Research on the classification method of unbalanced data for the assessment of traffic accident disability level_Journal of Biomedical Engineering

Authors：

YAN Hubin ¹ ,  WANG Shaohua ^1,2 , JIA Jianlin ³ , QIN Wenhui ¹

1. School of Automotive and Transportation, Tianjin University of Technology and Education, Tianjin 300222, P. R. China;
2. Judicial Appraisal Institute, Tianjin Yize Testing Technology Co. Ltd., Tianjin 300210, P. R. China;
3. Inner Mongolia Autonomous Region Key Laboratory of Green Construction and Intelligent Operation and Maintenance of Civil Engineering, Inner Mongolia University of Technology, Huhhot 010051, P. R. China;

Corresponding?author：

WANG Shaohua, Email: shaohuawang@tute.edu.cn

Keywords：

Traffic accidents; Disability level; Loss function; Unbalanced data; Self-attention mechanism

DOI：

10.7507/1001-5515.202508042

Video：

Export PDF Favorites Scan Get Citation

Abstract Full text Figures/Tables Video References Cited by

The evaluation of disability grades in traffic accidents is a professional forensic clinical appraisal matter, and its results directly affect the fairness of judicial compensation. In the construction of automated disability grade evaluation models, the imbalanced distribution of disability cases leads to low recognition accuracy for minority categories, becoming a key bottleneck restricting the technology's implementation. In response, this paper proposes an imbalanced data classification method based on a hybrid parameter scaling weight optimization mechanism. First, a loss weight calculation model is constructed based on category proportion, category sparsity, and category diversity. Second, the loss weight calculation model is designed by integrating the focal loss function's ability to focus on hard samples with the cross-entropy loss function’s global gradient stability advantage. Then, at the early stages of training, the model proposed in this paper aligns sensitivity to imbalanced categories and constructs a low-computational-demand hybrid parameter scaling weight optimization mechanism. Experimental results show that, compared with the best-performing baseline methods, the proposed method significantly improves both accuracy and macro-F1 score on the traffic accident disability grade dataset. It can effectively enhance the classification performance of minority grade categories in imbalanced data and help improve the accuracy of automated appraisal in judicial identification of traffic accident disability grades.

Citation： YAN Hubin, WANG Shaohua, JIA Jianlin, QIN Wenhui. Research on the classification method of unbalanced data for the assessment of traffic accident disability level. Journal of Biomedical Engineering, 2026, 43(3): 607-614, 624. doi: 10.7507/1001-5515.202508042 Copy

