Semantic search is powerful. It has the capability to overcome many intractable problems across diverse functional areas and teams, including early phase research, competitive intelligence, drug safety monitoring and more. While the power of semantic search is incontestable, has it eliminated the need for manual indexing? Michael Iarrobino, Product Management Leader at Copyright Clearance Center, provides the answer.
Semantic search has the ability to enhance the relevancy of query results by disambiguating concepts and applying the context surrounding the text to retrieve results that align with the user’s intent. In addition, it leverages automated methods such as named entity recognition or graph-based models, to achieve the result.
However, the capabilities of semantic search do not mean that indexing has no relevancy in content discovery. MEDLINE, a bibliographic database of life sciences and biomedical information, is a testament to the continued relevance of indexing. Hierarchical Medical Subject Headings (MeSH) are applied to the records in the MEDLINE database records manually. These headings facilitate information retrieval workflows for users of PubMed.
As the volume of research continues to grow, the volume of data processing tasks done through machines will reflect an upward trend. However, as the example of MEDLINE and how its manually indexed database facilitates information retrieval in PubMed demonstrates, there is always a place for high-quality, manually curated data sets. Hence, both semantic search and manual indexing are still relevant as each address singular needs in their respective domains.
Click here to read the article.
Please give your feedback on this article or share a similar story for publishing by clicking here.
