=sqrt(meanB);
        data_z_score(j,i)=(datarow(j,i)-meanA)/xigmad;
    end
end
save data_z_score

%narmalization by decimal scaling 
data_decimal=zeros(993,6);
for i=1:6
    maxA=max(datarow(:,i));
    if maxA/10<1
        for j=1:993
            data_decimal(j,i)=datarow(j,i)/10;
        end
    else if maxA/100<1
            for j=1:993
               data_decimal(j,i)=datarow(j,i)/100;
            end
        else
            for j=1:993
                data_decimal(j,i)=datarow(j,i)/1000;
            end
        end
    end
end
save  data_decimal;