- Code: Select all
#include <iostream>
//#include <conio.h>
#include <math.h>
#define row 10
#define col 2
using namespace std;
// int row=10,col=2;
int d [row][col];
int dataset[row][col]={1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20},g[row][col],g1[row][col],centroid1[1][2],centroid2[1][2];
//***************Method for calculating Eucledian distances***************//
void dist()
{
int x,y,i;
int v;
for (i=0;i<row;i++)
{
x=dataset[i][0]-centroid1[0][0];//Eucledian distance at x from centroid1
y=dataset[i][1]-centroid1[0][1];//Eucledian distance at y from centroid1
x=x*x;//squaring Eucledian distance at x
y=y*y;//squaring Eucledian distance at y
v=(int)(x+y);//sum of squared distance
d[i][0]=(int)sqrt(v);//mean squaring of distance
x=dataset[i][0]-centroid2[0][0];//Eucledian distance at x from centroid2
y=dataset[i][1]-centroid2[0][1];//Eucledian distance at y from centroid2
x=x*x;//squaring Eucledian distance at x
y=y*y;//squaring Eucledian distance at y
v=(int)(x+y);//sum of squared distance
d[i][1]=(int)sqrt(v);//mean squaring of distance
}
}
//***********************************************************************//
//*****************Method to perform k-means clustering*****************//
void clust()
{
int i;
for(i=0;i<row;i++)
{
g[i][0]=g[i][1]=0;
if(d[i][0]<=d[i][1])
g[i][0]=1;
else
g[i][1]=1;
}
}
void copy()
{
for(int i=0;i<row;i++)
for(int j=0;j<col;j++)
g1[i][j]=g[i][j];
}
int main(void)
{
int x,y,s,ch,i,j;
ch=x=y=s=0;
// clrscr();
std::cout<<"\n****## Program for K-Means Clustering ##****";
// cout<<"\n\nEnter values : ";
//********************Reading dataset********************//
for(i=0;i<row;i++)
for(j=0;j<col;j++)
{
//cin>>dataset[i][j];
g[i][j]=g1[i][j]=0;
}
//******************************************************//
//********Randomly assigning cluster centres************//
centroid1[0][0]=dataset[0][0];
centroid1[0][1]=dataset[0][1];
centroid2[0][0]=dataset[1][0];
centroid2[0][1]=dataset[1][1];
//******************************************************//
dist();
clust();
copy();
while(ch==0)
{
x=y=s=0;
for(i=0;i<row;i++)
{
if(g[i][0]==1)
{
x=x+dataset[i][0];
y=y+dataset[i][1];
s=s+1;
}
}
centroid1[0][0]=x/s;
centroid1[0][1]=y/s;
x=y=s=0;
for(i=0;i<row;i++)
{
if(g[i][1]==1)
{
x=x+dataset[i][0];
y=y+dataset[i][1];
s=s+1;
}
}
centroid2[0][0]=x/s;
centroid2[0][1]=y/s;
dist();
clust();
for(i=0;i<row;i++)
for(j=0;j<col;j++)
if(g[i][j]!=g1[i][j])
goto xyz;
ch=1;
xyz:copy();
}
cout<<"\n---We get the final grouping as the results---";
cout<<"\n\n\nObject\t Value1 Value2 Cluster"<<endl;
for(i=0;i<row;i++)
{
if(g[i][0]==1)
cout<<"\n\nmedicine"<<i<<" "<<dataset[i][0]<<"\t\t"<<dataset[i][1]<<"\t1";
if(g[i][1]==1)
cout<<"\n\nmedicine"<<i<<" "<<dataset[i][0]<<"\t\t"<<dataset[i][1]<<"\t2";
}
// getch();
return 0;
}
I want OpenMP version of this algo... Any help plzzz