research-article

Differential prioritization in feature selection and classifier aggregation for multiclass microarray datasets

Authors:

Chia Huey Ooi,

Madhu Chetty,

Shyh Wei TengAuthors Info & Claims

Data Mining and Knowledge Discovery, Volume 14, Issue 3

Pages 329 - 366

https://doi.org/10.1007/s10618-006-0055-5

Published: 01 June 2007 Publication History

Abstract

The high dimensionality of microarray datasets endows the task of multiclass tissue classification with various difficulties—the main challenge being the selection of features deemed relevant and non-redundant to form the predictor set for classifier training. The necessity of varying the emphases on relevance and redundancy, through the use of the degree of differential prioritization (DDP) during the search for the predictor set is also of no small importance. Furthermore, there are several types of decomposition technique for the feature selection (FS) problem—all-classes-at-once, one-vs.-all (OVA) or pairwise (PW). Also, in multiclass problems, there is the need to consider the type of classifier aggregation used—whether non-aggregated (a single machine), or aggregated (OVA or PW). From here, first we propose a systematic approach to combining the distinct problems of FS and classification. Then, using eight well-known multiclass microarray datasets, we empirically demonstrate the effectiveness of the DDP in various combinations of FS decomposition types and classifier aggregation methods. Aided by the variable DDP, feature selection leads to classification performance which is better than that of rank-based or equal-priorities scoring methods and accuracies higher than previously reported for benchmark datasets with large number of classes. Finally, based on several criteria, we make general recommendations on the optimal choice of the combination of FS decomposition type and classifier aggregation method for multiclass microarray datasets.

References

[1]

Ambroise C and McLachlan GJ Selection bias in gene extraction on the basis of microarray gene-expression data Proc Natl Acad Sci USA 2002 99 6562-6566

Crossref

Google Scholar

[2]

Armstrong SA, Staunton JE, Silverman LB, Pieters R, den Boer ML, Minden MD, Sallan SE, Lander ES, Golub TR, and Korsmeyer SJ MLL translocations specify a distinct gene expression profile that distinguishes a unique leukemia Nat Genet 2002 30 41-47

Crossref

Google Scholar

[3]

Bhattacharjee A, Richards WG, Staunton JE, Li C, Monti S, Vasa P, Ladd C, Beheshti J, Bueno R, Gillette M, Loda M, Weber G, Mark EJ, Lander ES, Wong W, Johnson BE, Golub TR, Sugarbaker DJ, and Meyerson M Classification of human lung carcinomas by mRNA expression profiling reveals distinct adenocarcinoma subclasses Proc Natl Acad Sci USA 2001 98 13790-13795

Crossref

Google Scholar

[4]

Breiman L Bagging predictors Mach Learn 1996 24 123-140

Google Scholar

[5]

Decoste D and Schölkopf B Training invariant support vector machines Mach Learn 2002 46 161-190

Crossref

Google Scholar

[6]

Ding C, Peng H (2003) Minimum redundancy feature selection from microarray gene expression data. In: Proceedings of the 2nd IEEE computational systems bioinformatics conference, pp 523–529

Google Scholar

[7]

Dudoit S, Fridlyand J, and Speed T Comparison of discrimination methods for the classification of tumors using gene expression data J Am Stat Assoc 2002 97 77-87

Crossref

Google Scholar

[8]

Franc V (2005) Optimization algorithms for kernel methods. PhD thesis, Center for Machine Perception, Department of Cybernetics, Faculty of Electrical Engineering, Czech Technical University, 29 July, 2005. ftp://cmp.felk.cvut.cz/pub/cmp/articles/franc/Franc-PhD.pdf

Google Scholar

[9]

Garber ME, Troyanskaya OG, Schluens K, Petersen S, Thaesler Z, Pacyna-Gengelbach M, van de Rijn M, Rosen GD, Perou CM, Whyte RI, Altman RB, Brown PO, Botstein D, and Petersen I Diversity of gene expression in adenocarcinoma of the lung Proc Natl Acad Sci 2001 98 24 13784-13789

Crossref

Google Scholar

[10]

Golub TR, Slonim DK, Tamayo P, Huard C, Gaasenbeek M, Mesirov JP, Coller H, Loh ML, Downing JR, Caligiuri MA, Bloomfield CD, and Lander ES Molecular classification of cancer: class discovery and class prediction by gene expression monitoring Science 1999 286 531-537

Crossref

Google Scholar

[11]

