Hugging Face's logo

Datasets:
research-backup
/
semeval2012_relational_similarity_v5

Modalities:
Text
Languages:
English
Size:
1K - 10K
Libraries:
Datasets
License:
Dataset card Viewer Files Community
asahi417 commited on
Commit
a4e3fbc
•
1 Parent(s): 71049f9

README

Browse files
Files changed (10) hide show
  1. .gitattributes +1 -0
  2. .gitignore +1 -0
  3. README.md +0 -0
  4. dataset/train.jsonl +3 -0
  5. dataset/valid.jsonl +0 -0
  6. get_stats.py +22 -0
  7. process.py +165 -0
  8. semeval2012_relational_similarity_v5.py +84 -0
  9. stats.csv +169 -0
  10. stats.md +170 -0
.gitattributes CHANGED
@@ -51,3 +51,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
51
  *.jpg filter=lfs diff=lfs merge=lfs -text
52
  *.jpeg filter=lfs diff=lfs merge=lfs -text
53
  *.webp filter=lfs diff=lfs merge=lfs -text
 
 
51
  *.jpg filter=lfs diff=lfs merge=lfs -text
52
  *.jpeg filter=lfs diff=lfs merge=lfs -text
53
  *.webp filter=lfs diff=lfs merge=lfs -text
54
+ dataset/train.jsonl filter=lfs diff=lfs merge=lfs -text
.gitignore ADDED
@@ -0,0 +1 @@
 
 
1
+ cache
README.md ADDED
File without changes
dataset/train.jsonl ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:488d0497eb7f0eff79ce06b95c471dbccb8481e5bdf2dd157b23cc6fb80fe183
3
+ size 11376392
dataset/valid.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
get_stats.py ADDED
@@ -0,0 +1,22 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import pandas as pd
2
+ from datasets import load_dataset
3
+
4
+ data = load_dataset('relbert/semeval2012_relational_similarity_v5')
5
+ stats = []
6
+ for k in data.keys():
7
+ for i in data[k]:
8
+ stats.append(
9
+ {
10
+ 'relation_type': i['relation_type'],
11
+ 'split': k,
12
+ 'positives': len(i['positives']),
13
+ 'negatives': len(i['negatives']),
14
+ 'level': i['level']
15
+ })
16
+ df = pd.DataFrame(stats)
17
+ g = df.groupby(['relation_type', 'level', 'split']).sum()
18
+ g.to_csv('stats.csv')
19
+ with open('stats.md', 'w') as f:
20
+ f.write(g.to_markdown())
21
+
22
+
process.py ADDED
@@ -0,0 +1,165 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import json
2
+ import os
3
+ import tarfile
4
+ import zipfile
5
+ import gzip
6
+ import requests
7
+ from itertools import chain
8
+ from glob import glob
9
+ import gdown
10
+
11
+ from datasets import load_dataset
12
+
13
+ k = 10 # the 3rd level negative-distance ranking
14
+ m = 5 # the 3rd level negative-distance ranking
15
+ top_n = 10 # threshold of positive pairs in the 1st and 2nd relation
16
+
17
+
18
+ def wget(url, cache_dir: str = './cache', gdrive_filename: str = None):
19
+ """ wget and uncompress data_iterator """
20
+ os.makedirs(cache_dir, exist_ok=True)
21
+ if url.startswith('https://drive.google.com'):
22
+ assert gdrive_filename is not None, 'please provide fileaname for gdrive download'
23
+ gdown.download(url, f'{cache_dir}/{gdrive_filename}', quiet=False)
24
+ filename = gdrive_filename
25
+ else:
26
+ filename = os.path.basename(url)
27
+ with open(f'{cache_dir}/{filename}', "wb") as f:
28
+ r = requests.get(url)
29
+ f.write(r.content)
30
+ path = f'{cache_dir}/{filename}'
31
+
32
+ if path.endswith('.tar.gz') or path.endswith('.tgz') or path.endswith('.tar'):
33
+ if path.endswith('.tar'):
34
+ tar = tarfile.open(path)
35
+ else:
36
+ tar = tarfile.open(path, "r:gz")
37
+ tar.extractall(cache_dir)
38
+ tar.close()
39
+ os.remove(path)
40
+ elif path.endswith('.zip'):
41
+ with zipfile.ZipFile(path, 'r') as zip_ref:
42
+ zip_ref.extractall(cache_dir)
43
+ os.remove(path)
44
+ elif path.endswith('.gz'):
45
+ with gzip.open(path, 'rb') as f:
46
+ with open(path.replace('.gz', ''), 'wb') as f_write:
47
+ f_write.write(f.read())
48
+ os.remove(path)
49
+
50
+
51
+ def get_training_data():
52
+ """ Get RelBERT training data
53
+
54
+ Returns
55
+ -------
56
+ pairs: dictionary of list (positive pairs, negative pairs)
57
+ {'1b': [[0.6, ('office', 'desk'), ..], [[-0.1, ('aaa', 'bbb'), ...]]
