Newer
Older
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
## Quantity Expressions Dataset
This dataset contains the results of Ulrich Rabenstein's
[master thesis](https://gl.kwarc.info/supervision/MSc-archive/blob/master/2017/urabenstein/Rabenstein.pdf),
in which he developed a framework for the detection of quantity expressions in STEM documents.
### Accessibility and License
The content of this Dataset is licensed to [SIGMathLing members](/member/) for research
and tool development purposes.
Access is restricted to [SIGMathLing members](/member/) under the
[SIGMathLing Non-Disclosure-Agreement](/nda/) as for most [arXiv](http://arxiv.org)
articles, the right of distribution was only given (or assumed) to arXiv itself.
### Contents
* `Annotations.zip`: All quantity expressions detected by the spotter in a format suitable for the [Kwarc Annotation Tool (KAT)](https://github.com/kwarc/kat).
* `Documents.zip`: The documents in which quantity expressions were searched. These are modified arXMLiv documents in which each word is wrapped by a `<span>`. This was required by KAT to annotate words.
* `Harvest.zip`: Data for math web search.
* `screen-reader-documents.zip`: The documents prepared in a way that enables screen readers to read out units ("two kilometers" instead of "two k m" for "2km").
### Remarks on Annotation Format
The annotations are stored as RDF in a way suitable for the [Kwarc Annotation Tool (KAT)](https://github.com/kwarc/kat).
For more information on KAT and the KAT format consider reading [this](https://gl.kwarc.info/KAT/papers/blob/master/cicm14/paper.pdf)
and [this](https://gl.kwarc.info/KAT/papers/blob/master/cicm16/paper.pdf) paper.
In the example annotation below,
```
cse(%2F%2F*%5B%40id%3D'S1.p10.1'%5D%2C%2F%2F*%5B%40id%3D'S1.p10.1.w270'%5D%2C%2F%2F*%5B%40id%3D'S1.p10.1.w272'%5D)
```
describes the annotated quantity expression.
URL-decoding the expression in the parentheses, we can obtain the three comma-separated XPaths
```
//*[@id='S1.p10.1'],//*[@id='S1.p10.1.w270'],//*[@id='S1.p10.1.w272']
```
where the first path is the common parent, the second path is the start of the annotated range,
and the third path is the end of the annotated range.
```
<rdf:Description rdf:nodeID="KAT_5764208381">
<kat:run rdf:nodeID="kat_run"/>
<kat:kannspec rdf:nodeID="KAT_1_QuantityExpression"/>
<kat:concept>QuantityExpression</kat:concept>
<kat:type rdf:resource="http://kwarc.info/semanticextraction/KAnnSpec#quantityexpression"/>
<kat:annotates rdf:resource="http://localhost/procl.html#cse(%2F%2F*%5B%40id%3D'S1.p10.1'%5D%2C%2F%2F*%5B%40id%3D'S1.p10.1.w270'%5D%2C%2F%2F*%5B%40id%3D'S1.p10.1.w272'%5D)"/>
<kat:contentmathml rdf:parseType="Literal" score="1">
<apply>
<times/>
<cn>21</cn>
<apply>
<times/>
<apply>
<csymbol cd="Prefix">Prefix</csymbol>
<csymbol cd="centi">c</csymbol>
<csymbol cd="meter">m</csymbol>
</apply>
</apply>
</apply>
</kat:contentmathml>
</rdf:Description>
```
### Download
From [this repository](https://gl.kwarc.info/SIGMathLing/quantity-expressions) (only for [SIGMathLing members](/member/)).
### Evaluation
According to the thesis, a manual validation of 50 randomly selected documents containing in total 646 quantity expressions yielded the following values:
* Precision: 75%
* Recall: 93%
* F-Score: 83%