Boruta is a feature selection algorithm based on random forests, whose core principle involves introducing shadow features to evaluate the importance of original features, thereby obtaining a subset of features with higher importance than their shadow counterparts. However, Boruta’s shadow feature generation mechanism involves randomly shuffling the original features, exhibiting significant randomness and uncertainty. Consequently, focusing on the shadow feature generation mechanism, a regularised autoencoder perturbation feature selection algorithm (RAE-PFS) injected with Gaussian noise was proposed. The proposed algorithm learned structural information from original features via a regularised autoencoder, encoded the original features into the latent space latent space, introduced Gaussian noise to perturb the latent space, and decoded them to generate shadow features. After merging with original features, a feature subset with higher importance than shadow features was obtained using a random forest. Experiments across eight public datasets and five classifiers demonstrate that the proposed RAE-PFS reduces the feature selection rate by 8.34% compared to traditional Boruta, while significantly improving classification accuracy, precision, and recall. Finally, interpretability analysis based on SHAP was performed on medical datasets to verify the rationality of the features selected by RAE-PFS.
OmuyaE O, OkeyoG O, KimweleM W. Feature selection for classification using principal component analysis and information gain[J]. Expert Systems with Applications, 2021, 174: 114765.
[2]
BarbieriM C, GrisciB I, DornM. Analysis and comparison of feature selection methods towards performance and stability[J]. Expert Systems with Applications, 2024, 249: 123667.
[3]
LiJ, ChengK, WangS, et al. Feature selection: a data perspective[J]. ACM Computing Surveys (CSUR), 2017, 50(6): 1-45.
[4]
DhalP, AzadC. A comprehensive survey on feature selection in the various fields of machine learning[J]. Applied Intelligence, 2022, 52(4): 4543-4581.
[5]
KursaM B, JankowskiA, RudnickiW R. Boruta—a system for feature selection[J]. Fundamenta Informaticae, 2010, 101(4): 271-285.
LiChao, LiMing, LiYuanhua, et al. Short-term prediction of offshore wind power based on Boruta-SSA-CKCformer[J]. Science Technology and Engineering, 2025, 25(23): 9840-9850.
[8]
ZhouH, XinY, LiS. A diabetes prediction model based on Boruta feature selection and ensemble learning[J]. BMC Bioinformatics, 2023, 24(1): 224.
[9]
GholamiH, MohammadifarA, GolzariS, et al. Using the Boruta algorithm and deep learning models for mapping land susceptibility to atmospheric dust emissions in Iran[J]. Aeolian Research, 2021, 50: 100682.
[10]
SzulT, TaborS, PancerzK. Application of the Boruta algorithm to input data selection for a model based on rough set theory (RST) to prediction energy consumption for building heating[J]. Energies, 2021, 14(10): 2779.
NieBin. Research on partial least squares structural equation model analysis method for pharmacodynamic substances and effects of traditional Chinese medicine formulas[D]. Nanchang: Jiangxi University of Chinese Medicine, 2024.
[13]
YuanX, ChenF, XiaZ, et al. A novel feature susceptibility approach for a PEMFC control system based on an improved XGBoost-Boruta algorithm[J]. Energy and AI, 2023, 12: 100229.
HuangHao, CheHengda, KongXiangwei, et al. Factors affecting the temporary plugging effect of tight sandstone gas based on random forest and SHAP algorithm[J]. Science Technology and Engineering, 2025, 25(26): 11135-11143.
LiQiang, ZhangLimei, BaiMuke. Identification of abnormal data in smart distribution network based on multivariate data features and im proved random forest[J]. Science Technology and Engineering, 2023, 23(5): 2007-2015.
LiP, PeiY, LiJ. A comprehensive survey on design and application of autoencoder in deep learning[J]. Applied Soft Computing, 2023, 138: 110176.
[20]
MengQ, CatchpooleD, SkillicornD, et al. Relational autoencoder for feature extraction[C]// Proceedings of the 2017 International Joint Conference on Neural Networks (IJCNN). Piscataway, NJ: IEEE, 2017: 364-371.
[21]
MienyeI D, SwartT G. Deep autoencoder neural networks: a comprehensive review and new perspectives[J]. Archives of Computational Methods in Engineering, 2025, 32(1): 1-20.
[22]
ManjunathaB A, ShastryK A, NareshE, et al. A network intrusion detection framework on sparse deep denoising auto-encoder for dimensionality reduction[J]. Soft Computing, 2024, 28(5): 4503-4517.
[23]
MajumdarA. Blind denoising autoencoder[J]. IEEE Transactions on Neural Networks and Learning Systems, 2018, 30(1): 312-317.
[24]
FanC, XiaoF, ZhaoY, et al. Analytical investigation of autoencoder-based methods for unsupervised anomaly detection in building energy data[J]. Applied Energy, 2018, 211: 1123-1135.
[25]
AlainG, BengioY. What regularized auto-encoders learn from the data-generating distribution[J]. The Journal of Machine Learning Research, 2014, 15(1): 3563-3593.
[26]
LuoW, LiJ, YangJ, et al. Convolutional sparse autoencoders for image classification[J]. IEEE Transactions on Neural Networks and Learning Systems, 2017, 29(7): 3289-3294.
[27]
PuY, GanZ, HenaoR, et al. Variational autoencoder for deep learning of images, abels and captions[J]. Advances in Neural Information Processing Systems, 2016, 29: 2352-2360.
[28]
ShenC, QiY, WangJ, et al. An automatic and robust features learning method for rotating machinery fault diagnosis based on contractive autoencoder[J]. Engineering Applications of Artificial Intelligence, 2018, 76: 170-184.
[29]
ZhouH F, ZhangJ W, ZhouY Q, et al. A feature selection algorithm of decision tree based on feature weight[J]. Expert Systems with Applications, 2021, 164: 113842.
[30]
ZengZ, ZhangH, ZhangR, et al. A novel feature selection method considering feature interaction[J]. Pattern Recognition, 2015, 48(8): 2656-2666.
[31]
PedregosaF, VaroquauxG, GramfortA, et al. Scikit-learn: machine learning in Python[J]. Journal of Machine Learning Research, 2011, 12: 2825-2830.
[32]
FriedmanM. A comparison of alternative tests of significance for the problem of m rankings[J]. The Annals of Mathematical Statistics, 1940, 11(1): 86-92.
[33]
KorkmazS, ŞahmanM A, CinarA C, et al. Boosting the oversampling methods based on differential evolution strategies for imbalanced learning[J]. AApplied Soft Computing, 2021, 112:107787.
[34]
LiuZ, DuJ, LuoJ, et al. A stable feature selection method based on majority voting and SHAP for high-dimensional metabolomics data[J]. Computer Methods and Programs in Biomedicine, 2025, 265: 109170.
ZhouYinjuan, WangWen, WeiHaiping, Assessment of disease burden and trend prediction of ischemic stroke related to high fasting plasma glucose in China: prevention and control strategies for key populations—a public health early warning based on GBD 2021[J]. Journal of Army Medical University, 2026, 48(2): 181-190.
[37]
HongR, XuB. Breast cancer: an up-to-date review and future perspectives[J]. Cancer Communications, 2022, 42(10): 913-936.