.

RESULTS.txt

@@ -0,0 +1,34 @@
+5-26
+timesteps      features                        f1-score   FN
+n=12  ['r1 s1','r1 s4','r1 s5','pa1 apiii']   0.966  [51 0 0 0 65]
+n=12  ['r1 s1','r1 s4','r1 s5']    0.9799   [0 0 0 0 57]
+n=18   ['r1 s1','r1 s4','r1 s5']  0.996   [0 0 0 0 10]
+n=24  error
+
+5-22   poques dades
+n=12  ['r1 s1','r1 s4','r1 s5','pa1 apiii']    
+n=12  ['r1 s1','r1 s4','r1 s5']    0.9799    
+n=18   ['r1 s1','r1 s4','r1 s5']  0.996    
+n=24  error
+
+
+
+5-18
+
+
+n=6 ['r1 s1','r1 s4','r1 s5']   0.941     [0 0 0 0 103]
+n=12  ['r1 s1','r1 s4','r1 s5','pa1 apiii']   0.951  [0 0 0 0 86]
+n=12  ['r1 s1','r1 s4','r1 s5']    0.934   [0 0 0 0 112]
+n=18   error
+n=24   ['r1 s1','r1 s4','r1 s5']  0.943   [0 0 0 0 92]
+
+
+
+
+30
+
+n=6  ['r1 s1','r1 s4','r1 s5']   error   [0 0 0 0 112]
+n=12  ['r1 s1','r1 s4','r1 s5','pa1 apiii']     error  [0 0 0 0 86]
+n=12  ['r1 s1','r1 s4','r1 s5']   error   [0 0 0 0 112]
+n=18     error
+n=24   ['r1 s1','r1 s4','r1 s5']   error   [0 0 0 0 92]

v2_unsupervised.py

@@ -203,7 +203,8 @@ def plotData():
     )
     for i in range(NumberOfFailures+1):
         for  j in range(NumFeatures):
-            axes[i][j].plot(np.concatenate((dataTrainNorm[i][:,j],dataTestNorm[i][:,j])),label="Fail "+str(i)+",  feature 0")
+            #axes[i][j].plot(np.concatenate((dataTrainNorm[i][:,j],dataTestNorm[i][:,j])),label="Fail "+str(i)+",  feature 0")
+            axes[i][j].plot(np.concatenate((dataTrain[i][:,j],dataTest[i][:,j])),label="Fail "+str(i)+",  feature 0")
             if i==(NumberOfFailures):
                 axes[i][j].set_xlabel(features[j])
             if j==0:
@@ -373,6 +374,7 @@ def anomalyMetric(labels,ranges):
     # Precision: Rate of positive results:  TP/(TP+FP)  
     # F1-score: predictive performance measure: 2*Precision*Sensitity/(Precision+Sensitity)
 
+    print("Labels on test: ",set(labels))
     lab=[]    #   Labels are assigned randomly by classifer
     TP=[]
     FN=[]