FOR LEI LEI

#include <cassert>
#include <cstdio>
#include <cstring>
#include <cstdlib>

int num_cmp(const void * e1, const void * e2)
{
    return (*(int *)e1) - (*(int *)e2);
}

void GetMinMax(const char * input, char * char_tok, int *num_max, int *num_min)
{
    int nCount = 0;
    int nIndex = 0;
    int * num_input;

    int len = strlen(input);
    char *str_input = new char [len];
    char *str_input2 = new char [len];

    strcpy(str_input, input);
    strcpy(str_input2, input);

    char *pch;
    pch = strtok(str_input, char_tok);

    while (pch != NULL)
    {
        ++nCount;
        pch = strtok(NULL, char_tok);
    }


    num_input = new int[nCount];

    pch = strtok(str_input2, char_tok);
    while (pch != NULL)
    {
        *(num_input+nIndex) = atoi(pch);
        ++nIndex;

        pch = strtok(NULL, char_tok);
    }
    assert(nIndex == nCount);

    qsort(num_input, nCount, sizeof(int), num_cmp);

    *num_min = num_input[0];
    *num_max = num_input[nCount-1];

    delete [] num_input;

    delete [] str_input;
    delete [] str_input2;
}

int main()
{
    int num_max;
    int num_min;

    char input[] = "3;9;10;11;1;8;2;4";
    char *char_tok = ";";

    GetMinMax(input, char_tok, &num_max, &num_min);

    printf("min=%d, max=%d\n", num_min, num_max);

    return 0;
}
x=read.table("D:\\大二下\\多元统计分析\\shuju\\距离判别.txt",header = T) x class=factor(x[,1])#转化为因子型 x=x[,-1] g=length(levels(class))#类别数 L=ncol(x)#指标数 nx=nrow(x)#样品数 mu=matrix(0,nrow=g,ncol=L)#均值 s=list()#协方差 for (i in 1:g) { mu[i,]=colMeans(x[class==i,]) s[[i]]=cov(x[class==i,]) } shf=matrix(0,nrow=L,ncol=L) for (i in 1:length(s)) { n=length(class[class==i]) shf=shf+(n-1)*s[[i]] } sh=shf/(nx-g) D=matrix(0,nrow = nx,ncol=g)#马氏平方距离 for (i in 1:g) { for (j in 1:nx) { #D[j,i]=as.matrix(x[j,]-mu[i,])%*%solve(sh)%*%t(x[j,]-mu[i,]) D[j,i]=mahalanobis(as.matrix(x[j,]),mu[i,],sh) } } D x=c(8.06,231.03,14.41,5.72,6.15) x1=c(9.89,409.42,19.47,5.19,10.49) matrix(x,ncol=L) mahalanobis(matrix(x1,ncol=L),mu[1,],sh) #回代估计法 lei=c() for (i in 1:nx) { lei[i]=which.min(D[i,]) } lei for (i in 1:nx) { n[i]=ifelse(class[i]==lei[i],0,1) } p=sum(n)/nx#回代误判率 #交叉确认估计法 y=read.table("D:\\大二下\\多元统计分析\\shuju\\距离判别.txt",header = T) L=ncol(y[,-1])#指标数 nx=nrow(y)#样品数 lei=c() nn=c() for (k in 1:nx) { x=y[-k,] class=factor(x[,1]) g=length(levels(class))#类别数 x=x[,-1] nnx=nx-1 mu=matrix(0,nrow=g,ncol=L)#均值 s=list()#协方差 for (i in 1:g) { mu[i,]=colMeans(x[class==i,]) s[[i]]=cov(x[class==i,]) } shf=matrix(0,nrow=L,ncol=L) for (j in 1:length(s)) { n=length(class[class==j]) shf=shf+(n-1)*s[[j]] } sh=shf/(nnx-g) D=c()#剔除样品的马氏平方距离 for (m in 1:g) { #D[m]=as.matrix(y[k,-1]-mu[m,])%*%solve(sh)%*%t(y[k,-1]-mu[m,]) D[m]=mahalanobis(as.matrix(y[k,-1]),mu[m,],sh) } lei[k]=which.min(D)#剔除样本判断的所属类别 nn[k]=ifelse(y[k,1]==lei[k],0,1)#误判时n为1 } x[which(class!=lei)] p=sum(nn)/nx#交叉确认误判率 nn lei.如果假定各个总体的协方差不相等,又该如何修改距离判别的代码?
06-02
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值