An Efficient Parallelized Ontology Network-Based Semantic Similarity Measure for Big Biomedical Document Clustering

Area of Science:

• Biomedical Informatics
• Computational Biology
• Data Science

Background:

• Semantic mining of large biomedical text corpora presents significant computational challenges.
• Existing ontology-based semantic similarity calculations are often too complex for big data analysis.
• Scalability and efficiency are critical for processing extensive biomedical document collections.

Purpose of the Study:

• To develop a parallelized semantic similarity measurement method for big biomedical text data.
• To address the limitations of traditional methods in handling large-scale datasets.
• To enable efficient semantic analysis and clustering of biomedical documents.

Main Methods:

• Document preprocessing and semantic feature extraction.
• Ontology network structure-based semantic similarity calculation within the Hadoop MapReduce framework.
• Document clustering using generated semantic similarity measures.

Main Results:

• Traditional methods fail with datasets exceeding ten thousand biomedical documents.
• The proposed parallelized method demonstrates efficiency and accuracy on large datasets.
• The approach exhibits high parallelism and scalability for big data analysis.

Conclusions:

• The Hadoop MapReduce-based parallelized semantic similarity method effectively handles big biomedical text data.
• This approach overcomes the scalability and complexity issues of traditional semantic analysis techniques.
• The method provides a robust solution for semantic mining and clustering in large biomedical corpora.