ArneBinder commited on
Commit
b5db26b
1 Parent(s): 45cf7a6

add image with example

Browse files
Files changed (1) hide show
  1. README.md +43 -11
README.md CHANGED
@@ -1,14 +1,15 @@
1
  ---
2
  language:
3
- - en
4
  size_categories:
5
- - n<1K
6
  ---
7
- # Dataset Card for "CDCP"
 
8
 
9
  ### Dataset Summary
10
 
11
- CDCP (a.k.a. *Cornell eRulemaking Corpus*; [Park and Cardie, 2018](https://aclanthology.org/L18-1257.pdf)) consists of 731 user comments from an eRulemaking platform in the English language. There are five types of components (`Fact`, `Testimony`, `Reference`, `Value`, and `Policy`) and two types of supporting relations (`Reason` and `Evidence`) are annotated on the basis of the study by Park et al. (2015). The resulting dataset contains 4931 elementary unit and 1221 support relation annotations. (pp. 1623-1624)
12
 
13
  ### Supported Tasks and Leaderboards
14
 
@@ -55,11 +56,42 @@ The language in the dataset is English (AmE).
55
 
56
  ### Data Splits
57
 
58
- | | train | test |
59
- | ------------------------------------------------------------------------------------------------ | ---------------------------------------: | -------------------------------------: |
60
- | No. of instances | 581 | 150 |
61
- | No. of span labels<br/>- `Fact`<br/>- `Testimony`<br/>- `Reference`<br/>- `Value`<br/>- `Policy` | <br/>654<br/>873<br/>31<br/>1686<br/>662 | <br/>132<br/>244<br/>1<br/>496<br/>153 |
62
- | No. of relation labels<br/>- `reason`<br/>- `evidence` | <br/>1055<br/>47 | <br/>298<br/>26 |
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
63
 
64
  ## Dataset Creation
65
 
@@ -111,7 +143,7 @@ General public participants, implying American citizens.
111
 
112
  ### Discussion of Biases
113
 
114
- About 45% of the elementary units are `VALUE` type. A significant portion, roughly 75%, of support relation annotations are between adjacent elementary units. While commenters certainly tend to provide reasons immediately after the proposition to be supported, it is also easier for annotators to identify support relations in proximity. Thus, support relations in the wild may be not as skewed toward those between adjacent elementary units. (pp. 1626-1627)
115
 
116
  ### Other Known Limitations
117
 
@@ -140,4 +172,4 @@ About 45% of the elementary units are `VALUE` type. A significant portion, rough
140
 
141
  ### Contributions
142
 
143
- Thanks to [@idalr](https://github.com/idalr) for adding this dataset.
 
1
  ---
2
  language:
3
+ - en
4
  size_categories:
5
+ - n<1K
6
  ---
7
+
8
+ # Dataset Card for "cdcp"
9
 
10
  ### Dataset Summary
11
 
12
+ CDCP (a.k.a. *Cornell eRulemaking Corpus*; [Park and Cardie, 2018](https://aclanthology.org/L18-1257.pdf)) consists of 731 user comments from an eRulemaking platform in the English language. There are five types of components (`fact`, `testimony`, `reference`, `value`, and `policy`) and two types of supporting relations (`reason` and `evidence`) are annotated on the basis of the study by Park et al. (2015). The resulting dataset contains 4931 elementary unit and 1221 support relation annotations (pp. 1623-1624). The spans are segmented into elementary units with a proposition consisting of a sentence or a clause, as well as a few non-argumentative units (Morio et al., 2022, p. 642).
13
 
14
  ### Supported Tasks and Leaderboards
15
 
 
56
 
57
  ### Data Splits
58
 
59
+ | | train | test |
60
+ | ---------------- | ----: | ---: |
61
+ | No. of instances | 581 | 150 |
62
+
63
+ ### Label Description and Statistics
64
+
65
+ In this section, we report our own statistics of the corpus. However, there are yet discrepancies between our report, the author's report (see Park & Cardie, 2017, p. 1627, Table 2), and Morio et al. (2022)'s, who also utilized this corpus.
66
+
67
+ #### Components
68
+
69
+ | Components | train | test | total | percentage |
70
+ | ------------------------------------------------------------------- | ----------------------------------: | --------------------------------: | -----------------------------------: | -------------------------------------------: |
71
+ | `fact`<br/> `testimony`<br/> `reference`<br/> `value`<br/> `policy` | 654<br/>873<br/>31<br/>1686<br/>662 | 132<br/>244<br/>1<br/>496<br/>153 | 786<br/>1117<br/>32<br/>2182<br/>815 | 15.9%<br/>22.6%<br/>0.6%<br/>44.2%<br/>16.5% |
72
+
73
+ - `value`: "judgments without making specific claims about what should be done"
74
+ - `fact`: "expressing or dealing with facts or conditions as perceived without distortion by personal feelings, prejudices, or interpretations"
75
+ - `testimony`: "an objective proposition about the author’s personal state or experience"; "often practically impossible to provide objective evidence in online commenting setting"
76
+ - `policy`: "a specific course of action to be taken"; "typically contains modal verbs like “should” and “ought to.”"
77
+ - `reference`: "a source of objective evidence"
78
+
79
+ (Park & Cardie, 2018, p. 1625)
80
+
81
+ #### Relations
82
+
83
+ | Relations | train | test | total | percentage |
84
+ | :----------------------- | ----------: | ---------: | ----------: | -------------: |
85
+ | `reason`<br/> `evidence` | 1055<br/>47 | 298<br/>26 | 1353<br/>73 | 94.9%<br/>5.1% |
86
+
87
+ - `reason`: "X (source) is `reason` for a proposition Y (target; `policy`, `value`, `fact`, `testimony`) if X provides rationale for Y"
88
+ - `evidence`: "X (`testimony`, `fact`, `reference`) is `evidence` for a proposition Y if X proves whether proposition Y is true or not"
89
+
90
+ (Park & Cardie, 2018, pp. 1625-1626)
91
+
92
+ #### Examples
93
+
94
+ ![Examples](img/cdcp-sam.png)
95
 
96
  ## Dataset Creation
97
 
 
143
 
144
  ### Discussion of Biases
145
 
146
+ About 45% of the elementary units are `value` type. A significant portion, roughly 75%, of support relation annotations are between adjacent elementary units. While commenters certainly tend to provide reasons immediately after the proposition to be supported, it is also easier for annotators to identify support relations in proximity. Thus, support relations in the wild may be not as skewed toward those between adjacent elementary units. (pp. 1626-1627)
147
 
148
  ### Other Known Limitations
149
 
 
172
 
173
  ### Contributions
174
 
175
+ Thanks to [@idalr](https://github.com/idalr) for adding this dataset.