Add purity measure to the script

Mathias
1 parent 503bfd9274
Showing 1 changed file with 15 additions and 2 deletions Side-by-side Diff
scripts/evaluations/clustering.py
@@ -18,7 +18,7 @@
 import matplotlib.pyplot as plt
  
 from volia.data_io import read_features,read_lst
-from volia.measures import entropy_score
+from volia.measures import entropy_score, purity_score
  
 '''
 TODO: 
@@ -77,6 +77,12 @@
     # V-Measure
     v_measure = v_measure_score(labels, predicted_labels)
  
+    # Purity
+    purity_scores = purity_score(labels, predicted_labels)
+    purity_class_score = purity_scores["purity_class_score"]
+    purity_cluster_score = purity_scores["purity_cluster_score"]
+    K = purity_scores["K"]
+
     # Write results
     with open(os.path.join(outdir, f"_" + args.prefix + "eval_clustering.log"), "w") as fd:
         print(f"F1-scores for each classes:\n{fscores_str}", file=fd)
@@ -85,6 +91,9 @@
         print(f"Homogeneity: {homogeneity}", file=fd)
         print(f"completeness: {completeness}", file=fd)
         print(f"v-measure: {v_measure}", file=fd)
+        print(f"purity class score: {purity_class_score}", file=fd)
+        print(f"purity cluster score: {purity_cluster_score}", file=fd)
+        print(f"purity overall evaluation criterion (K): {K}", file=fd)
  
     # Process t-SNE and plot
     tsne_estimator = TSNE()
  
@@ -128,8 +137,12 @@
         "entropy": entropy,
         "homogeneity": homogeneity,
         "completeness": completeness,
-        "v-measure": v_measure
+        "v-measure": v_measure,
+        "purity_class_score": purity_class_score,
+        "purity_cluster score": purity_cluster_score,
+        "K": K
     }
+
  
 if __name__ == "__main__":
     # Argparse
...	...	@@ -18,7 +18,7 @@
18	18	import matplotlib.pyplot as plt
19	19
20	20	from volia.data_io import read_features,read_lst
21		-from volia.measures import entropy_score
	21	+from volia.measures import entropy_score, purity_score
22	22
23	23	'''
24	24	TODO:
...	...	@@ -77,6 +77,12 @@
77	77	# V-Measure
78	78	v_measure = v_measure_score(labels, predicted_labels)
79	79
	80	+ # Purity
	81	+ purity_scores = purity_score(labels, predicted_labels)
	82	+ purity_class_score = purity_scores["purity_class_score"]
	83	+ purity_cluster_score = purity_scores["purity_cluster_score"]
	84	+ K = purity_scores["K"]
	85	+
80	86	# Write results
81	87	with open(os.path.join(outdir, f"_" + args.prefix + "eval_clustering.log"), "w") as fd:
82	88	print(f"F1-scores for each classes:\n{fscores_str}", file=fd)
...	...	@@ -85,6 +91,9 @@
85	91	print(f"Homogeneity: {homogeneity}", file=fd)
86	92	print(f"completeness: {completeness}", file=fd)
87	93	print(f"v-measure: {v_measure}", file=fd)
	94	+ print(f"purity class score: {purity_class_score}", file=fd)
	95	+ print(f"purity cluster score: {purity_cluster_score}", file=fd)
	96	+ print(f"purity overall evaluation criterion (K): {K}", file=fd)
88	97
89	98	# Process t-SNE and plot
90	99	tsne_estimator = TSNE()
91	100
...	...	@@ -128,8 +137,12 @@
128	137	"entropy": entropy,
129	138	"homogeneity": homogeneity,
130	139	"completeness": completeness,
131		- "v-measure": v_measure
	140	+ "v-measure": v_measure,
	141	+ "purity_class_score": purity_class_score,
	142	+ "purity_cluster score": purity_cluster_score,
	143	+ "K": K
132	144	}
	145	+
133	146
134	147	if __name__ == "__main__":
135	148	# Argparse