Guyon I and Elisseeff A An introduction to variable and feature selection J Mach Learn Res 2003 3 1157-1182

Crossref

Google Scholar

[12]

Hall MA, Smith LA (1998) Practical feature subset selection for machine learning. In: Proceedings of the 21st Australasian computer science conference, pp 181–191

Google Scholar

[13]

Holm S A simple sequentially rejective multiple test procedure Scand J Stat 1979 6 65-70

Google Scholar

[14]

Hsu CW and Lin CJ A comparison of methods for multiclass support vector machines IEEE Trans Neural Netw 2002 13 2 415-425

Crossref

Google Scholar

[15]

Jirapech-Umpai T and Aitken S Feature selection and classification for microarray data analysis: evolutionary methods for identifying predictive genes BMC Bioinform 2005 6 148

Crossref

Google Scholar

[16]

Khan J, Wei JS, Ringner M, Saal LH, Ladanyi M, Westermann F, Berthold F, Schwab M, Antonescu CR, Peterson C, and Meltzer PS Classification and diagnostic prediction of cancers using expression profiling and artificial neural networks Nat Med 2001 7 673-679

Crossref

Google Scholar

[17]

Knijnenburg TA, Reinders MJT, Wessels LFA (2005) The selection of relevant and non-redundant features to improve classification performance of microarray gene expression data. In: Procedings of the 11th annual conference of the advanced school for computing and imaging, Heijen, NL. http://www.ict.ewi.tudelft.nl/pub/marcel/Knij05a.pdf

Google Scholar

[18]

Li T, Zhang C, and Ogihara M A comparative study of feature selection and multiclass classification methods for tissue classification based on gene expression Bioinformatics 2004 20 2429-2437

Crossref

Google Scholar

[19]

Linder R, Dew D, Sudhoff H, Theegarten D, Remberger K, Poppl SJ, and Wagner M The subsequent artificial neural network (SANN) approach might bring more classificatory power to ANN-based DNA microarray analyses Bioinformatics 2004 20 3544-3552

Crossref

Google Scholar

[20]

Massart DL, Vandeginste BGM, Deming SN, Michotte Y, Kaufman L (1988) The k-nearest neighbor method. Chemometrics: a textbook (Data handling in science and technology) vol 2, pp 395–397

Google Scholar

[21]

Mitchell T (1997) Machine learning. McGraw-Hill

Google Scholar

[22]

Munagala K, Tibshirani R, and Brown P Cancer characterization and feature set extraction by discriminative margin clustering BMC Bioinform 2004 5 21

Crossref

Google Scholar

[23]

Ooi CH, Chetty M, Gondal I (2004) The role of feature redundancy in tumor classification. In: Proceedings of the international conference bioinformatics and its applications (ICBA’04). Advances in bioinformatics and its applications (Mathematical Biology and Medicine), vol 8, pp 197–208

Google Scholar

[24]

Ooi CH, Chetty M, Teng SW (2005a) Relevance, redundancy and differential prioritization in feature selection for multiclass gene expression data. In: Proceedings of the 6th international symposium on biological and medical data analysis. Lecture notes in computer science, vol 3745, pp 367–378

Google Scholar

[25]

Ooi CH, Chetty M, Teng SW. (2005b) Modeling microarray datasets for efficient feature selection. In: Proceedings of the 4th Australasian conference on knowledge discovery and data mining (AusDM05), pp 115–129

Google Scholar

[26]

Park M, Hastie T (2005) Hierarchical classification using shrunken centroids. Department of Statistics, Stanford University. Technical report. http://www-stat.stanford.edu/~hastie/Papers/hpam.pdf

Google Scholar

[27]

Platt JC Schölkopf B, Burges CJC, and Smola AJ Fast training of support vector machines using sequential minimal optimization Advances in Kernel methods 1998 Cambridge MIT Press 185-208

Google Scholar

[28]

Platt JC, Cristianini N, and Shawe-Taylor J Large margin DAGs for multiclass classification Adv Neural Inf Process Syst 2000 12 547-553

Google Scholar

[29]

Ramaswamy S, Tamayo P, Rifkin R, Mukherjee S, Yeang CH, Angelo M, Ladd C, Reich M, Latulippe E, Mesirov JP, Poggio T, Gerald W, Loda M, Lander ES, and Golub TR Multi-class cancer diagnosis using tumor gene expression signatures Proc Natl Acad Sci USA 2001 98 15149-15154

Crossref

Google Scholar

[30]

