【chatgpt代码系列】DBSCAN聚类算法

以下是简单的C++实现DBSCAN聚类算法的示例代码，供参考：

#include <iostream>
#include <vector>
#include <cmath>

double getDistance(std::vector<double> &p, std::vector<double> &q){
    double distance = 0.0;
    for (int i = 0; i < p.size(); i++){
        distance += pow(p[i] - q[i], 2);
    }
    return sqrt(distance);
}

std::vector<int> rangeQuery(std::vector<std::vector<double>> &dataset, int point, double eps){
    std::vector<int> neighborIndexes;
    for (int i = 0; i < dataset.size(); i++){
        if (getDistance(dataset[point], dataset[i]) <= eps){
            neighborIndexes.push_back(i);
        }
    }
    return neighborIndexes;
}

std::vector<std::vector<int>> DBSCAN(std::vector<std::vector<double>> &dataset, double eps, int minPoints){
    std::vector<std::vector<int>> clusters;
    std::vector<bool> visited(dataset.size(), false);
    std::vector<int> noise;

    for (int i = 0; i < dataset.size(); i++){
        if (!visited[i]){
            visited[i] = true;
            std::vector<int> neighbors = rangeQuery(dataset, i, eps);
            if (neighbors.size() < minPoints){
                noise.push_back(i);
            }else{
                std::vector<int> cluster;
                cluster.push_back(i);
                for (int j = 0; j < neighbors.size(); j++){
                    if (!visited[neighbors[j]]){
                        visited[neighbors[j]] = true;
                        std::vector<int> newNeighbors = rangeQuery(dataset, neighbors[j], eps);
                        if (newNeighbors.size() >= minPoints){
                            for (int k = 0; k < newNeighbors.size(); k++){
                                neighbors.push_back(newNeighbors[k]);
                            }
                        }
                    }
                    bool is_cluster_point = false;
                    for (int k = 0; k < cluster.size(); k++){
                        if (cluster[k] == neighbors[j]){
                            is_cluster_point = true;
                            break;
                        }
                    }
                    if (!is_cluster_point){
                        cluster.push_back(neighbors[j]);
                    }
                }
                clusters.push_back(cluster);
            }
        }
    }
    return clusters;
}

int main()
{
    std::vector<std::vector<double>> dataset = {{0.0, 0.0}, {1.0, 0.0}, {2.0, 0.0}, {0.0, 1.0}, {1.0, 1.0}, {5.0,5.0}};
    double eps = 1.0;
    int minPoints = 2;
    std::vector<std::vector<int>> clusters = DBSCAN(dataset, eps, minPoints);

    for (int i = 0; i < clusters.size(); i++){
        std::cout << "Cluster " << i << ": ";
        for (int j = 0; j < clusters[i].size(); j++){
            std::cout << clusters[i][j] << ", ";
        }
        std::cout << "\n";
    }

    return 0;
}

这个示例代码使用输入数据dataset、epsilon(eps)和minPoints作为参数，并返回聚类。其中，输入数据dataset是一个向量集合，每个向量表示一个n维数据点。

运行示例后，输出如下：

Cluster 0: 0, 1, 2, 
Cluster 1: 3, 4, 
Cluster 2: 5,

这些结果表明， DBSCAN算法成功地将坐标值小于1.0的数据点划分到簇0中，并将另外三个点分别分配给簇1和簇2。