58
+ """
59
+ cache_dir = 'cache'
60
+ os.makedirs(cache_dir, exist_ok=True)
61
+ remove_relation = None
62
+ path_answer = f'{cache_dir}/Phase2Answers'
63
+ path_scale = f'{cache_dir}/Phase2AnswersScaled'
64
+ url = 'https://drive.google.com/u/0/uc?id=0BzcZKTSeYL8VYWtHVmxUR3FyUmc&export=download'
65
+ filename = 'SemEval-2012-Platinum-Ratings.tar.gz'
66
+ if not (os.path.exists(path_scale) and os.path.exists(path_answer)):
67
+ wget(url, gdrive_filename=filename, cache_dir=cache_dir)
68
+ files_answer = [os.path.basename(i) for i in glob(f'{path_answer}/*.txt')]
69
+ files_scale = [os.path.basename(i) for i in glob(f'{path_scale}/*.txt')]
70
+ assert files_answer == files_scale, f'files are not matched: {files_scale} vs {files_answer}'
71
+ positives = {}
72
+ negatives = {}
73
+ positives_limit = {}
74
+ all_relation_type = {}
75
+ # score_range = [90.0, 88.7] # the absolute value of max/min prototypicality rating
76
+ for i in files_scale:
77
+ relation_id = i.split('-')[-1].replace('.txt', '')
78
+ if remove_relation and int(relation_id[:-1]) in remove_relation:
79
+ continue
80
+ with open(f'{path_answer}/{i}', 'r') as f:
81
+ lines_answer = [_l.replace('"', '').split('\t') for _l in f.read().split('\n')
82
+ if not _l.startswith('#') and len(_l)]
83
+ relation_type = list(set(list(zip(*lines_answer))[-1]))
84
+ assert len(relation_type) == 1, relation_type
85
+ relation_type = relation_type[0]
86
+ with open(f'{path_scale}/{i}', 'r') as f:
87
+ # list of tuple [score, ("a", "b")]
88
+ scales = [[float(_l[:5]), _l[6:].replace('"', '')] for _l in f.read().split('\n')
89
+ if not _l.startswith('#') and len(_l)]
90
+ scales = sorted(scales, key=lambda _x: _x[0])
91
+ # positive pairs are in the reverse order of prototypicality score
92
+ positive_pairs = [[s, tuple(p.split(':'))] for s, p in filter(lambda _x: _x[0] > 0, scales)]
93
+ positive_pairs = sorted(positive_pairs, key=lambda x: x[0], reverse=True)
94
+ positives[relation_id] = list(list(zip(*positive_pairs))[1])
95
+ positives_limit[relation_id] = list(list(zip(*positive_pairs[:min(top_n, len(positive_pairs))]))[1])
96
+ negatives[relation_id] = [tuple(p.split(':')) for s, p in filter(lambda _x: _x[0] < 0, scales)]
97
+ all_relation_type[relation_id] = relation_type
98
+ parent = list(set([i[:-1] for i in all_relation_type.keys()]))
99
+
100
+ # 1st level relation contrast (among parent relations)
101
+ relation_pairs_1st = []
102
+ relation_pairs_1st_validation = []
103
+ for p in parent:
104
+ child_positive = list(filter(lambda x: x.startswith(p), list(all_relation_type.keys())))
105
+ child_negative = list(filter(lambda x: not x.startswith(p), list(all_relation_type.keys())))
106
+ positive_pairs = []
107
+ negative_pairs = []
108
+ for c in child_positive:
109
+ positive_pairs += positives_limit[c]
110
+ for c in child_negative:
111
+ negative_pairs += positives_limit[c]
112
+
113
+ relation_pairs_1st += [{
114
+ "positives": positive_pairs, "negatives": negative_pairs, "relation_type": p, "level": "parent"
115
+ }]
116
+
117
+ # 2nd level relation contrast (among child relations) & 3rd level relation contrast (within child relations)
118
+ relation_pairs_2nd = []
119
+ relation_pairs_2nd_validation = []
120
+ for p in all_relation_type.keys():
121
+ positive_pairs = positives_limit[p]
122
+ negative_pairs = []
123
+ for n in all_relation_type.keys():
124
+ if p == n:
