KNN的C语言代码，要能够读取数据文件

您好，我理解您的问题。以下是一个简单的KNN算法的C语言实现，它可以从数据文件中读取训练数据和测试数据并对其进行分类：

#include <stdio.h>
#include <stdlib.h>
#include <math.h>

#define K 5 // 设置K值为5

typedef struct {
    int label; // 样本标签
    double* features; // 样本特征
} Sample;

int numFeatures; // 样本特征数
int numSamples; // 样本总数

double euclideanDistance(double* f1, double* f2) {
    double dist = 0.0;
    for (int i = 0; i < numFeatures; i++)
        dist += pow(f1[i] - f2[i], 2);
    return sqrt(dist);
}

int classify(Sample* samples, double* testFeatures) {
    double distances[numSamples];
    for (int i = 0; i < numSamples; i++)
        distances[i] = euclideanDistance(samples[i].features, testFeatures);
    int indexes[numSamples];
    for (int i = 0; i < numSamples; i++)
        indexes[i] = i;
    for (int i = 1; i < numSamples; i++) {
        int keyIndex = indexes[i];
        double key = distances[i];
        int j = i - 1;
        while (j >= 0 && distances[j] > key) {
            distances[j + 1] = distances[j];
            indexes[j + 1] = indexes[j];
            j--;
        }
        distances[j + 1] = key;
        indexes[j + 1] = keyIndex;
    }
    int freq[numSamples];
    for (int i = 0; i < numSamples; i++)
        freq[i] = 0;
    for (int i = 0; i < K; i++)
        freq[samples[indexes[i]].label]++;
    int maxFreqIndex = 0;
    for (int i = 1; i < numSamples; i++)
        if (freq[maxFreqIndex] < freq[i])
            maxFreqIndex = i;
    return samples[indexes[maxFreqIndex]].label;
}

int main(void) {
    char trainFilename[] = "train.csv"; // 训练数据文件
    char testFilename[] = "test.csv"; // 测试数据文件

    // 读取训练数据
    FILE* trainFile = fopen(trainFilename, "r");
    fscanf(trainFile, "%d,%d", &numSamples, &numFeatures);
    Sample* trainSamples = (Sample*)malloc(numSamples * sizeof(Sample));
    for (int i = 0; i < numSamples; i++) {
        trainSamples[i].features = (double*)malloc(numFeatures * sizeof(double));
        for (int j = 0; j < numFeatures; j++)
            fscanf(trainFile, "%lf", &trainSamples[i].features[j]);
        fscanf(trainFile, "%d", &trainSamples[i].label);
    }
    fclose(trainFile);

    // 读取测试数据
    FILE* testFile = fopen(testFilename, "r");
    int numTestSamples;
    fscanf(testFile, "%d,%d", &numTestSamples, &numFeatures);
    double testFeatures[numFeatures];
    for (int i = 0; i < numTestSamples; i++) {
        for (int j = 0; j < numFeatures; j++)
            fscanf(testFile, "%lf", &testFeatures[j]);
        int predictedLabel = classify(trainSamples, testFeatures);
        printf("%d\n", predictedLabel);
    }
    fclose(testFile);

    // 释放内存
    for (int i = 0; i < numSamples; i++)
        free(trainSamples[i].features);
    free(trainSamples);

    return 0;
}

请注意，这只是一个简单的实现，还有很多可以改进的地方。另外，这个实现只能处理数值特征，如果您的数据集包含非数值特征，则需要进行预处理。