Dataset Card for Elsevier OA CC-By

Dataset Summary

Elsevier OA CC-By: This is a corpus of 40k (40,091) open access (OA) CC-BY articles from across Elsevier’s journals representing a large scale, cross-discipline set of research data to support NLP and ML research. The corpus include full-text articles published in 2014 to 2020 and are categorized in 27 Mid Level ASJC Code (subject classification).

Distribution of Publication Years

Publication Year	Number of Articles
2014	3018
2015	4438
2016	5913
2017	6419
2018	8016
2019	10135
2020	2159

Distribution of Articles Per Mid Level ASJC Code. Each article can belong to multiple ASJC codes.

Discipline	Count
General	3847
Agricultural and Biological Sciences	4840
Arts and Humanities	982
Biochemistry, Genetics and Molecular Biology	8356
Business, Management and Accounting	937
Chemical Engineering	1878
Chemistry	2490
Computer Science	2039
Decision Sciences	406
Earth and Planetary Sciences	2393
Economics, Econometrics and Finance	976
Energy	2730
Engineering	4778
Environmental Science	6049
Immunology and Microbiology	3211
Materials Science	3477
Mathematics	538
Medicine	7273
Neuroscience	3669
Nursing	308
Pharmacology, Toxicology and Pharmaceutics	2405
Physics and Astronomy	2404
Psychology	1760
Social Sciences	3540
Veterinary	991
Dentistry	40
Health Professions	821

Supported Tasks and Leaderboards

[More Information Needed]

Languages

English ( en ).

Dataset Structure

Data Instances

The original dataset was published with the following json structure:

{
    "docId": <str>,
    "metadata":{
        "title": <str>,
        "authors": [
            {
                "first": <str>,
                "initial": <str>,
                "last": <str>,
                "email": <str>
            },
            ...
        ],
        "issn": <str>,
        "volume": <str>,
        "firstpage": <str>,
        "lastpage": <str>,
        "pub_year": <int>,
        "doi": <str>,
        "pmid": <str>,
        "openaccess": "Full",
        "subjareas": [<str>],
        "keywords": [<str>],
        "asjc": [<int>],
    },
    "abstract":[
        {
          "sentence": <str>,
          "startOffset": <int>,
          "endOffset": <int>
        },
        ...
    ],
    "bib_entries":{
        "BIBREF0":{
            "title":<str>,
            "authors":[
                {
                "last":<str>,
                "initial":<str>,
                "first":<str>
                },
                ...
            ],
            "issn": <str>,
            "volume": <str>,
            "firstpage": <str>,
            "lastpage": <str>,
            "pub_year": <int>,
            "doi": <str>,
            "pmid": <str>
        },
        ...
    },
    "body_text":[
        {
        "sentence": <str>,
        "secId": <str>,
        "startOffset": <int>,
        "endOffset": <int>,
        "title": <str>,
        "refoffsets": {
            <str>:{
                "endOffset":<int>,
                "startOffset":<int>
                }
            },
        "parents": [
            {
            "id": <str>,
            "title": <str>
            },
            ...
        ]
    },
    ...
    ]
}

docId The docID is the identifier of the document. This is unique to the document, and can be resolved into a URL for the document through the addition of https//www.sciencedirect.com/science/pii/<docId>

abstract This is the author provided abstract for the document

body_text The full text for the document. The text has been split on sentence boundaries, thus making it easier to use across research projects. Each sentence has the title (and ID) of the section which it is from, along with titles (and IDs) of the parent section. The highest-level section takes index 0 in the parents array. If the array is empty then the title of the section for the sentence is the highest level section title. This will allow for the reconstruction of the article structure. References have been extracted from the sentences. The IDs of the extracted reference and their respective offset within the sentence can be found in the “refoffsets” field. The complete list of references are can be found in the “bib_entry” field along with the references’ respective metadata. Some will be missing as we only keep ‘clean’ sentences,

bib_entities All the references from within the document can be found in this section. If the meta data for the reference is available, it has been added against the key for the reference. Where possible information such as the document titles, authors, and relevant identifiers (DOI and PMID) are included. The keys for each reference can be found in the sentence where the reference is used with the start and end offset of where in the sentence that reference was used.

metadata Meta data includes additional information about the article, such as list of authors, relevant IDs (DOI and PMID). Along with a number of classification schemes such as ASJC and Subject Classification.

author_highlights Author highlights were included in the corpus where the author(s) have provided them. The coverage is 61% of all articles. The author highlights, consisting of 4 to 6 sentences, is provided by the author with the aim of summarising the core findings and results in the article.

Data Fields

• title : This is the author provided title for the document. 100% coverage.
• abstract : This is the author provided abstract for the document. 99.25% coverage.
• keywords : This is the author and publisher provided keywords for the document. 100% coverage.
• asjc : This is the disciplines for the document as represented by 334 ASJC (All Science Journal Classification) codes. 100% coverage.
• subjareas : This is the Subject Classification for the document as represented by 27 ASJC top-level subject classifications. 100% coverage.
• body_text : The full text for the document. 100% coverage.
• author_highlights : This is the author provided highlights for the document. 61.31% coverage.

Data Splits

Distribution of Publication Years

Train	Test	Validation
All Articles	32072	4009	4008
With Author Highlights	19644	2420	2514

Dataset Creation

Curation Rationale

[More Information Needed]

Source Data

Initial Data Collection and Normalization

Date the data was collected: 2020-06-25T11:00:00.000Z

See the original paper for more detail on the data collection process.

Who are the source language producers?

See 3.1 Data Sampling in the original paper .

Annotations

Annotation process

[More Information Needed]

Who are the annotators?

[More Information Needed]

Personal and Sensitive Information

[More Information Needed]

Considerations for Using the Data

Social Impact of Dataset

[More Information Needed]

Discussion of Biases

[More Information Needed]

Other Known Limitations

[More Information Needed]

Additional Information

Dataset Curators

[More Information Needed]

Licensing Information

CC BY 4.0

Citation Information

@article{Kershaw2020ElsevierOC,
  title     = {Elsevier OA CC-By Corpus},
  author    = {Daniel James Kershaw and R. Koeling},
  journal   = {ArXiv},
  year      = {2020},
  volume    = {abs/2008.00774},
  doi       = {https://doi.org/10.48550/arXiv.2008.00774},
  url       = {https://elsevier.digitalcommonsdata.com/datasets/zm33cdndxs},
  keywords  = {Science, Natural Language Processing, Machine Learning, Open Dataset},
  abstract  = {We introduce the Elsevier OA CC-BY corpus. This is the first open
               corpus of Scientific Research papers which has a representative sample
               from across scientific disciplines. This corpus not only includes the
               full text of the article, but also the metadata of the documents, 
               along with the bibliographic information for each reference.}
}

@dataset{https://10.17632/zm33cdndxs.3,
  doi       = {10.17632/zm33cdndxs.2},
  url       = {https://data.mendeley.com/datasets/zm33cdndxs/3},
  author    = "Daniel Kershaw and Rob Koeling",
  keywords  = {Science, Natural Language Processing, Machine Learning, Open Dataset},
  title     = {Elsevier OA CC-BY Corpus},
  publisher = {Mendeley},
  year      = {2020},
  month     = {sep}
}

Contributions

Thanks to @orieg for adding this dataset.