125
+ continue
126
+ negative_pairs += positives[n]
127
+
128
+ relation_pairs_2nd += [{
129
+ "positives": positive_pairs, "negatives": negative_pairs, "relation_type": p, "level": "child"
130
+ }]
131
+
132
+ relation_pairs_3rd = []
133
+ for p in all_relation_type.keys():
134
+ positive_pairs = positives[p]
135
+ negative_pairs = positive_pairs + negatives[p]
136
+ for n, anchor in enumerate(positive_pairs):
137
+ if n > m:
138
+ continue
139
+ for _n, posi in enumerate(positive_pairs):
140
+ if n < _n and len(negative_pairs) > _n + k:
141
+ relation_pairs_3rd += [{
142
+ "positives": [(anchor, posi)],
143
+ "negatives": [(anchor, neg) for neg in negative_pairs[_n+k:]],
144
+ "relation_type": p,
145
+ "level": "child_prototypical"
146
+ }]
147
+
148
+ train = relation_pairs_1st + relation_pairs_2nd + relation_pairs_3rd
149
+
150
+ # conceptnet as the validation set
151
+ cn = load_dataset('relbert/conceptnet_high_confidence_v2')
152
+ valid = list(chain(*cn.values()))
153
+ for i in valid:
154
+ i['level'] = 'N/A'
155
+ return train, valid
156
+
157
+
158
+ if __name__ == '__main__':
159
+ data_train, data_validation = get_training_data()
160
+ print(f"- training data : {len(data_train)}")
161
+ print(f"- validation data : {len(data_validation)}")
162
+ with open('dataset/train.jsonl', 'w') as f_writer:
163
+ f_writer.write('\n'.join([json.dumps(i) for i in data_train]))
164
+ with open('dataset/valid.jsonl', 'w') as f_writer:
165
+ f_writer.write('\n'.join([json.dumps(i) for i in data_validation]))
semeval2012_relational_similarity_v5.py ADDED
@@ -0,0 +1,84 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import json
2
+ import datasets
3
+
4
+
5
+ logger = datasets.logging.get_logger(__name__)
6
+ _DESCRIPTION = """[SemEVAL 2012 task 2: Relational Similarity](https://aclanthology.org/S12-1047/)"""
7
+ _NAME = "semeval2012_relational_similarity_v5"
8
+ _VERSION = "1.1.0"
9
+ _CITATION = """
10
+ @inproceedings{jurgens-etal-2012-semeval,
11
+ title = "{S}em{E}val-2012 Task 2: Measuring Degrees of Relational Similarity",
12
+ author = "Jurgens, David and
13
+ Mohammad, Saif and
14
+ Turney, Peter and
15
+ Holyoak, Keith",
16
+ booktitle = "*{SEM} 2012: The First Joint Conference on Lexical and Computational Semantics {--} Volume 1: Proceedings of the main conference and the shared task, and Volume 2: Proceedings of the Sixth International Workshop on Semantic Evaluation ({S}em{E}val 2012)",
17
+ month = "7-8 " # jun,
18
+ year = "2012",
19
+ address = "Montr{\'e}al, Canada",
20
+ publisher = "Association for Computational Linguistics",
21
+ url = "https://aclanthology.org/S12-1047",
22
+ pages = "356--364",
23
+ }
24
+ """
25
+
26
+ _HOME_PAGE = "https://github.com/asahi417/relbert"
27
+ _URL = f'https://huggingface.co/datasets/relbert/{_NAME}/raw/main/dataset'
28
+ _URLS = {
29
+ str(datasets.Split.TRAIN): [f'{_URL}/train.jsonl'],
30
+ str(datasets.Split.VALIDATION): [f'{_URL}/valid.jsonl'],
31
+ }
32
+
33
+
34
+ class SemEVAL2012RelationalSimilarityV5Config(datasets.BuilderConfig):
35
+ """BuilderConfig"""
36
+
37
+ def __init__(self, **kwargs):
38
+ """BuilderConfig.
39
+ Args:
40
+ **kwargs: keyword arguments forwarded to super.
41
+ """
42
+ super(SemEVAL2012RelationalSimilarityV5Config, self).__init__(**kwargs)
43
+
44
+
45
+ class SemEVAL2012RelationalSimilarityV5(datasets.GeneratorBasedBuilder):
46
+ """Dataset."""
