High-dimensional microarray data presents a major challenge for accurate disease classification due to feature redundancy and limited sample sizes. We hypothesize that modeling multivariate gene relationships using a graph-based structure can enhance the effectiveness of gene selection without relying on labelled data. This paper proposes Feature Graph-based Unsupervised Gene Selection (FGUGS), a novel unsupervised filter method that constructs a threshold-free directed graph to represent pairwise gene similarities and selects a compact subset of genes by identifying high in-degree nodes, thereby minimizing redundancy. FGUGS eliminates the need for manual parameter tuning, enhances interpretability through graph analysis, and scales efficiently with the number of genes while preserving structural relationships. Experimental results on four benchmark datasets show that FGUGS outperforms both traditional and recent state-of-the-art gene selection methods, achieving up to a 20% improvement in classification accuracy and demonstrating strong clustering performance. FGUGS provides a reproducible and scalable solution for biomedical data analysis, particularly in scenarios where class labels are scarce or unavailable.
DashRMisraB. Gene selection and classification of microarray data: A pareto de approach. Intell Dec Technol2017; 11: 93–107. DOI:https://doi.org/10.3233/IDT-160280.
2.
ChamlalHOuaderhmanTRebbahFE. A hybrid feature selection approach for microarray datasets using graph theoretic-based method. Inf Sci (Ny)2022; 615: 449–474. DOI: https://doi.org/10.1016/j.ins.2022.10.001.
3.
KishoreAVenkataramanaLPrasadDVV, et al.Enhancing the prediction of idc breast cancer staging from gene expression profiles using hybrid feature selection methods and deep learning architecture. Med Biol Eng Comput2023; 61: 2895–2919. DOI: https://doi.org/10.1007/s11517-023-02892-1.
4.
FlaounasINIakovidisDKMaroulisDE. Cascading svms as a tool for medical diagnosis using multi-class gene expression data. Int J Artif Intell Tools2006; 15: 335–352. DOI: https://doi.org/10.1142/S0218213006002709.
5.
RostamiMForouzandehSBerahmandK, et al.Gene selection for microarray data classification via multi-objective graph theoretic-based method. Artif Intell Med2022; 123: 102228. DOI: https://doi.org/10.1016/j.artmed.2021.102228.
6.
AlhenawiEAl-SayyedRHudaibA, et al.Feature selection methods on gene expression microarray data for cancer classification: A systematic review. Comput Biol Med2022; 140: 105051.DOI: https://doi.org/10.1016/j.compbiomed.2021.105051.
7.
MoslemiAAhmadianA. Dual regularized subspace learning using adaptive graph learning and rank constraint: Unsupervised feature selection on gene expression microarray datasets. Comput Biol Med2023; 167: 107659.
8.
Bolón-CanedoVAlonso-BetanzosAMorán-FernándezL, et al.Feature selection: from the past to the future, Vol. 24, 2022. DOI: 10.1007/978-3-030-93052-3_2.
9.
AlzubiOAQiqiehIAlzubiJA. Fusion of deep learning based cyberattack detection and classification model for intelligent systems. Cluster Comput2023; 26: 1363–1374. DOI: https://doi.org/10.1007/s10586-022-03686-0.
10.
HallMA. Correlation-based feature selection for machine learning. PhD Thesis, The University of Waikato, Hamilton, New Zealand, 1999
11.
GuyonIWestonJBarnhillS, et al.Gene selection for cancer classification using support vector machines. Mach Learn2002; 46: 389–422. DOI:https://doi.org/10.1023/A:1012487302797.
12.
Saberi-MovahedFRostamiMBerahmandK, et al.Dual regularized unsupervised feature selection based on matrix factorization and minimum redundancy with application in gene selection. Knowl Based Syst2022; 256: 109884. DOI: https://doi.org/10.1016/j.knosys.2022.109884.
13.
NagarajanGBabuLDD. A hybrid feature selection model based on improved squirrel search algorithm and rank aggregation using fuzzy techniques for biomedical data classification. Netw Model Anal Health Inform Bioinform2021; 10: 39.
14.
XuKMaungCAraiH, et al.Two-stage feature selection with unsupervised second stage. Int J Artif Intell Tools2018; 27: 1860014. DOI: https://doi.org/10.1142/S021821301860014X.
15.
DjellaliHGhoualmi-ZineNGuessoumS. Hybrid adapted fast correlation fcbf-support vector machine recursive feature elimination for feature selection. Intell Dec Technol2020; 14: 182–198. DOI: https://doi.org/10.3233/IDT-190014.
16.
