#ifndef _KMEANS_H_#define _KMEANS_H_#include <vector>#include "Index.h"#include "Cluster.h"#include "DocumentVector.h"class Kmeans {private:	int num_clusters;	std::vector<Cluster> clusters;	std::vector<DocumentVector> documents;	Index index;public:	Kmeans(Index index_, std::vector<DocumentVector> documents_, int k) : num_clusters(k), documents(documents_), index(index_) {}			std::vector<Cluster> run() 	{		// initialization		for(int i = 0; i < num_clusters; i++)			clusters.insert(clusters.end(), Cluster(documents[i], index));				//stop the algorithm after 5 iterations		for (int a = 0; a < 5; a++) 		{			for(int i = 0; i < num_clusters; i++) 				clusters[i] = Cluster(clusters[i].get_centroid(), index);					// put each document in the appropriate cluster			for(unsigned int i = 0; i < documents.size(); i++) 			{				int max_index = 0;				double max_similarity = 0;							for(int j = 0; j < num_clusters; j++) 				{					if (clusters[j].similarity(documents[i]) > max_similarity)					{						max_index = j;						max_similarity = clusters[j].similarity(documents[i]);					}					}				// cout << "max index: " << max_index << endl;							 	clusters[max_index].add_document(documents[i]);			}								for(int i = 0; i < num_clusters; i++)				clusters[i].update_centroid();					}				return clusters;	}};#endif /* _KMEANS_H_ */