47
+
48
+ BUILDER_CONFIGS = [
49
+ SemEVAL2012RelationalSimilarityV5Config(
50
+ name=_NAME, version=datasets.Version(_VERSION), description=_DESCRIPTION
51
+ ),
52
+ ]
53
+
54
+ def _split_generators(self, dl_manager):
55
+ downloaded_file = dl_manager.download_and_extract(_URLS)
56
+ return [datasets.SplitGenerator(name=i, gen_kwargs={"filepaths": downloaded_file[str(i)]})
57
+ for i in [datasets.Split.TRAIN, datasets.Split.VALIDATION]]
58
+
59
+ def _generate_examples(self, filepaths):
60
+ _key = 0
61
+ for filepath in filepaths:
62
+ logger.info(f"generating examples from = {filepath}")
63
+ with open(filepath, encoding="utf-8") as f:
64
+ _list = [i for i in f.read().split('\n') if len(i) > 0]
65
+ for i in _list:
66
+ data = json.loads(i)
67
+ yield _key, data
68
+ _key += 1
69
+
70
+ def _info(self):
71
+ return datasets.DatasetInfo(
72
+ description=_DESCRIPTION,
73
+ features=datasets.Features(
74
+ {
75
+ "level": datasets.Value("string"),
76
+ "relation_type": datasets.Value("string"),
77
+ "positives": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
78
+ "negatives": datasets.Sequence(datasets.Sequence(datasets.Value("string"))),
79
+ }
80
+ ),
81
+ supervised_keys=None,
82
+ homepage=_HOME_PAGE,
83
+ citation=_CITATION,
84
+ )
stats.csv ADDED
@@ -0,0 +1,169 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ relation_type,level,positive (train),negative (train),positive (validation),negative (validation)
2
+ 1,parent,110,680,129,760
3
+ 10,parent,60,730,66,823
4
+ 10a,child,10,780,14,875
5
+ 10a,child_prototypical,1,18,1,22
6
+ 10b,child,10,780,13,876
7
+ 10b,child_prototypical,1,16,1,19
8
+ 10c,child,10,780,11,878
9
+ 10c,child_prototypical,1,19,1,20
10
+ 10d,child_prototypical,1,18,1,18
11
+ 10d,child,10,780,10,879
12
+ 10e,child,10,780,8,881
13
+ 10e,child_prototypical,1,14,1,12
14
+ 10f,child,10,780,10,879
15
+ 10f,child_prototypical,1,18,1,18
16
+ 1a,child,10,780,14,875
17
+ 1a,child_prototypical,1,16,1,20
18
+ 1b,child,10,780,14,875
19
+ 1b,child_prototypical,1,16,1,20
20
+ 1c,child_prototypical,1,19,1,20
21
+ 1c,child,10,780,11,878
22
+ 1d,child,10,780,16,873
23
+ 1d,child_prototypical,1,16,1,22
24
+ 1e,child,10,780,8,881
25
+ 1e,child_prototypical,1,13,1,11
26
+ 2,parent,100,690,117,772
27
+ 2a,child,10,780,15,874
28
+ 2a,child_prototypical,1,18,1,23
29
+ 2b,child_prototypical,1,15,1,16
30
+ 2b,child,10,780,11,878
31
+ 2c,child,10,780,13,876
32
+ 2c,child_prototypical,1,17,1,20
33
+ 2d,child,10,780,10,879
34
+ 2d,child_prototypical,1,17,1,17
35
+ 2e,child,10,780,11,878
36
+ 2e,child_prototypical,1,18,1,19
37
+ 2f,child,10,780,11,878
38
+ 2f,child_prototypical,1,21,1,22
39
+ 2g,child,10,780,16,873
40
+ 2g,child_prototypical,1,15,1,21
41
+ 2h,child_prototypical,1,18,1,19
42
+ 2h,child,10,780,11,878
43
+ 2i,child,10,780,9,880
44
+ 2i,child_prototypical,1,19,1,18
45
+ 2j,child,10,780,10,879
46
+ 2j,child_prototypical,1,20,1,20
47
+ 3,parent,80,710,80,809
48
+ 3a,child,10,780,11,878
49
+ 3a,child_prototypical,1,18,1,19
50
+ 3b,child,10,780,11,878
