An Explainable Neural Oblivious Decision Ensemble Learning for Baldness Prediction Using Multidimensional Risk Factors
International Journal of Medical Toxicology and Forensic Medicine,
Vol. 16 (2026),
1 January 2026
,
Page 1-9
https://doi.org/10.22037/ijmtfm.v16.53282
Abstract
Background: Baldness is characterized by significant hair loss on the scalp and can be influenced by factors such as age, lifestyle, genetics, and personal habits. Explainable artificial intelligence methods can help address this issue by showing the factors that contribute to individual predictions. Therefore, this study proposes an Explainable Risk-Based Baldness Prediction model using Neural Oblivious Decision Ensembles (XRB-NODE). The proposed model is an ensemble-based framework for baldness prediction that incorporates multidimensional risk factors, including genetic, medical, lifestyle, nutritional, environmental, and demographic characteristics.
Methods: An Improved ReliefF method is applied to select the most relevant factors associated with baldness classification. Subsequently, the selected features are used as input to the neural oblivious decision ensembles model, which categorizes individuals by baldness risk. Moreover, the model is trained using the AdaBelief optimizer to improve learning performance. In addition, SHAP is integrated to explain the model's prediction by showing how individual risk factors contribute to the predicted baldness category.
Results: The proposed XRB-NODE methodology was evaluated on the Baldness Risk Factor Dataset and achieved 95.06% accuracy in baldness risk classification.
Conclusion: The proposed XRB-NODE method offers accurate and interpretable baldness risk prediction, supporting modified, faster risk assessment and preventive decision-making.
- Baldness risk classification, Explainable artificial intelligence, Neural oblivious decision ensembles, Deep learning, AdaBelief optimizer
How to Cite
References
[1] Li M, Wang X, Wang Q, Fu C, Shrestha N, Virani SS, et al. Male sex hormone concentrations, puberty timing, baldness patterns, and risk of cardiovascular disease. Hormones. 2025;24(3):701-10. [DOI: 10.1007/s42000-025-00655-1]
[2] [2] Sambandam RP, Kaliyaperumal D, Dileep JE. Androgenetic alopecia in men: an update on genetics. Indian J Dermatol. 2024;69(3):282. [DOI: 10.4103/ijd.ijd_729_23]
[3] [3] Maben M, Kumar A, Rao B. Hair Baldness Progression Prediction Using XGBoost: Risk Forecasting Model. In: 2026 International Conference on Artificial Intelligence and Data Engineering (AIDE). IEEE; 2026. p. 614-9. [DOI: 10.1109/AIDE69088.2026.11544272]
[4] Xi H, Yuan X, Yuan H, Lin K, Wang C, Wang Z, et al. Enhanced stratification of male pattern hair loss using AI through novel loss region ratio analysis. Sci Rep. 2025;15(1):38280. [DOI: 10.1038/s41598-025-23561-3]
[5] Kim J, Song SY, Sung JH. Recent advances in drug development for hair loss. Int J Mol Sci. 2025;26(8):3461. [DOI: 10.3390/ijms26083461]
[6] [6] Dachawar M, Sampathi S, Ladkat VV, Tong VB, Gaikwad VB, Bhagwat OE. Innovative approaches to baldness detection and management with artificial intelligence and machine learning. Arch Dermatol Res. 2026;318(1):36. [DOI: 10.1007/s00403-025-04477-4]
[7] Zaman S, Perven S, Moni JB, Rahmat NS, Chowdhury M. A Deep Learning and XAI-Based Alopecia Disorder Prediction Scheme Using Temperature-Weighted Softmax Ensemble Architectures. In: 2026 International Conference on Data Science and Business Systems (ICDSBS). Vol 3. IEEE; 2026. p. 1-6. [DOI: 10.1109/ICDSBS69077.2026.11634775]
[8] Pinon MU, Emon AS, Tafsir M, Hassan MM, Shah MS. An Innovative Ensemble Deep Learning Method with Explainable AI for Resilient Multi-Class Hair and Scalp Disease Classification. In: 2026 IEEE 2nd International Conference on Quantum Photonics, Artificial Intelligence & Networking (QPAIN). IEEE; 2026. p. 1-6. [DOI: 10.1109/QPAIN69676.2026.11546472]
[9] Behal R, Priya P, Kapoor A, Mishra A. Hair loss stage prediction using deep learning. In: 2024 IEEE International Conference on Computing, Power and Communication Technologies (IC2PCT). Vol 5. IEEE; 2024. p. 1263-7. [DOI: 10.1109/IC2PCT60090.2024.10486442]
[10] Putri FP, Al-Hadi IAA, Ali A, Ratnavelu K. Enhancing ReliefF for Multi-Label Text Classification via High-Order Label Correlation. Results Eng. 2026;30:109411. [DOI: 10.1016/j.rineng.2026.109411]
[11] Wang H, Tian Z, Wang L, Wang H, Zhang Y, Ban W. Hybrid computational model for analysis of drug release from biomaterial matrix: machine learning-aided material design. Case Stud Therm Eng. 2025;73:106505. [DOI: 10.1016/j.csite.2025.106505]
[12] Altınel D. Development of deep learning optimizers: Approaches, concepts, and update rules. arXiv preprint arXiv:2509.18396. 2025. [DOI: 10.48550/arXiv.2509.18396]
[13] Baytar O, Zontul M, Orak C, Karateke S, Aydın H, Horoz S. Interpretable Machine Learning Approach for Photocatalytic Degradation in Mn-Doped Semiconductors Using Multilayer Perceptron and SHAP Analysis. Catalysts. 2026;16(6):530. [DOI: 10.3390/catal16060530]
[14] Baldness Risk Factor Dataset [Internet]. Kaggle; 2026 [cited 2026 Oct 1]. [Link]
[15] Chen A, Shang L, Ju Y, Zhuo F. Early prediction of alopecia areata using machine learning modeling of neuro stress immune signatures from multi datasets. Sci Rep. 2025;16(1):3823. [DOI: 10.1038/s41598-025-33927-2]
[16] Ahmad M, Mir A, Permana A. An Early Hair Loss Detection and Prediction Method Based on Machine Learning. Eng Proc. 2025;107(1):126. [DOI: 10.3390/engproc2025107126]
- Abstract Viewed: 26 times
- 53282 Downloaded: 28 times