Chemical-text hybrid search engines

Area of Science:

• Chemistry
• Information Science
• Computer Science

Background:

• Increasing volume of chemical literature necessitates improved document retrieval.
• Existing text and chemical search engines produce inaccurate results (false positives/negatives).
• Current search methods struggle with diverse chemical document formats.

Purpose of the Study:

• To develop an advanced search approach for chemical literature.
• To overcome limitations of traditional text and chemical search engines.
• To enhance the accuracy and efficiency of locating specific chemical information.

Main Methods:

• Developed Entity-Canonical Keyword Indexing (ECKI) to convert chemical entities into canonical keywords.
• Implemented ECKI within Microsoft Office SharePoint Server Search.
• Created a hybrid search engine integrating chemical and keyword indexing.

Main Results:

• The hybrid search engine successfully supported complex mixed chemical and keyword queries.
• ECKI-based search was effective in both intranet and internet environments.
• Demonstrated improved accuracy and speed in retrieving relevant chemical documents.

Conclusions:

• Entity-Canonical Keyword Indexing (ECKI) offers a significant advancement in chemical information retrieval.
• This approach empowers researchers to conduct more sophisticated and precise searches.
• ECKI enhances the ability to locate specific chemical information, improving research efficiency.