cluster SSW data by C

Parallel_Algorithm/OpenMP/Kmean_omp.c

@@ -7,7 +7,8 @@
 #include "../shared/math_util.h"
 
 /* This is the name of the data file we will read. */
-#define FILE_NAME "../test_data/Blobs_smp20000_fea30_cls8.nc"
+//#define FILE_NAME "../test_data/Blobs_smp20000_fea30_cls8.nc"
+#define FILE_NAME "../../Data_Analysis/data/SSWdata.nc"
 #define TOL 0.0001 
 #define MAX_ITER 100 
 
@@ -93,9 +94,11 @@ int main() {
  #pragma omp for schedule(static)
  for (i = 0; i < N_samples; i++) {
  k_best = 0;//assume cluster no.0 is the nearest
- dist_min = distance(N_features, X[i], old_cluster_centers[k_best]); 
+ //dist_min = distance(N_features, X[i], old_cluster_centers[k_best]); 
+ dist_min = correlation(N_features, X[i], old_cluster_centers[k_best]); 
  for (k = 1; k < N_clusters; k++){
- dist = distance(N_features, X[i], old_cluster_centers[k]); 
+ //dist = distance(N_features, X[i], old_cluster_centers[k]); 
+ dist = correlation(N_features, X[i], old_cluster_centers[k]); 
  if (dist < dist_min){
  dist_min = dist;
  k_best = k; 

Parallel_Algorithm/python_reference/check_SSWdata.py

@@ -0,0 +1,18 @@
+import sys 
+import numpy as np
+from IO_util import Raw_to_NetCDF
+import xarray as xr
+
+dirname = '../../Data_Analysis/data/'
+filename='SSWdata.nc'
+
+ds = xr.open_dataset(dirname+filename)
+
+print('total data size',ds["Y_TRUE"].size)
+print('size of 2nd cluster by MATLAB',ds["Y_TRUE"].sum())
+print('size of 2nd cluster by C',ds["Y_C"].sum())
+
+mismatch = (ds["Y_TRUE"].values != ds["Y_C"].values) 
+print("inconsistent labels: ",mismatch.sum())
+
+#ds.close()

Parallel_Algorithm/python_reference/convert_SSWdata.py

@@ -0,0 +1,42 @@
+import sys 
+import numpy as np
+from IO_util import Raw_to_NetCDF
+
+ndata = 17878
+nfeatures = 252 
+
+dirname = '../../Data_Analysis/data/'
+# Read data points
+file1=open(dirname+'SSWdata.bin','rb')
+X=np.fromfile(file1)
+if sys.byteorder=='little':
+ X.byteswap(True)
+X=X.reshape(ndata,nfeatures)
+
+# Read python label
+file1=open(dirname+'Label_py.bin','rb')
+Y_py=np.fromfile(file1,np.int32)
+if sys.byteorder=='little':
+ Y_py.byteswap(True)
+
+# Read matlab label
+file1=open(dirname+'Label_matlab.bin','rb')
+Y_matlab=np.fromfile(file1,np.int32)
+Y_matlab -= 1 # 1~2 to 0~1
+
+# ========================
+# convert the NetCDF format
+# ========================
+N_clusters = 2
+N_samples = ndata 
+N_features = nfeatures 
+N_repeat = 20
+
+initial_ind = np.zeros([N_repeat,N_clusters],dtype=np.int32)
+for i in range(N_repeat):
+ initial_ind[i,:] = np.random.choice(np.arange(N_samples),
+ N_clusters,replace=False)
+
+filename='SSWdata.nc'
+Raw_to_NetCDF(X,initial_ind,dirname+filename,y_true=Y_matlab)
+