51
+ 3b,child_prototypical,1,21,1,22
52
+ 3c,child_prototypical,1,17,1,19
53
+ 3c,child,10,780,12,877
54
+ 3d,child,10,780,14,875
55
+ 3d,child_prototypical,1,17,1,21
56
+ 3e,child,10,780,5,884
57
+ 3e,child_prototypical,1,21,1,16
58
+ 3f,child,10,780,11,878
59
+ 3f,child_prototypical,1,22,1,23
60
+ 3g,child,10,780,6,883
61
+ 3g,child_prototypical,1,20,1,16
62
+ 3h,child_prototypical,1,20,1,20
63
+ 3h,child,10,780,10,879
64
+ 4,parent,80,710,82,807
65
+ 4a,child,10,780,11,878
66
+ 4a,child_prototypical,1,21,1,22
67
+ 4b,child,10,780,7,882
68
+ 4b,child_prototypical,1,16,1,13
69
+ 4c,child,10,780,12,877
70
+ 4c,child_prototypical,1,19,1,21
71
+ 4d,child_prototypical,1,15,1,9
72
+ 4d,child,10,780,4,885
73
+ 4e,child,10,780,12,877
74
+ 4e,child_prototypical,1,21,1,23
75
+ 4f,child,10,780,9,880
76
+ 4f,child_prototypical,1,21,1,20
77
+ 4g,child,10,780,15,874
78
+ 4g,child_prototypical,1,17,1,22
79
+ 4h,child_prototypical,1,20,1,22
80
+ 4h,child,10,780,12,877
81
+ 5,parent,90,700,105,784
82
+ 5a,child,10,780,14,875
83
+ 5a,child_prototypical,1,17,1,21
84
+ 5b,child_prototypical,1,20,1,18
85
+ 5b,child,10,780,8,881
86
+ 5c,child,10,780,11,878
87
+ 5c,child_prototypical,1,18,1,19
88
+ 5d,child,10,780,15,874
89
+ 5d,child_prototypical,1,16,1,21
90
+ 5e,child,10,780,8,881
91
+ 5e,child_prototypical,1,20,1,18
92
+ 5f,child,10,780,11,878
93
+ 5f,child_prototypical,1,20,1,21
94
+ 5g,child_prototypical,1,21,1,20
95
+ 5g,child,10,780,9,880
96
+ 5h,child,10,780,15,874
97
+ 5h,child_prototypical,1,19,1,24
98
+ 5i,child,10,780,14,875
99
+ 5i,child_prototypical,1,19,1,23
100
+ 6,parent,80,710,99,790
101
+ 6a,child,10,780,15,874
102
+ 6a,child_prototypical,1,17,1,22
103
+ 6b,child_prototypical,1,20,1,21
104
+ 6b,child,10,780,11,878
105
+ 6c,child_prototypical,1,20,1,23
106
+ 6c,child,10,780,13,876
107
+ 6d,child,10,780,10,879
108
+ 6d,child_prototypical,1,23,1,23
109
+ 6e,child,10,780,11,878
110
+ 6e,child_prototypical,1,20,1,21
111
+ 6f,child,10,780,12,877
112
+ 6f,child_prototypical,1,18,1,20
113
+ 6g,child,10,780,12,877
114
+ 6g,child_prototypical,1,17,1,19
115
+ 6h,child_prototypical,1,18,1,23
116
+ 6h,child,10,780,15,874
117
+ 7,parent,80,710,91,798
118
+ 7a,child,10,780,14,875
119
+ 7a,child_prototypical,1,19,1,23
120
+ 7b,child,10,780,7,882
121
+ 7b,child_prototypical,1,15,1,12
122
+ 7c,child,10,780,11,878
123
+ 7c,child_prototypical,1,16,1,17
124
+ 7d,child_prototypical,1,19,1,23
125
+ 7d,child,10,780,14,875
126
+ 7e,child_prototypical,1,16,1,16
127
+ 7e,child,10,780,10,879
128
+ 7f,child,10,780,12,877
129
+ 7f,child_prototypical,1,15,1,17
130
+ 7g,child,10,780,9,880
131
+ 7g,child_prototypical,1,13,1,12
132
+ 7h,child,10,780,14,875
133
+ 7h,child_prototypical,1,14,1,18
134
+ 8,parent,80,710,90,799
135
+ 8a,child,10,780,14,875
136
+ 8a,child_prototypical,1,16,1,20
137
+ 8b,child_prototypical,1,20,1,17
138
+ 8b,child,10,780,7,882
139
+ 8c,child,10,780,12,877
140
+ 8c,child_prototypical,1,15,1,17
141
+ 8d,child,10,780,13,876
142
+ 8d,child_prototypical,1,15,1,18
143