Sosa-CabreraGGarcía-TorresMGómez-GuerreroS, et al.A multivariate approach to the symmetrical uncertainty measure: Application to feature selection problem. Inf Sci (Ny)2019; 494: 1–20. DOI: https://doi.org/10.1016/j.ins.2019.04.046.
SunHJieWLooJ, et al.A parallel self-organizing overlapping community detection algorithm based on swarm intelligence for large scale complex networks. Future Gener Comput Sys2018; 89: 265–285.
19.
QiuPZhangCGaoD, et al.A fusion of centrality and correlation for feature selection. Expert Syst Appl2024; 241: 122548.
20.
LiangSMaAYangS, et al.A review of matched-pairs feature selection methods for gene expression data analysis, 2018. DOI: 10.1016/j.csbj.2018.02.005.
21.
ZhengLChaoFParthaláinNM, et al.Feature grouping and selection: A graph-based approach. Inf Sci (Ny)2021; 546: 1256–1272. DOI: https://doi.org/10.1016/j.ins.2020.09.022.
22.
AzadifarSRostamiMBerahmandK, et al.Graph-based relevancy-redundancy gene selection method for cancer diagnosis. Comput Biol Med2022; 147: 105766. DOI: https://doi.org/10.1016/j.compbiomed.2022.105766.
23.
SongQNiJWangG. A fast clustering-based feature subset selection algorithm for high-dimensional data. IEEE Trans Knowl Data Eng2013; 25: 1–14. DOI: https://doi.org/10.1109/TKDE.2011.181..
24.
MandalKSarmahR. Sgaclust: Semi-supervised graph attraction clustering of gene expression data. Netw Model Anal Health Inform Bioinform2022; 11: 24.
25.
TaheriKMoradiHTavassolipourM. Collaboration graph for feature set partitioning in data classification. Expert Syst Appl2023; 213: 118988. DOI: https://doi.org/10.1016/j.eswa.2022.118988.
26.
HuangPKongZXieM, et al.Robust unsupervised feature selection via data relationship learning. Pattern Recognit2023; 142: 109676. DOI: https://doi.org/10.1016/j.patcog.2023.109676.
Al-GhaferIAAlAfeshatNAlshomaliL, et al.Nmf-guided feature selection and genetic algorithm-driven framework for tumor mutational burden classification in bladder cancer using multi-omics data. Netw Model Anal Health Inform Bioinformat2024; 13: 26.
29.
ThawkarSSinghLKKhannaM. Multi-objective techniques for feature selection and classification in digital mammography. Intell Dec Technol2021; 15: 115–125. DOI: https://doi.org/10.3233/IDT-200049.
CaiHZhengVWChangKCC. A comprehensive survey of graph embedding: Problems, techniques, and applications. IEEE Trans Knowl Data Eng2018; 30: 1616–1637.
32.
LiJChengKWangS, et al.Feature selection: a data perspective, 2017. DOI: 10.1145/3136625.
33.
AlzubiOAAlzubiJAAlzubiTM, et al.Quantum mayfly optimization with encoder-decoder driven lstm networks for malware detection and classification model. Mob Netw Appl2023; 28: 795–807. DOI: https://doi.org/10.1007/s11036-023-02105-x.
34.
RingnérM. What is principal component analysis?, 2008. DOI: 10.1038/nbt0308-303.
35.
ZhuXWangYLiY, et al.A new unsupervised feature selection algorithm using similarity-based feature clustering. Comput Intell2019; 35: 2–22. DOI: https://doi.org/10.1111/coin.12192.
36.
KaramiSSaberi-MovahedFTiwariP, et al.Unsupervised feature selection based on variance–covariance subspace distance. Neural Netw2023; 166: 105877. DOI: https://doi.org/10.1016/j.neunet.2023.06.018.
37.
DashtbanMBalafarMSuravajhalaP. Gene selection for tumor classification using a novel bio-inspired multi-objective approach. Genomics2018; 110: 10–17. DOI: https://doi.org/10.1016/j.ygeno.2017.07.010.
38.
ShunmugapriyaPKanmaniS. A hybrid algorithm using ant and bee colony optimization for feature selection and classification (ac-abc hybrid). Swarm Evol Comput2017; 36: 27–36. DOI: https://doi.org/10.1016/j.swevo.2017.04.002.
39.
SalemHAttiyaGEl-FishawyN. Classification of human cancer diseases by gene expression profiles. Appl Soft Comput J2017; 50: 124–134. DOI: https://doi.org/10.1016/j.asoc.2016.11.026.
40.
KangCHuoYXinL, et al.Feature selection and tumor classification for microarray data using relaxed lasso and generalized multi-class support vector machine. J Theor Biol2019; 463: 77–91. DOI:https://doi.org/10.1016/j.jtbi.2018.12.010.
