Datasets:
Tasks:
Text Classification
Modalities:
Text
Formats:
parquet
Languages:
English
Size:
10K<n<100K
Tags:
License:
Renamed parquet files and updated readme.
Browse files
README.md
CHANGED
@@ -1,4 +1,12 @@
|
|
1 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
2 |
dataset_info:
|
3 |
features:
|
4 |
- name: id
|
@@ -20,28 +28,179 @@ dataset_info:
|
|
20 |
'4': non_domain
|
21 |
splits:
|
22 |
- name: train
|
23 |
-
num_bytes:
|
24 |
num_examples: 4969
|
25 |
- name: test_ua
|
26 |
-
num_bytes:
|
27 |
num_examples: 540
|
28 |
- name: test_uq
|
29 |
-
num_bytes:
|
30 |
num_examples: 733
|
31 |
- name: test_ud
|
32 |
-
num_bytes:
|
33 |
num_examples: 4562
|
34 |
-
|
35 |
-
dataset_size: 4466195
|
36 |
configs:
|
37 |
- config_name: default
|
38 |
data_files:
|
39 |
- split: train
|
40 |
path: data/train-*
|
41 |
- split: test_ua
|
42 |
-
path: data/
|
43 |
- split: test_uq
|
44 |
-
path: data/
|
45 |
- split: test_ud
|
46 |
-
path: data/
|
47 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
+
pretty_name: SciEntsBank
|
3 |
+
license: cc-by-4.0
|
4 |
+
language:
|
5 |
+
- en
|
6 |
+
task_categories:
|
7 |
+
- text-classification
|
8 |
+
size_categories:
|
9 |
+
- 10K<n<100K
|
10 |
dataset_info:
|
11 |
features:
|
12 |
- name: id
|
|
|
28 |
'4': non_domain
|
29 |
splits:
|
30 |
- name: train
|
31 |
+
num_bytes: 232655
|
32 |
num_examples: 4969
|
33 |
- name: test_ua
|
34 |
+
num_bytes: 52730
|
35 |
num_examples: 540
|
36 |
- name: test_uq
|
37 |
+
num_bytes: 35716
|
38 |
num_examples: 733
|
39 |
- name: test_ud
|
40 |
+
num_bytes: 177307
|
41 |
num_examples: 4562
|
42 |
+
dataset_size: 498408
|
|
|
43 |
configs:
|
44 |
- config_name: default
|
45 |
data_files:
|
46 |
- split: train
|
47 |
path: data/train-*
|
48 |
- split: test_ua
|
49 |
+
path: data/test-ua-*
|
50 |
- split: test_uq
|
51 |
+
path: data/test-uq-*
|
52 |
- split: test_ud
|
53 |
+
path: data/test-ud-*
|
54 |
---
|
55 |
+
|
56 |
+
# Dataset Card for "SciEntsBank"
|
57 |
+
|
58 |
+
SciEntsBank is one of the two distinct subsets within the Student Response Analysis (SRA) corpus, the other subset being the
|
59 |
+
[Beetle](https://huggingface.co/datasets/nkazi/Beetle) dataset. Derived from student answers gathered by Nielsen et al. [1],
|
60 |
+
this dataset comprises nearly 11K responses to 197 assessment questions spanning 15 diverse science domains. The dataset
|
61 |
+
features three labeling schemes: (a) 5-way, (b) 3-way, and (c) 2-way. The dataset includes a training set and three distinct
|
62 |
+
test sets: (a) Unseen Answers (`test_ua`), (b) Unseen Questions (`test_uq`), and (c) Unseen Domains (`test_ud`).
|
63 |
+
|
64 |
+
- **Authors:** Myroslava Dzikovska, Rodney Nielsen, Chris Brew, Claudia Leacock, Danilo Giampiccolo, Luisa Bentivogli, Peter Clark, Ido Dagan, Hoa Trang Dang
|
65 |
+
- **Paper:** [SemEval-2013 Task 7: The Joint Student Response Analysis and 8th Recognizing Textual Entailment Challenge](https://aclanthology.org/S13-2045)
|
66 |
+
|
67 |
+
## Loading Dataset
|
68 |
+
|
69 |
+
```python
|
70 |
+
from datasets import load_dataset
|
71 |
+
dataset = load_dataset('nkazi/SciEntsBank')
|
72 |
+
```
|
73 |
+
|
74 |
+
## Labeling Schemes
|
75 |
+
|
76 |
+
The authors released the dataset with annotations using five labels (i.e., 5-way labeling scheme) for Automated Short-Answer Grading (ASAG).