+ 8e,child,10,780,11,878
144
+ 8e,child_prototypical,1,15,1,16
145
+ 8f,child,10,780,12,877
146
+ 8f,child_prototypical,1,16,1,18
147
+ 8g,child_prototypical,1,12,1,9
148
+ 8g,child,10,780,7,882
149
+ 8h,child,10,780,14,875
150
+ 8h,child_prototypical,1,17,1,21
151
+ 9,parent,90,700,96,793
152
+ 9a,child,10,780,14,875
153
+ 9a,child_prototypical,1,14,1,18
154
+ 9b,child,10,780,12,877
155
+ 9b,child_prototypical,1,18,1,20
156
+ 9c,child,10,780,7,882
157
+ 9c,child_prototypical,1,9,1,6
158
+ 9d,child_prototypical,1,22,1,21
159
+ 9d,child,10,780,9,880
160
+ 9e,child,10,780,8,881
161
+ 9e,child_prototypical,1,23,1,21
162
+ 9f,child,10,780,10,879
163
+ 9f,child_prototypical,1,18,1,18
164
+ 9g,child,10,780,14,875
165
+ 9g,child_prototypical,1,15,1,19
166
+ 9h,child,10,780,13,876
167
+ 9h,child_prototypical,1,18,1,21
168
+ 9i,child,10,780,9,880
169
+ 9i,child_prototypical,1,18,1,17
stats.md ADDED
@@ -0,0 +1,170 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ | relation_type | level | positive (train) | negative (train) | positive (validation) | negative (validation) |
2
+ |:----------------|:-------------------|-------------------:|-------------------:|------------------------:|------------------------:|
3
+ | 1 | parent | 110 | 680 | 129 | 760 |
4
+ | 10 | parent | 60 | 730 | 66 | 823 |
5
+ | 10a | child | 10 | 780 | 14 | 875 |
6
+ | 10a | child_prototypical | 1 | 18 | 1 | 22 |
7
+ | 10b | child | 10 | 780 | 13 | 876 |
8
+ | 10b | child_prototypical | 1 | 16 | 1 | 19 |
9
+ | 10c | child | 10 | 780 | 11 | 878 |
10
+ | 10c | child_prototypical | 1 | 19 | 1 | 20 |
11
+ | 10d | child_prototypical | 1 | 18 | 1 | 18 |
12
+ | 10d | child | 10 | 780 | 10 | 879 |
13
+ | 10e | child | 10 | 780 | 8 | 881 |
14
+ | 10e | child_prototypical | 1 | 14 | 1 | 12 |
15
+ | 10f | child | 10 | 780 | 10 | 879 |
16
+ | 10f | child_prototypical | 1 | 18 | 1 | 18 |
17
+ | 1a | child | 10 | 780 | 14 | 875 |
18
+ | 1a | child_prototypical | 1 | 16 | 1 | 20 |
19
+ | 1b | child | 10 | 780 | 14 | 875 |
20
+ | 1b | child_prototypical | 1 | 16 | 1 | 20 |
21
+ | 1c | child_prototypical | 1 | 19 | 1 | 20 |
22
+ | 1c | child | 10 | 780 | 11 | 878 |
23
+ | 1d | child | 10 | 780 | 16 | 873 |
24
+ | 1d | child_prototypical | 1 | 16 | 1 | 22 |
25
+ | 1e | child | 10 | 780 | 8 | 881 |
26
+ | 1e | child_prototypical | 1 | 13 | 1 | 11 |
27
+ | 2 | parent | 100 | 690 | 117 | 772 |
28
+ | 2a | child | 10 | 780 | 15 | 874 |
29
+ | 2a | child_prototypical | 1 | 18 | 1 | 23 |
30
+ | 2b | child_prototypical | 1 | 15 | 1 | 16 |
31
+ | 2b | child | 10 | 780 | 11 | 878 |
32
+ | 2c | child | 10 | 780 | 13 | 876 |
33
+ | 2c | child_prototypical | 1 | 17 | 1 | 20 |
34
+ | 2d | child | 10 | 780 | 10 | 879 |
35
+ | 2d | child_prototypical | 1 | 17 | 1 | 17 |
36
+ | 2e | child | 10 | 780 | 11 | 878 |
37
+ | 2e | child_prototypical | 1 | 18 | 1 | 19 |
38
+ | 2f | child | 10 | 780 | 11 | 878 |
39
+ | 2f | child_prototypical | 1 | 21 | 1 | 22 |
40
+ | 2g | child | 10 | 780 | 16 | 873 |
41
+ | 2g | child_prototypical | 1 | 15 | 1 | 21 |