41.
KabirMMShahjahanMMuraseK. A new local search based hybrid genetic algorithm for feature selection. Neurocomputing2011; 74: 2914–2928. DOI: https://doi.org/10.1016/j.neucom.2011.03.034.
42.
WittenIHFrankEGellerJ. Data mining: practical machine learning tools and techniques with java implementations. SIGMOD Record2002; 31: 76–87. DOI: https://doi.org/10.1145/507338.507355.
43.
HuangA. Similarity measures for text document clustering. In: New Zealand computer science research student conference, NZCSRSC 2008 - Proceedings, 2008.
AzizRM. Nature-inspired metaheuristics model for gene selection and classification of biomedical microarray data. Med Biol Eng Comput2022; 60: 1627–1646. DOI: https://doi.org/10.1007/s11517-022-02555-7.
TangCZhengXZhangW, et al.Unsupervised feature selection via multiple graph fusion and feature weight learning. Sci China Inform Sci2023; 66: 152101. DOI: https://doi.org/10.1007/s11432-022-3579-1.
48.
ZhuXZhangSZhuY, et al.Unsupervised spectral feature selection with dynamic hyper-graph learning. IEEE Trans Knowl Data Eng2022; 34: 1–7. DOI: https://doi.org/10.1109/TKDE.2020.3017250.
49.
ZhangWShangRJiaoL. Large-scale community detection based on core node and layer-by-layer label propagation. Inf Sci (Ny)2023; 632: 1–18. DOI: https://doi.org/10.1016/j.ins.2023.02.090.
50.
RahmanianMMansooriEG. An unsupervised gene selection method based on multivariate normalized mutual information of genes. Chemometr Intell Lab Syst2022; 222: 104512. DOI: https://doi.org/10.1016/j.chemolab.2022.104512.
51.
BandyopadhyaySBhadraTMitraP, et al.Integration of dense subgraph finding with feature clustering for unsupervised feature selection. Pattern Recognit Lett2014; 40: 104–112. DOI: https://doi.org/10.1016/j.patrec.2013.12.008.
52.
Solorio-FernándezSCarrasco-OchoaJAMartínez-TrinidadJF. A review of unsupervised feature selection methods. Artif Intell Rev2020; 53: 907–948. DOI: https://doi.org/10.1007/s10462-019-09682-y.
53.
VapnikVN. The Nature of Statistical Learning Theory. New York, NY: Springer, 1995. DOI: 10.1007/978-1-4757-2440-0.
54.
ManningCDRaghavanPSchützeH. Introduction to Information Retrieval. Cambridge University Pres: Cambridge, 2008. DOI: 10.1017/cbo9780511809071.
55.
HastieTTibshiraniRFriedmanJ. The elements of statistical learning, second edition. New York, NY: Springer, 2009.
ArthurDVassilvitskiiS. K-means++: The advantages of careful seeding. In: Proceedings of the annual ACM-SIAM symposium on discrete algorithms, 2007, Volume 07-09-January-2007.
ShiJMalikJ. Normalized cuts and image segmentation. IEEE Trans Pattern Anal Mach Intell2000; 22: 888–905. DOI: https://doi.org/10.1109/34.868688.
61.
BriolaAAsteT. Topological feature selection. In: Doster T, Emerson T, Kvinge H et al. (eds.) Proceedings of 2nd Annual workshop on topology, algebra, and geometry in machine learning (TAG-ML), Vol. 221, pp.534–556. PMLR. https://proceedings.mlr.press/v221/briola23a.html.
62.
RoffoGMelziSCastellaniU, et al.Infinite feature selection: a graph-based feature filtering approach. IEEE Trans Pattern Anal Mach Intell2021; 43: 4396–4410. DOI: https://doi.org/10.1109/TPAMI.2020.3002843.
63.
WangCWangJGuZ, et al.Unsupervised feature selection by learning exponential weights. Pattern Recognit2024; 148: 110183.
64.
YouMYuanAHeD, et al.Unsupervised feature selection via neural networks and self-expression with adaptive graph constraint. Pattern Recognit2023; 135: 109173. DOI: https://doi.org/10.1016/j.patcog.2022.109173.
65.
FriedmanM. A comparison of alternative tests of significance for the problem of rankings. Ann Math Stat1940; 11: 86–92. DOI: https://doi.org/10.1214/aoms/1177731944.
66.
AlzubiOAAlzubiJAAlweshahM, et al.An optimal pruning algorithm of classifier ensembles: dynamic programming approach. Neur Comput Appl2020; 32: 16091–16107. DOI: https://doi.org/10.1007/s00521-020-04761-6.
