Related Experiment Video
Updated: Jul 17, 2026
07:41
Performing Data Mining And Integrative Analysis Of Biomarker in Breast Cancer Using Multiple Publicly Accessible Databases
Published on: May 17, 2019
Efficient gene expression analysis by linking multiple data mining algorithms
Nikola Bogunovic1, Viktor Marohnic, Zeljko Debeljak
1Fac. of Electr. Eng. & Comput., Zagreb Univ.
Summary
This study introduces a novel data mining method for leukemia classification using gene micro-arrays. The MIFS/SVM hybrid efficiently identifies key genes for accurate disease diagnosis.
Area of Science:
- Bioinformatics
- Computational Biology
- Medical Informatics
Background:
- Accurate leukemia classification is crucial for effective treatment.
- High-dimensional gene expression data presents challenges for disease classification.
- Identifying relevant gene subsets is essential for improving classification accuracy.
Purpose of the Study:
- To evaluate a novel integrated data mining classification process for leukemia.
- To reduce the dimensional complexity of gene micro-array data.
- To develop an efficient and reliable tool for disease classification and gene selection.
Main Methods:
- Utilized gene micro-arrays from two leukemia types.
- Implemented a filter based on mutual information feature selection.
- Employed a support vector machines classifier within a leave-one-out cross-validation loop.
- Developed the MIFS/SVM hybrid tool.
Main Results:
- Successfully reduced the dimensional complexity of the gene micro-array data.
- Achieved efficient and reliable classification of leukemia types.
- The MIFS/SVM hybrid demonstrated effectiveness in identifying relevant gene subsets.
- Optimal parameters for classification and attribute selection were determined within a practical timeframe.
Conclusions:
- The MIFS/SVM hybrid is an efficient and reliable tool for leukemia classification.
- Mutual information feature selection coupled with SVM is effective for high-dimensional data.
- The developed method facilitates accurate disease classification and gene discovery.