42
+ | 2h | child_prototypical | 1 | 18 | 1 | 19 |
43
+ | 2h | child | 10 | 780 | 11 | 878 |
44
+ | 2i | child | 10 | 780 | 9 | 880 |
45
+ | 2i | child_prototypical | 1 | 19 | 1 | 18 |
46
+ | 2j | child | 10 | 780 | 10 | 879 |
47
+ | 2j | child_prototypical | 1 | 20 | 1 | 20 |
48
+ | 3 | parent | 80 | 710 | 80 | 809 |
49
+ | 3a | child | 10 | 780 | 11 | 878 |
50
+ | 3a | child_prototypical | 1 | 18 | 1 | 19 |
51
+ | 3b | child | 10 | 780 | 11 | 878 |
52
+ | 3b | child_prototypical | 1 | 21 | 1 | 22 |
53
+ | 3c | child_prototypical | 1 | 17 | 1 | 19 |
54
+ | 3c | child | 10 | 780 | 12 | 877 |
55
+ | 3d | child | 10 | 780 | 14 | 875 |
56
+ | 3d | child_prototypical | 1 | 17 | 1 | 21 |
57
+ | 3e | child | 10 | 780 | 5 | 884 |
58
+ | 3e | child_prototypical | 1 | 21 | 1 | 16 |
59
+ | 3f | child | 10 | 780 | 11 | 878 |
60
+ | 3f | child_prototypical | 1 | 22 | 1 | 23 |
61
+ | 3g | child | 10 | 780 | 6 | 883 |
62
+ | 3g | child_prototypical | 1 | 20 | 1 | 16 |
63
+ | 3h | child_prototypical | 1 | 20 | 1 | 20 |
64
+ | 3h | child | 10 | 780 | 10 | 879 |
65
+ | 4 | parent | 80 | 710 | 82 | 807 |
66
+ | 4a | child | 10 | 780 | 11 | 878 |
67
+ | 4a | child_prototypical | 1 | 21 | 1 | 22 |
68
+ | 4b | child | 10 | 780 | 7 | 882 |
69
+ | 4b | child_prototypical | 1 | 16 | 1 | 13 |
70
+ | 4c | child | 10 | 780 | 12 | 877 |
71
+ | 4c | child_prototypical | 1 | 19 | 1 | 21 |
72
+ | 4d | child_prototypical | 1 | 15 | 1 | 9 |
73
+ | 4d | child | 10 | 780 | 4 | 885 |
74
+ | 4e | child | 10 | 780 | 12 | 877 |
75
+ | 4e | child_prototypical | 1 | 21 | 1 | 23 |
76
+ | 4f | child | 10 | 780 | 9 | 880 |
77
+ | 4f | child_prototypical | 1 | 21 | 1 | 20 |
78
+ | 4g | child | 10 | 780 | 15 | 874 |
79
+ | 4g | child_prototypical | 1 | 17 | 1 | 22 |
80
+ | 4h | child_prototypical | 1 | 20 | 1 | 22 |
81
+ | 4h | child | 10 | 780 | 12 | 877 |
82
+ | 5 | parent | 90 | 700 | 105 | 784 |
83
+ | 5a | child | 10 | 780 | 14 | 875 |
84
+ | 5a | child_prototypical | 1 | 17 | 1 | 21 |
85
+ | 5b | child_prototypical | 1 | 20 | 1 | 18 |
86
+ | 5b | child | 10 | 780 | 8 | 881 |
87
+ | 5c | child | 10 | 780 | 11 | 878 |
88
+ | 5c | child_prototypical | 1 | 18 | 1 | 19 |
89
+ | 5d | child | 10 | 780 | 15 | 874 |
90
+ | 5d | child_prototypical | 1 | 16 | 1 | 21 |
91
+ | 5e | child | 10 | 780 | 8 | 881 |
92
+ | 5e | child_prototypical | 1 | 20 | 1 | 18 |
93
+ | 5f | child | 10 | 780 | 11 | 878 |
94
+ | 5f | child_prototypical | 1 | 20 | 1 | 21 |
95
+ | 5g | child_prototypical | 1 | 21 | 1 | 20 |
96
+ | 5g | child | 10 | 780 | 9 | 880 |
97
+ | 5h | child | 10 | 780 | 15 | 874 |
98
+ | 5h | child_prototypical | 1 | 19 | 1 | 24 |
99
+ | 5i | child | 10 | 780 | 14 | 875 |
100
+ | 5i | child_prototypical | 1 | 19 | 1 | 23 |
101
+ | 6 | parent | 80 | 710 | 99 | 790 |
102
+ | 6a | child | 10 | 780 | 15 | 874 |
103
+ | 6a | child_prototypical | 1 | 17 | 1 | 22 |
104
+ | 6b | child_prototypical | 1 | 20 | 1 | 21 |
105
+ | 6b | child | 10 | 780 | 11 | 878 |
106
+ | 6c | child_prototypical | 1 | 20 | 1 | 23 |
