A Clustering Approach for Feature Selection in Microarray Data Classification Using Random Forest
2
조회
0
다운로드
1
인용
0
참고문헌
키워드
초록
Microarray data plays an essential role in diagnosing and detecting cancer. Microarray analysis allows theexamination of levels of gene expression in specific cell samples, where thousands of genes can be analyzedsimultaneously. However, microarray data have very little sample data and high data dimensionality. Therefore, to classify microarray data, a dimensional reduction process is required. Dimensional reductioncan eliminate redundancy of data; thus, features used in classification are features that only have a highcorrelation with their class. There are two types of dimensional reduction, namely feature selection andfeature extraction. In this paper, we used k-means algorithm as the clustering approach for feature selection. The proposed approach can be used to categorize features that have the same characteristics in one cluster, sothat redundancy in microarray data is removed. The result of clustering is ranked using the Relief algorithmsuch that the best scoring element for each cluster is obtained. All best elements of each cluster are selectedand used as features in the classification process. Next, the Random Forest algorithm is used. Based on thesimulation, the accuracy of the proposed approach for each dataset, namely Colon, Lung Cancer, and ProstateTumor, achieved 85.87%, 98.9%, and 89% accuracy, respectively. The accuracy of the proposed approach istherefore higher than the approach using Random Forest without clustering.
Abstract
Microarray data plays an essential role in diagnosing and detecting cancer. Microarray analysis allows theexamination of levels of gene expression in specific cell samples, where thousands of genes can be analyzedsimultaneously. However, microarray data have very little sample data and high data dimensionality. Therefore, to classify microarray data, a dimensional reduction process is required. Dimensional reductioncan eliminate redundancy of data; thus, features used in classification are features that only have a highcorrelation with their class. There are two types of dimensional reduction, namely feature selection andfeature extraction. In this paper, we used k-means algorithm as the clustering approach for feature selection. The proposed approach can be used to categorize features that have the same characteristics in one cluster, sothat redundancy in microarray data is removed. The result of clustering is ranked using the Relief algorithmsuch that the best scoring element for each cluster is obtained. All best elements of each cluster are selectedand used as features in the classification process. Next, the Random Forest algorithm is used. Based on thesimulation, the accuracy of the proposed approach for each dataset, namely Colon, Lung Cancer, and ProstateTumor, achieved 85.87%, 98.9%, and 89% accuracy, respectively. The accuracy of the proposed approach istherefore higher than the approach using Random Forest without clustering.
참고문헌 (13)
- American Cancer Society, "Cancer Facts & Figures 2015" (2015), American Cancer Society
- A. Nurfalah, "Cancer detection based on microarray data classification using PCA And modified back propagation", Far East Journal of Electronics and Communications, 16(2), pp.269~281 (2015)
- K. Moorthy, "Knowledge Technology", pp.174~183 (2012), Springer
- E. Pashaei, "A novel gene selection algorithm for cancer identification based on random forest and particle swarm optimization", Proceedings of 2015 IEEE Conference on Computational Intelligence in Bioinformatics and Computational Biology (CIBCB), pp.1~6 (2015)
- Z. M. Hira, "A review of feature selection and feature extraction methods applied on microarray data", Advances in Bioinformatics, 2015 (2015)
- P. K. Ammu, "Review on feature selection techniques of DNA microarray data", International Journal of Computer Applications, 61(12), pp.39~44 (2013)
- C. S. Tan, "A review of feature extraction software for microarray gene expression data", BioMed Research International, 2014 (2014)
- D. P. Ismi, "K-means clustering based filter feature selection on high dimensional data", International Journal of Advances in Intelligent Informatics, 2(1), pp.38~45 (2016)
- H. Aydadenta, "On the classification techniques in data mining for microarray data classification", Journal of Physics: Conference Series, 971(1) (2018)
- J. Biesiada, "Feature ranking methods based on information entropy with Parzen windows", Proceedings of International Conference on Research in Electrotechnology and Applied Informatics (2015)
- K. Kira, "The feature selection problem: traditional methods and a new algorithm", Proceedings of the 10th National Conference on Artificial Intelligence (AAAI), pp.129~134 (1992)
- R. Diaz-Uriarte, "Gene selection and classification of microarray data using random forest", BMC Bioinformatics, 7 (2006)
- J. Darbon, "Algorithms for overcoming the curse of dimensionality for certain Hamilton–Jacobi equations arising in control theory and elsewhere", Research in the Mathematical Sciences, 3 (2016)