A hybrid strategy to extract metadata from scholarly articles by utilizing support vector machine and heuristics

Abstract

The immense growth in online research publications has attracted the research community to extract valuable information from scientific resources by exploring online digital libraries and publishers’ websites. The metadata stored in a machine comprehendible form can facilitate a precise search to enlist most related articles by applying semantic queries to the document’s metadata and the structural elements. The online search engines and digital libraries offer only keyword-based search on full-body text, which creates excessive results. The research community in recent years has adopted different approaches to extract structural information from research documents. We have distributed the content of an article into two logical layouts and metadata levels. This strategy has given our technique an advantage over the state-of-the-art (SOTA) extracting metadata with diversified publication styles. The experimental results have revealed that the proposed approach has shown a significant gain in performance of 20.26% to 27.14%.

Appendix: Comparsion with Cermine and PDFX on GOLD-standard

XQuery for Cermine generated XML files

We used the following XQueries to get the desired metadata from Cermine system’s XML file output.

Journal name XQuery to find the name of the journal.

$<journaltitle>$

$\{data(\$articles/front/journal-meta/journal-title-group/journal-title)\}$

$</journaltitle>$

Title XQuery to find the title of the article.

$<title>$

$\{data(\$articles/front/article-meta/title-group/article-title)\}$

$</title>$

DOI XQuery to find the doi of the article.

$<doi>$

$\{data(\$articles/front/article-meta/article-id)\}$

$</doi>$

Year XQuery to find the year of the article.

$<pubyear>$

$\{data(\$articles/front/article-meta/pub-date/year)\}$

$</pubyear>$

Volume XQuery to find the volume of an article.

$<volume>$

$\{data(\$articles/front/article-meta/volume)\}$

$</volume>$

Issue XQuery to find the issue number of the article.

$<issue>$

$\{data(\$articles/front/article-meta/issue)\}$

$</issue>$

Pages XQuery to find the pages of the article.

$<firstpage>$

$\{data(\$articles/front/article-meta/fpage)\}$

$</firstpage>$

$<lastpage>$

$\{data(\$articles/front/article-meta/lpage)\}$

$</lastpage>$

Keywords XQuery to find the keywords of the article.

$<keywords>$ $\{ for \$keyword at \$j in$

$\$articles/front/article-meta/kwd-group/kwd$

$return keyword\{data(\$keyword)\} /keyword \}$

$\langle /keywords\rangle$

Authors XQuery to find the authors of the article.

$\langle authors\rangle$ $\{ for \$authors at \$j in$

$\$articles/front/article-meta/contrib-group/contrib/string-name$

$return \langle author\rangle \{data(\$authors)\} \langle /author\rangle \}$

$\langle /authors\rangle$

Affiliations XQuery to find the author affiliations of the article.

$\langle affiliations\rangle$

$\{$ $for \$affiliation at \$j in$

$\$articles/front/article-meta/contrib-group/aff$

$return \langle institution\rangle$

$\{data(\$affiliation/institution)\} \langle /institution\rangle$

$\}$ $\{$

$for \$affiliation at \$j in$

$\$articles/front/article-meta/contrib-group/aff$

$return \langle country\rangle$

$\{data(\$affiliation/country)\} \langle /country\rangle$

$\}$

$\langle /affiliations\rangle$

H1 XQuery to find the Heading level 1 of the article.

$\langle section1\rangle$

$\{$

$for \$h1 at \$j in \$articles/body/sec$

$return \langle h1\rangle \{data(\$h1/title)\} \langle /h1\rangle$

$\}$

$\langle /section1\rangle$

H2 XQuery to find the Heading level 2 of the article.

$\langle section2\rangle$

$\{$

$for \$h2 at \$j in \$articles/body/sec/sec$

$return \langle h2\rangle \{data(\$h2/title)\} \langle /h2\rangle$

$\}$

$\langle /section2\rangle$

H3 XQuery to find the Heading level 1 of the article.

$\langle section3\rangle$

$\{$ $for \$h3 at \$j in \$articles/body/sec/sec/sec$

$return \langle h3\rangle \{data(\$h3/title)\} \langle /h3\rangle$

$\}$

$\langle /section3\rangle$

References XQuery to find the number of references in the article.

$\langle refcnt\rangle$

$\{$

$for \$ref at \$j in \$articles/back/ref-list$

$return \langle ref\rangle \{count(\$ref/ref)\} \langle /ref\rangle$

$\}$

$\langle /refcnt\rangle$

Abstract XQuery to find the abstract of the article.

$\langle abstract\rangle$

$\{count(\$articles/front/article-meta/abstract)\}$

$\langle /abstract\rangle$

XQuery for PDFX generated XML files

We used the following XQueries to get the desired metadata from PDFX system’s XML file output.

Title XQuery to find the title of the article.

$\langle title\rangle$

$\{data(\$articles/article/front/title-group/article-title)\}$

$\langle /title\rangle$

DOI XQuery to find the doi of the article.

$\langle doi\rangle \{data(\$articles/meta/doi)\} \langle /doi\rangle$

Abstract XQuery to find the abstract of the article.

$\langle abstract\rangle$

$\{count(\$articles/article/front/abstract)\}$

$\langle /abstract\rangle$

Authors XQuery to find the authors of the article.

$\langle authors\rangle$ $\{ for \$authors at \$j in$

$\$articles/article/front/region/email$

$return \langle email\rangle \{data(\$authors)\} \langle /email\rangle \}$

$\langle /authors\rangle$

H1 XQuery to find the Heading level 1 of the article.

$\langle section1\rangle$

$\{$

$for \$h1 at \$j in \$articles/article/body/.//h1$

$return \langle h1\rangle \{data(\$h1)\} \langle /h1\rangle$

$\}$

$\langle /section1\rangle$

H2 XQuery to find the Heading level 2 of the article.

$\langle section2\rangle$

$\{$

$for \$h2 at \$j in \$articles/article/body/.//h2$

$return \langle h2\rangle \{data(\$h2)\} \langle /h2\rangle$

$\}$

$\langle /section2\rangle$

H3 XQuery to find the Heading level 3 of the article.

$\langle section3\rangle$

$\{$

$for \$h3 at \$j in \$articles/article/body/.//h3$

$return \langle h3\rangle \{data(\$h3)\} \langle /h3\rangle$

$\}$

$\langle /section3\rangle$

References XQuery to find the number of references in the article.

$\langle ref\rangle$

$\{count(\$articles/article/body/section/.//ref-list/ref)\}$

$\langle /ref\rangle$

Table caption XQuery to find the table captions in the article. $\langle tables\rangle \{$

$for \$captions at \$j in$

$\$articles/.//region[contains(@class, 'DoCO:TableBox')]$

$return \langle table\rangle \{data(\$captions/caption)\}\langle /table\rangle$ $\}\langle /tables\rangle$

Figure caption XQuery to find the Figure captions in the article.

$\langle figures\rangle \{$

$for \$captions at \$j in$

$\$articles/.//region[contains(@class, 'DoCO:FigureBox')]$

$return \langle fig\rangle \{data(\$captions/caption)\}\langle /fig\rangle$

$\}\langle /figures\rangle$