|
77 |
+
Additionally, the authors have introduced two alternative labeling schemes, namely the 3-way and 2-way schemes, both derived from the 5-way
|
78 |
+
labeling scheme designed for Recognizing Textual Entailment (RTE). In the 3-way labeling scheme, the categories "partially correct but
|
79 |
+
incomplete", "irrelevant", and "non-domain" are consolidated into a unified category labeled as "incorrect". On the other hand, the 2-way
|
80 |
+
labeling scheme simplifies the classification into a binary system where all labels except "correct" are merged under the "incorrect" category.
|
81 |
+
|
82 |
+
The `label` column in this dataset presents the 5-way labels. For 3-way and 2-way labels, use the code provided below to derive it
|
83 |
+
from the 5-way labels. After converting the labels, please verify the label distribution. A code to print the label distribution is
|
84 |
+
also given below.
|
85 |
+
|
86 |
+
### 5-way to 3-way
|
87 |
+
|
88 |
+
```python
|
89 |
+
from datasets import ClassLabel
|
90 |
+
|
91 |
+
dataset = dataset.align_labels_with_mapping({'correct': 0, 'contradictory': 1, 'partially_correct_incomplete': 2, 'irrelevant': 2, 'non_domain': 2}, 'label')
|
92 |
+
dataset = dataset.cast_column('label', ClassLabel(names=['correct', 'contradictory', 'incorrect']))
|
93 |
+
```
|
94 |
+
|
95 |
+
Using `align_labels_with_mapping()`, we are mapping "partially correct but incomplete", "irrelevant", and "non-domain" to the same id. Subsequently,
|
96 |
+
we are using `cast_column()` to redefine the class labels (i.e., the label feature) where the id 2 corresponds to the "incorrect" label.
|
97 |
+
|
98 |
+
### 5-way to 2-way
|
99 |
+
|
100 |
+
```python
|
101 |
+
from datasets import ClassLabel
|
102 |
+
|
103 |
+
dataset = dataset.align_labels_with_mapping({'correct': 0, 'contradictory': 1, 'partially_correct_incomplete': 1, 'irrelevant': 1, 'non_domain': 1}, 'label')
|
104 |
+
dataset = dataset.cast_column('label', ClassLabel(names=['correct', 'incorrect']))
|
105 |
+
```
|
106 |
+
|
107 |
+
In the above code, the label "correct" is mapped to 0 to maintain consistency with both the 5-way and 3-way labeling schemes. If the preference is to
|
108 |
+
represent "correct" with id 1 and "incorrect" with id 0, either adjust the label map accordingly or run the following to switch the ids:
|
109 |
+
|
110 |
+
```python
|
111 |
+
dataset = dataset.align_labels_with_mapping({'incorrect': 0, 'correct': 1}, 'label')
|
112 |
+
```
|
113 |
+
|
114 |
+
### Saving and loading 3-way and 2-way datasets
|
115 |
+
|
116 |
+
Use the following code to store the dataset with the 3-way (or 2-way) labeling scheme locally to eliminate the need to convert labels each time the dataset is loaded:
|
117 |
+
|
118 |
+
```python
|
119 |
+
dataset.save_to_disk('SciEntsBank_3way')
|
120 |
+
```
|
121 |
+
|
122 |
+
Here, `SciEntsBank_3way` depicts the path/directory where the dataset will be stored. Use the following code to load the dataset from the same local directory/path:
|
123 |
+
|
124 |
+
```python
|
125 |
+
from datasets import DatasetDict
|
126 |
+
dataset = DatasetDict.load_from_disk('SciEntsBank_3way')
|
127 |
+
```
|
128 |
+
|
129 |
+
### Printing Label Distribution
|
130 |
+
|
131 |
+
Use the following code to print the label distribution:
|
132 |
+
|
133 |
+
```python
|
134 |
+
def print_label_dist(dataset):
