Pointed out that non-English treatises are discriminated against by the treatise search service 'Google Scholar'

'Google Scholar' provided by Google is a treatise search engine that allows you to easily find academic treatises that exist on the Internet. However, it has been pointed out that the search algorithm used by Google Scholar may make it difficult for Google Scholar to search for non-English articles.

Future Internet | Free Full-Text | Language Bias in the Google Scholar Ranking Algorithm


Google Scholar renders documents not in English invisible

Google Scholar uses a search algorithm similar to the Google search engine. According to a research team at Pompeu Fabra University , SEO-conscious papers have appeared so that Google Scholar will be ranked high in search, just as websites perform search engine optimization (SEO) for the purpose of increasing the number of visits. It is said that it is.

In order to implement SEO to Google Scholar in your treatise, you need to understand the algorithm that evaluates the relevance ranking of Google Scholar. Of course, Google hasn't disclosed the details of Google Scholar's algorithm to prevent fraud.

The research team conducted 45 searches by author, year, and keyword to study Google Scholar's algorithm by reverse engineering based on statistical analysis, and obtained 1000 search results for each.

As a result, when I searched on Google Scholar, I found that 90% of articles written in languages other than English were displayed below the 900th in the search order and were hardly visualized.

The following is a graph of English articles (gray dots) and non-English articles (red dots) distributed on a graph with the citation ranking of the articles on the vertical axis and the search ranking on Google Scholar on the horizontal axis. It can be seen that non-English treatises have a search ranking of 900th or lower, regardless of the number of citations, which is clearly lower than English treatises.

The research team said, 'Google Scholar is originally a search engine that can obtain results in other languages, but it can be said that articles not written in English are discriminated against. This bias is not suitable for researchers around the world who do not speak English. It can be beneficial and makes non-English-speaking researchers believe that there is no literature in their own language when doing searches with multilingual results. '

in Web Service, Posted by log1i_yk