Related Experiment Videos
An environment for knowledge discovery in biology.
Junior Barrera1, Roberto M Cesar, João E Ferreira
1USP Center for Bioinformatics (BIOINFO-USP), University of São Paulo, Rua do Matão, 1010 São Paulo, SP 05508-900, Brazil. jb@ime.usp.br
Computers in Biology and Medicine
|May 18, 2004
Summary
This study introduces a data mining environment for bioinformatics, enabling knowledge discovery from biomedical data using a versatile kernel and high-performance computing. Experimental results demonstrate its effectiveness in classification tasks.
Area of Science:
- Bioinformatics
- Data Mining
- Knowledge Discovery
Background:
- Bioinformatics applications require efficient data mining for knowledge discovery.
- Existing systems may lack a unified environment for complex analyses.
Purpose of the Study:
- To describe a novel data mining environment for bioinformatics.
- To enable knowledge discovery from biomedical databases.
Main Methods:
- Developed a generic kernel with supervised and unsupervised classification capabilities.
- Utilized a biomedical information warehouse architecture.
- Incorporated a high-performance library for parallel processing.
Main Results:
- The environment successfully applied mining functions to extracted data.
- Results were stored in a complex object database for discovery.
- Experimental validation confirmed the kernel's utility.
Conclusions:
- The described data mining environment provides a robust platform for bioinformatics.
- The system facilitates efficient knowledge discovery through parallel computation.