control experiment analysis

jayadevbhaskaran · May 26, 2019 · ada6832 · ada6832
1 parent 663d5cc
commit ada6832
Show file tree

Hide file tree

Showing 2 changed files with 13 additions and 3 deletions.
diff --git a/src/analysis.py b/src/analysis.py
@@ -28,6 +28,7 @@
 print(results)
 np.savetxt(Config.RESULTS_FILE, np.array(results), 
  header="t prob f-m (models: logreg lstm bert)")
+print("\n")
 ###############################################################################
 #Ad-hoc analysis for gender differences in SST-2 train dataset
 train = pd.read_csv(Config.TSV_TRAIN, sep="\t", header=None, names=["idx", "class", "dummy", "text"])
@@ -51,6 +52,7 @@
  f_list.append(item['_1'])
  break
 print(len(f_list), np.mean(f_list)) 
+print("\n")
 
 '''
 for p in Config.PROFESSIONS:
@@ -68,7 +70,6 @@
 map1 = {}
 map2 = {}
 for p in Config.PROFESSIONS:
- print(p)
  male0 = preds0[20*i:20*(i+1)]
  male1 = preds1[20*i:20*(i+1)]
  male2 = preds2[20*i:20*(i+1)]
@@ -79,10 +80,18 @@
 
  #print(np.mean(female0 + male0), np.mean(female0) - np.mean(male0), "\n")
  #print(np.mean(female1 + male1), np.mean(female1) - np.mean(male1), "\n")
- print(np.mean(female2 + male2), np.mean(female2) - np.mean(male2), "\n")
+ print(p, np.mean(female2 + male2), np.mean(female2) - np.mean(male2))
 
  i = i+1
+
+control_preds = []
+control = preds2_with_control[-40:]
+male_c = control[:20]
+female_c = control[20:]
+print("control", np.mean(female_c + male_c), np.mean(female_c) - np.mean(male_c))
+print("\n")
 ###############################################################################
+'''
 sentences = utils.get_sentences()
 args0 = np.argsort(preds0)
 args1 = np.argsort(preds1)
@@ -93,6 +102,7 @@
 
 for i in range(10):
  print(sentences[args2[799 - i]])
+'''
 ###############################################################################
 for noun in Config.MALE_NOUNS:
  noun_preds = []

diff --git a/src/utils.py b/src/utils.py
@@ -30,7 +30,7 @@ def ttest(preds):
  female_probs = preds[n:]
  (t, prob) = ttest_rel(male_probs, female_probs)
  diff = np.mean(female_probs) - np.mean(male_probs)
- print(np.mean(male_probs), np.mean(female_probs), diff, prob)
+ #print(np.mean(male_probs), np.mean(female_probs), diff, prob)
  return (t, prob, diff)
 
 def glove2dict(src_filename):