Ross DT, Scherf U, Eisen MB, Perou CM, Rees C, Spellman P, Iyer V, Jeffrey SS, Vande Rijn M, Waltham M, Pergamenschikov A, Lee JCF, Lashkari D, Shalon D, Myers TG, Weinstein JN, Botstein D, and Brown PO Systematic variation in gene expression patterns in human cancer cell lines Nat Genet 2000 24 227-235

Crossref

Google Scholar

[31]

Schena M, Shalon D, Davis RW, and Brown PO Quantitative monitoring of gene expression patterns with a complementary DNA microarray Science 1995 270 467-470

Crossref

Google Scholar

[32]

Shalon D, Smith SJ, and Brown PO A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization Genome Res 1996 6 7 639-645

Google Scholar

[33]

Slonim DK, Tamayo P, Mesirov JP, Golub TR, Lander ES (2000) Class prediction and discovery using gene expression data. In: RECOMB 2000, pp 263–272

Google Scholar

[34]

Vapnik VN (1998) Statistical learning theory. John Wiley and Sons

Google Scholar

[35]

Yeoh E-J, Ross ME, Shurtleff SA, Williams WK, Patel D, Mahfouz R, Behm FG, Raimondi SC, Relling MV, Patel A, Cheng C, Campana D, Wilkins D, Zhou X, Li J, Liu H, Pui C-H, Evans WE, Naeve C, Wong L, and Downing JR Classification, subtype discovery, and prediction of outcome in pediatric lymphoblastic leukemia by gene expression profiling Cancer Cell 2002 1 133-143

Crossref

Google Scholar

Cited By

View all

Rahman MChetty MBulach DWangikar P(2015)Frequency Decomposition Based Gene ClusteringProceeings, Part II, of the 22nd International Conference on Neural Information Processing - Volume 949010.1007/978-3-319-26535-3_20(170-181)Online publication date: 9-Nov-2015
https://dl.acm.org/doi/10.1007/978-3-319-26535-3_20
Lutu PEngelbrecht A(2013)Base Model Combination Algorithm for Resolving Tied Predictions for K-Nearest Neighbor OVA Ensemble ModelsINFORMS Journal on Computing10.5555/3214394.321439525:3(517-526)Online publication date: 1-Aug-2013
https://dl.acm.org/doi/10.5555/3214394.3214395
(2013)Positive-versus-Negative Classification for Model Aggregation in Predictive Data MiningINFORMS Journal on Computing10.5555/2700769.270078425:4(792-807)Online publication date: 1-Nov-2013
https://dl.acm.org/doi/10.5555/2700769.2700784
Show More Cited By

Index Terms

Differential prioritization in feature selection and classifier aggregation for multiclass microarray datasets
1. Applied computing
  1. Life and medical sciences
    1. Computational biology
    2. Genetics
2. Computing methodologies
  1. Machine learning
    1. Learning paradigms
      1. Supervised learning
        Supervised learning by classification
    2. Machine learning approaches
      1. Classification and regression trees

Index terms have been assigned to the content through auto-classification.

Recommendations

Learning ECOC Code Matrix for Multiclass Classification with Application to Glaucoma Diagnosis

Classification of different mechanisms of angle closure glaucoma (ACG) is important for medical diagnosis. Error-correcting output code (ECOC) is an effective approach for multiclass classification. In this study, we propose a new ensemble learning ...
A hybrid feature selection method for DNA microarray data

Gene expression profiles, which represent the state of a cell at a molecular level, have great potential as a medical diagnosis tool. In cancer classification, available training data sets are generally of a fairly small sample size compared to the ...
A heuristic classifier ensemble for huge datasets
AMT'11: Proceedings of the 7th international conference on Active media technology

This paper proposes a heuristic classifier ensemble to improve the performance of learning in multiclass problems. Although the more accurate classifier leads to a better performance, there is another approach to use many inaccurate classifiers while ...

