Knn算法C语言实例

K近邻(K-Nearest Neighbors, KNN)是一种基于实例的学习方法，用于分类和回归分析。在C语言中实现KNN算法，首先需要准备数据结构存储样本点和它们对应的类别，然后计算新样本点与训练集中每个样本的距离，并找到距离最近的k个邻居。

以下是一个简单的KNN C语言示例，使用欧式距离（对于连续特征）：

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

// 计算两点之间的欧氏距离
double euclidean_distance(double* p1, double* p2, int n) {
    double distance = 0;
    for (int i = 0; i < n; i++) {
        distance += pow(p1[i] - p2[i], 2);
    }
    return sqrt(distance);
}

// 获取邻居
void get_neighbors(double** train_data, int* labels, int test_index, int k, double** neighbors, int* neighbor_labels, int data_size) {
    for (int i = 0; i < data_size; i++) {
        if (i != test_index) {
            double* temp = (double*)malloc(sizeof(double) * data_size);
            memcpy(temp, train_data[i], sizeof(double) * data_size);
            double dist = euclidean_distance(train_data[test_index], temp, data_size);
            // 将当前样本添加到邻居列表
            neighbors[i] = temp;
            neighbor_labels[i] = labels[i];
            free(temp);
        }
    }
    // 对邻居按距离排序并选择前k个
    qsort(neighbors, data_size, sizeof(double*), compare_distances);
    int count = 0;
    while (count < k && neighbor_labels[count]) {
        count++;
    }
    *neighbor_labels = neighbor_labels[0];
}

// 比较距离函数，用于qsort
int compare_distances(const void* a, const void* b) {
    double* pa = (double*)a;
    double* pb = (double*)b;
    return (pa[0] > pb[0]);
}

int main() {
    // 简单示例，实际应用中会包含更多细节
    int data_size = 5; // 假设有5个样本点
    double** train_data = ...; // 存储训练数据
    int* train_labels = ...; // 存储对应标签
    int test_index = ...; // 测试样本索引
    int k = 3; // 设置k值

    // 实例化邻居列表
    double** neighbors = malloc(data_size * sizeof(double*));
    int* neighbor_labels = malloc(data_size * sizeof(int));

    get_neighbors(train_data, train_labels, test_index, k, neighbors, neighbor_labels, data_size);

    printf("测试样本的预测类别: %d\n", neighbor_labels[0]);
    // 更多清理工作...
    return 0;
}