pubmed.mineR: an R package with text-mining algorithms to analyse PubMed abstracts

Area of Science:

• Bioinformatics
• Computational Biology
• Text Mining

Background:

• PubMed is a vast repository of biomedical literature, posing data-mining challenges.
• Existing text-mining algorithms have limitations in speed, flexibility, and open-source availability.

Purpose of the Study:

• To develop an R package, pubmed.mineR, for efficient and flexible text-mining of PubMed data.
• To combine advantages of existing algorithms while overcoming their limitations.

Main Methods:

• Development of the pubmed.mineR R package.
• Integration with Bioconductor and Comprehensive R Archive Network (CRAN) packages.
• Demonstration using three case studies: diabetes educators, cancer risk, and disease comorbidity.

Main Results:

• The pubmed.mineR package offers user flexibility and links with other R packages.
• The package demonstrates fast processing times on large datasets using standard workstations.
• Case studies illustrate the package's utility in analyzing evolving scientific concepts.

Conclusions:

• pubmed.mineR provides an efficient, flexible, and open-source solution for PubMed text-mining.
• The package enhances capabilities for multifaceted approaches in biomedical literature analysis.
• pubmed.mineR is available for download, promoting wider scientific application.