|
135 |
+
for split_name in dataset:
|
136 |
+
print(split_name, ':')
|
137 |
+
num_examples = 0
|
138 |
+
for label in dataset[split_name].features['label'].names:
|
139 |
+
count = dataset[split_name]['label'].count(dataset[split_name].features['label'].str2int(label))
|
140 |
+
print(' ', label, ':', count)
|
141 |
+
num_examples += count
|
142 |
+
print(' total :', num_examples)
|
143 |
+
|
144 |
+
print_label_dist(dataset)
|
145 |
+
```
|
146 |
+
|
147 |
+
## Label Distribution
|
148 |
+
|
149 |
+
<style>
|
150 |
+
.label-dist th:not(:first-child), .label-dist td:not(:first-child) {
|
151 |
+
width: 15%;
|
152 |
+
}
|
153 |
+
</style>
|
154 |
+
|
155 |
+
<div class="label-dist">
|
156 |
+
|
157 |
+
### 5-way
|
158 |
+
|
159 |
+
Label | Train | Test UA | Test UQ | Test UD
|
160 |
+
--- | --: | --: | --: | --:
|
161 |
+
Correct | 2,008 | 233 | 301 | 1,917
|
162 |
+
Contradictory | 499 | 58 | 64 | 417
|
163 |
+
Partially correct but incomplete | 1,324 | 113 | 175 | 986
|
164 |
+
Irrelevant | 1,115 | 133 | 193 | 1,222
|
165 |
+
Non-domain | 23 | 3 | - | 20
|
166 |
+
Total | 4,969 | 540 | 733 | 4,562
|
167 |
+
|
168 |
+
### 3-way
|
169 |
+
|
170 |
+
Label | Train | Test UA | Test UQ | Test UD
|
171 |
+
--- | --: | --: | --: | --:
|
172 |
+
Correct | 2,008 | 233 | 301 | 1,917
|
173 |
+
Contradictory | 499 | 58 | 64 | 417
|
174 |
+
Incorrect | 2,462 | 249 | 368 | 2,228
|
175 |
+
Total | 4,969 | 540 | 733 | 4,562
|
176 |
+
|
177 |
+
### 2-way
|
178 |
+
|
179 |
+
Label | Train | Test UA | Test UQ | Test UD
|
180 |
+
--- | --: | --: | --: | --:
|
181 |
+
Correct | 2,008 | 233 | 301 | 1,917
|
182 |
+
Incorrect | 2,961 | 307 | 432 | 2,645
|
183 |
+
Total | 4,969 | 540 | 733 | 4,562
|
184 |
+
|
185 |
+
</div>
|
186 |
+
|
187 |
+
## Citation
|
188 |
+
|
189 |
+
```tex
|
190 |
+
@inproceedings{dzikovska2013semeval,
|
191 |
+
title = {{S}em{E}val-2013 Task 7: The Joint Student Response Analysis and 8th Recognizing Textual Entailment Challenge},
|
192 |
+
author = {Dzikovska, Myroslava and Nielsen, Rodney and Brew, Chris and Leacock, Claudia and Giampiccolo, Danilo and Bentivogli, Luisa and Clark, Peter and Dagan, Ido and Dang, Hoa Trang},
|
193 |
+
year = 2013,
|
194 |
+
month = jun,
|
195 |
+
booktitle = {Second Joint Conference on Lexical and Computational Semantics ({SEM}), Volume 2: Proceedings of the Seventh International Workshop on Semantic Evaluation ({S}em{E}val 2013)},
|
196 |
+
editor = {Manandhar, Suresh and Yuret, Deniz}
|
197 |
+
publisher = {Association for Computational Linguistics},
|
198 |
+
address = {Atlanta, Georgia, USA},
|
199 |
+
pages = {263--274},
|
200 |
+
url = {https://aclanthology.org/S13-2045},
|
201 |
+
}
|
202 |
+
```
|
203 |
+
|
204 |
+
## References
|
205 |
+
1. Rodney D. Nielsen, Wayne Ward, James H. Martin, and Martha Palmer. 2008. Annotating students' understanding of science
|
206 |
+
concepts. In *Proceedings of the Sixth International Language Resources and Evaluation Conference*, Marrakech, Morocco.
|
data/{test_ua-00000-of-00001.parquet → test-ua-00001.parquet}
RENAMED
File without changes
|
data/{test_ud-00000-of-00001.parquet → test-ud-00001.parquet}
RENAMED
File without changes
|
data/{test_uq-00000-of-00001.parquet → test-uq-00001.parquet}
RENAMED
File without changes
|
data/{train-00000-of-00001.parquet → train-00001.parquet}
RENAMED
File without changes
|