-
Notifications
You must be signed in to change notification settings - Fork 1
/
utils.py
181 lines (144 loc) · 5.68 KB
/
utils.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
import numpy as np
from sklearn.metrics import confusion_matrix
from sklearn.utils.multiclass import unique_labels
from sklearn.linear_model import LinearRegression
import nltk
import syllables
import networkx as nx
from nltk.corpus import wordnet as wn
AOA_FILE = 'AoA_MCE.csv'
def load_scores(metrics):
df = pd.read_csv(AOA_FILE)
df = df[metrics + ['Word']]
scores = {}
for metric in metrics:
scores[metric] = {}
for _, row in df[[metric, 'Word']].dropna().iterrows():
scores[metric][row['Word'].lower()] = row[metric]
return scores
def linear_fit(true, pred):
model = LinearRegression()
model.fit(pred.reshape(-1, 1), true.reshape(-1, 1))
fit_line = np.arange(true.min(), true.max() + 1).reshape(-1, 1)
return model.predict(fit_line)
def scatterplot(y_true_train, y_pred_train,
y_true_test, y_pred_test,
xlabel='AoA',
ylabel='Predicted AoA',
filepath=None):
predictions_df_train = pd.DataFrame.from_records({
xlabel: y_true_train.ravel(),
ylabel: y_pred_train.ravel(),
'error': np.abs(y_true_train.ravel() - y_pred_train.ravel()).astype(np.int32)
})
predictions_df_test = pd.DataFrame.from_records({
xlabel: y_true_test.ravel(),
ylabel: y_pred_test.ravel(),
'error': np.abs(y_true_test.ravel() - y_pred_test.ravel()).astype(np.int32)
})
fig, ax = plt.subplots(1, 2)
fig.set_figwidth(12)
fig.set_figheight(6)
fig.set_tight_layout('tight')
ax[0].plot(np.arange(min(y_true_train), max(y_true_train) + 1),
np.arange(min(y_true_train), max(y_true_train) + 1), label='ideal fit')
ax[0].plot(np.arange(min(y_true_train), max(y_true_train) + 1),
linear_fit(y_true_train, y_pred_train), label='observed linear fit')
ax[0].set_title('Train')
ax[1].plot(np.arange(min(y_true_test), max(y_true_test) + 1),
np.arange(min(y_true_test), max(y_true_test) + 1), label='ideal fit')
ax[1].plot(np.arange(min(y_true_test), max(y_true_test) + 1),
linear_fit(y_true_test, y_pred_test), label='observed linear fit')
ax[1].set_title('Test')
sns.scatterplot(
x=xlabel, y=ylabel, data=predictions_df_train, hue='error', size_norm=True, ax=ax[0])
sns.scatterplot(
x=xlabel, y=ylabel, data=predictions_df_test, hue='error', size_norm=True, ax=ax[1])
fig.set_label('Scatterplot of AoA and predicted AoA values')
if filepath:
fig.savefig(filepath)
def plot_confusion_matrix(y_true, y_pred, classes,
normalize=False,
title=None,
cmap=plt.cm.Blues):
"""
This function prints and plots the confusion matrix.
Normalization can be applied by setting `normalize=True`.
"""
if not title:
if normalize:
title = 'Normalized confusion matrix'
else:
title = 'Confusion matrix, without normalization'
# Compute confusion matrix
cm = confusion_matrix(y_true, y_pred)
# Only use the labels that appear in the data
classes = classes[unique_labels(y_true, y_pred)]
if normalize:
cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]
print("Normalized confusion matrix")
else:
print('Confusion matrix, without normalization')
print(cm)
fig, ax = plt.subplots()
fig.set_figheight(12)
fig.set_figwidth(12)
im = ax.imshow(cm, interpolation='nearest', cmap=cmap)
ax.figure.colorbar(im, ax=ax)
# We want to show all ticks...
ax.set(xticks=np.arange(cm.shape[1]),
yticks=np.arange(cm.shape[0]),
# ... and label them with the respective list entries
xticklabels=classes, yticklabels=classes,
title=title,
ylabel='True label',
xlabel='Predicted label')
# Rotate the tick labels and set their alignment.
plt.setp(ax.get_xticklabels(), rotation=45, ha="right",
rotation_mode="anchor")
# Loop over data dimensions and create text annotations.
fmt = '.2f' if normalize else 'd'
thresh = cm.max() / 2.
for i in range(cm.shape[0]):
for j in range(cm.shape[1]):
ax.text(j, i, format(cm[i, j], fmt),
ha="center", va="center",
color="white" if cm[i, j] > thresh else "black")
fig.tight_layout()
return ax
def get_no_syllables(word):
return syllables.estimate(word)
def closure_graph(synset, fn):
visited = set([synset])
G = nx.DiGraph()
S = [synset]
while S != []:
u = S.pop(-1)
for v in fn(u):
if v in visited:
continue
visited.add(v)
G.add_node(v.name())
G.add_edge(u.name(), v.name())
S.append(v)
return G
def get_hypnonymy_tree_sizes(word):
hypernym_trees = []
hyponym_trees = []
hypernym_eccentricities = []
hyponym_eccentricities = []
for synset in wn.synsets(word):
hypernym_trees.append(closure_graph(synset, lambda s: s.hypernyms()))
hyponym_trees.append(closure_graph(synset, lambda s: s.hyponyms()))
hypernym_tree = hypernym_trees[-1]
hyponym_tree = hyponym_trees[-1]
if len(hypernym_tree) != 0:
hypernym_eccentricities.append(
nx.eccentricity(hypernym_tree, v=synset.name()))
if len(hyponym_tree) != 0:
hyponym_eccentricities.append(
nx.eccentricity(hyponym_tree, v=synset.name()))
return np.mean(hypernym_eccentricities), np.mean(hyponym_eccentricities)