107
+ | 6c | child | 10 | 780 | 13 | 876 |
108
+ | 6d | child | 10 | 780 | 10 | 879 |
109
+ | 6d | child_prototypical | 1 | 23 | 1 | 23 |
110
+ | 6e | child | 10 | 780 | 11 | 878 |
111
+ | 6e | child_prototypical | 1 | 20 | 1 | 21 |
112
+ | 6f | child | 10 | 780 | 12 | 877 |
113
+ | 6f | child_prototypical | 1 | 18 | 1 | 20 |
114
+ | 6g | child | 10 | 780 | 12 | 877 |
115
+ | 6g | child_prototypical | 1 | 17 | 1 | 19 |
116
+ | 6h | child_prototypical | 1 | 18 | 1 | 23 |
117
+ | 6h | child | 10 | 780 | 15 | 874 |
118
+ | 7 | parent | 80 | 710 | 91 | 798 |
119
+ | 7a | child | 10 | 780 | 14 | 875 |
120
+ | 7a | child_prototypical | 1 | 19 | 1 | 23 |
121
+ | 7b | child | 10 | 780 | 7 | 882 |
122
+ | 7b | child_prototypical | 1 | 15 | 1 | 12 |
123
+ | 7c | child | 10 | 780 | 11 | 878 |
124
+ | 7c | child_prototypical | 1 | 16 | 1 | 17 |
125
+ | 7d | child_prototypical | 1 | 19 | 1 | 23 |
126
+ | 7d | child | 10 | 780 | 14 | 875 |
127
+ | 7e | child_prototypical | 1 | 16 | 1 | 16 |
128
+ | 7e | child | 10 | 780 | 10 | 879 |
129
+ | 7f | child | 10 | 780 | 12 | 877 |
130
+ | 7f | child_prototypical | 1 | 15 | 1 | 17 |
131
+ | 7g | child | 10 | 780 | 9 | 880 |
132
+ | 7g | child_prototypical | 1 | 13 | 1 | 12 |
133
+ | 7h | child | 10 | 780 | 14 | 875 |
134
+ | 7h | child_prototypical | 1 | 14 | 1 | 18 |
135
+ | 8 | parent | 80 | 710 | 90 | 799 |
136
+ | 8a | child | 10 | 780 | 14 | 875 |
137
+ | 8a | child_prototypical | 1 | 16 | 1 | 20 |
138
+ | 8b | child_prototypical | 1 | 20 | 1 | 17 |
139
+ | 8b | child | 10 | 780 | 7 | 882 |
140
+ | 8c | child | 10 | 780 | 12 | 877 |
141
+ | 8c | child_prototypical | 1 | 15 | 1 | 17 |
142
+ | 8d | child | 10 | 780 | 13 | 876 |
143
+ | 8d | child_prototypical | 1 | 15 | 1 | 18 |
144
+ | 8e | child | 10 | 780 | 11 | 878 |
145
+ | 8e | child_prototypical | 1 | 15 | 1 | 16 |
146
+ | 8f | child | 10 | 780 | 12 | 877 |
147
+ | 8f | child_prototypical | 1 | 16 | 1 | 18 |
148
+ | 8g | child_prototypical | 1 | 12 | 1 | 9 |
149
+ | 8g | child | 10 | 780 | 7 | 882 |
150
+ | 8h | child | 10 | 780 | 14 | 875 |
151
+ | 8h | child_prototypical | 1 | 17 | 1 | 21 |
152
+ | 9 | parent | 90 | 700 | 96 | 793 |
153
+ | 9a | child | 10 | 780 | 14 | 875 |
154
+ | 9a | child_prototypical | 1 | 14 | 1 | 18 |
155
+ | 9b | child | 10 | 780 | 12 | 877 |
156
+ | 9b | child_prototypical | 1 | 18 | 1 | 20 |
157
+ | 9c | child | 10 | 780 | 7 | 882 |
158
+ | 9c | child_prototypical | 1 | 9 | 1 | 6 |
159
+ | 9d | child_prototypical | 1 | 22 | 1 | 21 |
160
+ | 9d | child | 10 | 780 | 9 | 880 |
161
+ | 9e | child | 10 | 780 | 8 | 881 |
162
+ | 9e | child_prototypical | 1 | 23 | 1 | 21 |
163
+ | 9f | child | 10 | 780 | 10 | 879 |
164
+ | 9f | child_prototypical | 1 | 18 | 1 | 18 |
165
+ | 9g | child | 10 | 780 | 14 | 875 |
166
+ | 9g | child_prototypical | 1 | 15 | 1 | 19 |
167
+ | 9h | child | 10 | 780 | 13 | 876 |
168
+ | 9h | child_prototypical | 1 | 18 | 1 | 21 |
169
+ | 9i | child | 10 | 780 | 9 | 880 |
170
+ | 9i | child_prototypical | 1 | 18 | 1 | 17 |