Reviews

Reviewer: Yongxi Tan

The classification of tumor samples, and biomarker discovery, using DNA microarray gene expression data intensifies the need for feature (gene) selection prior to classification. This is due to the high dimensionality of microarray gene expression data, which typically has thousands of variables (genes), and fewer observations (samples), in which severe co-linearity is observed. Generally speaking, a good feature selection method should select features (genes) of high correlation with the sample class labels (high relevancy), and low correlation with each other (low redundancy). In existing correlation-based feature selection techniques, either no redundancy in the selected features is taken into account, or both redundancy and relevancy play equal roles in the feature selection. In contrast, an important parameter, degree of differential prioritization (DDP, where 0 < DDP lt;=1), is introduced, which characterizes the importance of relevancy and redundancy in feature selection and classification. Typically, high DDP puts more priority on maximizing relevance, at the cost of increasing redundancy, while low DDP puts more emphasis on minimizing redundancy, at the cost of decreasing relevancy. To systematically investigate the effect of DDP on feature selection and classification performance, three different feature-selection decomposition methods (all-classes-at-once, one-versus-all, and pair-wise) were combined with three conventional classifier aggregation methods (nonaggregated single-machine, aggregated one-versus-all, and aggregated pair-wise) to classify eight well-known multiclass microarray benchmark datasets, using varying DDP values and fixed sizes of selected features. It was found that the optimal value of DDP depends on both combination type and the number of classes in the dataset, and the classification accuracy depends more on the feature-selection decomposition method than the classifier aggregation method. In summary, this paper is well written and easy to understand. For those interested in feature selection and classification using high-dimensional microarray gene expression data, this paper is worth reading. Online Computing Reviews Service

Access critical reviews of Computing literature here

Become a reviewer for Computing Reviews.

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

cover image Data Mining and Knowledge Discovery

Data Mining and Knowledge Discovery Volume 14, Issue 3

Jun 2007

126 pages

ISSN:1384-5810

Issue’s Table of Contents

Publisher

Kluwer Academic Publishers

United States

Publication History

Published: 01 June 2007

Accepted: 10 July 2006

Received: 16 February 2006

Author Tags

Qualifiers

Research-article

Contributors

Other Metrics

View Article Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

6
Total Citations
View Citations
0
Total Downloads

Downloads (Last 12 months)0
Downloads (Last 6 weeks)0

Reflects downloads up to 01 Jan 2025

Other Metrics

View Author Metrics

Citations

Cited By

View all

Rahman MChetty MBulach DWangikar P(2015)Frequency Decomposition Based Gene ClusteringProceeings, Part II, of the 22nd International Conference on Neural Information Processing - Volume 949010.1007/978-3-319-26535-3_20(170-181)Online publication date: 9-Nov-2015
https://dl.acm.org/doi/10.1007/978-3-319-26535-3_20
Lutu PEngelbrecht A(2013)Base Model Combination Algorithm for Resolving Tied Predictions for K-Nearest Neighbor OVA Ensemble ModelsINFORMS Journal on Computing10.5555/3214394.321439525:3(517-526)Online publication date: 1-Aug-2013
https://dl.acm.org/doi/10.5555/3214394.3214395
(2013)Positive-versus-Negative Classification for Model Aggregation in Predictive Data MiningINFORMS Journal on Computing10.5555/2700769.270078425:4(792-807)Online publication date: 1-Nov-2013
https://dl.acm.org/doi/10.5555/2700769.2700784
Lutu PEngelbrecht A(2013)Base Model Combination Algorithm for Resolving Tied Predictions for K-Nearest Neighbor OVA Ensemble ModelsINFORMS Journal on Computing10.5555/2508911.250892225:3(517-526)Online publication date: 1-Jul-2013
https://dl.acm.org/doi/10.5555/2508911.2508922
Ferraro MIrpino AVerde RGuarracino M(2013)A Novel Feature Selection Method for Classification Using a Fuzzy CriterionRevised Selected Papers of the 7th International Conference on Learning and Intelligent Optimization - Volume 799710.1007/978-3-642-44973-4_49(455-467)Online publication date: 7-Jan-2013
https://dl.acm.org/doi/10.1007/978-3-642-44973-4_49
Lutu PEngelbrecht A(2010)A decision rule-based method for feature selection in predictive data miningExpert Systems with Applications: An International Journal10.1016/j.eswa.2009.06.03137:1(602-609)Online publication date: 1-Jan-2010
https://dl.acm.org/doi/10.1016/j.eswa.2009.06.031

Abstract

References

Cited By

Index Terms

Recommendations

Learning ECOC Code Matrix for Multiclass Classification with Application to Glaucoma Diagnosis

A hybrid feature selection method for DNA microarray data

A heuristic classifier ensemble for huge datasets

Reviews

Access critical reviews of Computing literature here

Comments

Information

Published In

Publisher

Publication History

Author Tags

Qualifiers

Contributors

Other Metrics

Bibliometrics

Article Metrics

Other Metrics

Citations

Cited By

View options

Figures

Other

Share

Share this Publication link

Share on social media

Affiliations