Fixing SVM Issue #32. There were two area that called the ML modules.…

… The first around line 61 was edited correctly, but the second around line 106 was missed. I corrected the problem by copying the code from the first instance to the second. This only seemed to be an issue with the evaluation dataset.

Chrono/TimePhraseToChrono/PeriodInterval.py

@@ -205,6 +205,9 @@ def buildPeriodInterval(s, chrono_id, chrono_list, ref_list, classifier, feats):
             if(classifier[1] == "NN"):
                 my_class = ChronoKeras.keras_classify(classifier[0], np.array(list(my_features.values())))
                 #print("Class: " + str(my_class) + " : Start: " + str(abs_Sspan) + " : End: "+ str(abs_Espan))
+            elif classifier[1] in ("SVM", "RF"):
+                feat_array = [int(i) for i in my_features.values()]
+                my_class = classifier[0].predict([feat_array])[0]
             else:
                 my_class = classifier[0].classify(my_features)
                 #print("Class: " + str(my_class) + " : Start: " + str(abs_Sspan) + " : End: "+ str(abs_Espan))