Table 1 Comparative summary of related existing studies.
Authors | Year | Method | Results | Limitation |
|---|---|---|---|---|
Yuan et al.19 | 2020 | SVM, RF | An accuracy of 100% | The algorithms used were not explicitly designed to address the unique challenges of gene expression data |
Wang et al.20 | 2018 | PDX | Â | They didn't use the latest classifiers, and they used small samples |
Danaee et al.21 | 2017 | SDAE | An accuracy of 98.26% | The study does not explore using different optimization techniques or hyperparameter tuning to improve deep learning performance |
Jia et al.22 | 2021 | WGCNA | An accuracy of 97.36% | The study did not use hybrid classifiers based on deep learning, which is optimized using metaheuristic hyperparameter optimizers |
Elbashir et al.41 | 2019 | lightweight CNN | An accuracy of 98.76 | The study does not explore using different optimization techniques or hyperparameter tuning to improve deep learning performance |
Alshareef et al.23 | 2022 | AIFSDL-PCD | An accuracy of 97.19% | Their study did not utilize a hybrid DL-based classifier with metaheuristics-based hyperparameter optimizers |
MotieGhader et al.33 | 2020 | PSO, ACO, ICA, WCC, LCA, GA, LA, FOA, DSOS, HTS, and CUK, with an SVM classifier | An accuracy of 90% | The study used a relatively small dataset |
Wei et al.42 | 2022 | GANs | An accuracy of 92.6% | The study did not use a combination of deep learning-based classifiers and metaheuristic-based hyperparameter optimizers |
Deng et al.43 | 2022 | XGBoost-MOGA | An accuracy of 56.67% | The number of samples used is less than 300 |
Houssein et al.44 | 2021 | BMO-SVM | An accuracy of 99.36% | The study utilized small datasets, which may limit the generalizability of the findings |
Devi et al.45 | 2023 | IWOA | An accuracy of 97.7% | They evaluate their model on a small dataset (128 samples) |