1.	賀佳, 羅淑婧, 李冰清, 等. 得失之禍: 股市下跌對交通事故的溢出影響. 經濟學(季刊), 2025, 25(2): 394-409..
2.	Wang C, Chen F, Zhang Y, et al. Analysis of injury severity in rear-end crashes on an expressway involving different types of vehicles using random-parameters logit models with heterogeneity in means and variances. Transp Lett, 2023, 15(7): 742-753..
3.	俞曉英, 夏晴, 趙方宇, 等. 法醫臨床司法鑒定爭議意見及其形成原因的分析. 中國法醫學雜志, 2025, 40(1): 12-18..
4.	Musslick S, Bartlett L K, Chandramouli S H, et al. Automating the practice of science: opportunities, challenges, and implications. Proc Natl Acad Sci USA, 2025, 122(5): 240-251..
5.	Abdel-Aty M, Ding S. A matched case-control analysis of autonomous vs human-driven vehicle accidents. Nat Commun, 2024, 15(1): 31-49..
6.	王少華, 黃建玲, 陳艷艷, 等. 自行車駕駛員交通行為方式判定研究. 重慶交通大學學報(自然科學版), 2020, 39(6): 19-24..
7.	Xu Z Y, Shen Z F, Li Y, et al. Enhanced DeepLab algorithm and adaptive loss function for high-resolution remote sensing image classification. J Remote Sens, 2022, 26(2): 406-415..
8.	Hong B W. Research on pedestrian object detection model based on improved loss function. Electron Prod, 2024, 32(19): 54-56,112-120..
9.	Chen L F, Wei M R. 3D point cloud classification and segmentation network based on enhanced local features. J Appl Sci, 2022, 40(2): 328-337..
10.	Tak J H, Hong B W. Enhancing model agnostic meta-learning via gradient similarity loss. Electronics, 2024, 13(3): 5-35..
11.	Fernando K R M, Tsokos C P. Dynamically weighted balanced loss: class imbalanced learning and confidence calibration of deep neural networks. IEEE Trans Neural Netw Learn Syst, 2022, 33(7): 2940-2951..
12.	Fan S, Zhang X, Song Z. Imbalanced sample selection with deep reinforcement learning for fault diagnosis. IEEE Trans Ind Inf, 2022, 18(4): 2518-2527..
13.	Li Q, Gao X, Lu H, et al. A meta-learning imbalanced classification framework via boundary enhancement strategy with Bayes imbalance impact index. Neural Netw, 2025, 185: 107126..
14.	Lian Z, Li S, Huang Q, et al. UBMF: uncertainty-aware bayesian meta-learning framework for fault diagnosis with imbalanced industrial data. Knowl Based Syst, 2025, 323: 113772..
15.	Qian L, Zuo Q, Li D, et al. DGMSCL: a dynamic graph mixed supervised contrastive learning approach for class imbalanced multivariate time series classification. Neural Netw, 2025, 185: 107131..
16.	Li Y, Zhao X, Zeng Y, et al. Node transfer with graph contrastive learning for class-imbalanced node classification. Neural Netw, 2025, 190: 107674..
17.	Cabezas M, Diez Y. An analysis of loss functions for heavily imbalanced lesion segmentation. Sensors, 2024, 24(6): 1981..
18.	Johnson J M, Khoshgoftaar T M. Survey on deep learning with class imbalance. J Big Data, 2019, 6(1): 27..
19.	Razali M N, Arbaiy N, Lin P C, et al. Optimizing multiclass classification using convolutional neural networks with class weights and early stop for imbalanced datasets. Electronics, 2025, 14(4): 705..
20.	Yousef M, Voskergian D. TextNetTopics: text classification based word grouping as topics and topics’ scoring. Front Genet, 2022, 13: 907378..
21.	He L, Zhou Y, Liu L, et al. Research and application of deep learning object detection methods for forest fire smoke recognition. Sci Rep, 2025, 15(1): 16628..
22.	Carvalho M, Pinho A J, Brás S. Resampling approaches to handle class imbalance: a review from a data perspective. J Big Data, 2025, 12(1): 71..
23.	Yang Z, Liu Y. A steel surface defect detection method based on improved RetinaNet. Sci Rep, 2025, 15(1): 14560..
24.	Ding H, Huang N, Cui X. Leveraging GANs data augmentation for imbalanced medical image classification. Appl Soft Comput, 2024, 165: 112050..
25.	Bai J, Shi W, Xiao Z, et al. Achieving better category separability for hyperspectral image classification: a spatial-spectral approach. IEEE Trans Neural Netw Learn Syst, 2023, 35(7): 9621-9635..
26.	Prasad A, Mohammad Alenazy W, Ahmad N, et al. Optimizing IoT intrusion detection with cosine similarity based dataset balancing and hybrid deep learning. Sci Rep, 2025, 15(1): 13039..
27.	Huang J, Wu W, Hu D, et al. Distributed swarm learning for communication-efficient capsule endoscopy classification on solving Non-iid data. Alex Eng J, 2025, 128: 92-99..
28.	Zhao W, Li W, Li Y, et al. Constructing balanced training samples: a new perspective on long-tailed classification. IEEE Trans Multimedia, 2025, 16(28): 72-79..
29.	LeCun Y, Bengio Y, Hiton G. Deep learning. Nature, 2015, 524(7553): 436-444..
30.	Li J, Yuan Y, Luo X. Learning error refinement in stochastic gradient descent-based latent factor analysis via diversified pid controllers. IEEE Trans Emerg Top Comput Intell, 2025, 41(6): 236-264..
31.	Lin M, Yang H, Xu X, et al. Momentum-accelerated and biased unconstrained non-negative latent factor model for handling high-dimensional and incomplete data. ACM Trans Knowl Discov Data, 2025, 19(3): 1-25..
32.	Xiong N, Ding L, Du S S. How over-parameterization slows down gradient descent in matrix sensing: the curses of symmetry and initialization. arXiv, 2023: 2310.01769..
33.	Zhu W, Chen X, Wu W B. Online covariance matrix estimation in stochastic gradient descent. J Am Stat Assoc, 2023, 118(541): 393-404..
34.	Luo X, Chen J, Yuan Y, et al. Pseudo gradient-adjusted particle swarm optimization for accurate adaptive latent factor analysis. IEEE Trans Syst Man Cybern Syst, 2024, 54(4): 2213-2226..
35.	Meng F, Wang Z, Zhang M. Pissa: principal singular values and singular vectors adaptation of large language models. Adv Neural Inf Process Syst, 2024, 37: 121038-121072..
36.	Zhou L, Ran G, Tan H, et al. Asymmetric gradient penalty based on power exponential function for imbalanced data classification. Complex Intell Syst, 2024, 10(1): 1333-1348..
37.	Geiersbach C, Scarinci T. Stochastic proximal gradient methods for nonconvex problems in Hilbert spaces. Comput Optim Appl, 2021, 78(3): 705-740..
38.	Shen S, Hong T, Shen Y, et al. RMAAC: joint Markov games and robust multiagent actor-critic for explainable malware defense in social IoT. IEEE Trans Dependable Secure Comput, 2025, 22(6): 7091-7106..
39.	Lu S, Yang H. Convergence analysis of log-linear learning of games with irrationality. IEEE Trans Syst Man Cybern Syst, 2025, 55(9): 5013-5017..
40.	Iannelli A. Adaptive control mechanisms in gradient descent algorithms//2025 IEEE 64th Conference on Decision and Control, Rio de Janeiro: IEEE, 2025: 1664-1669..
41.	Xie Z, Tian W, Li R, et al. Horse-MinHash: high-performance and secure jaccard similarity estimation for cloud storage. IEEE Trans Dependable Secure Comput, 2025, 22(6): 6685-6697..
42.	Wang Z, Xu Q, Yang Z, et al. Top-K pairwise ranking: bridging the gap among ranking-based measures for multi-label classification. Int J Comput Vis, 2025, 133(1): 211-253..
43.	Yang J, Lu Y, Zhang Z, et al. A deep learning method coupling a channel attention mechanism and weighted dice loss function for water extraction in the Yellow River Basin. Water, 2025, 17(4): 511..

1. 賀佳, 羅淑婧, 李冰清, 等. 得失之禍: 股市下跌對交通事故的溢出影響. 經濟學(季刊), 2025, 25(2): 394-409..
2. Wang C, Chen F, Zhang Y, et al. Analysis of injury severity in rear-end crashes on an expressway involving different types of vehicles using random-parameters logit models with heterogeneity in means and variances. Transp Lett, 2023, 15(7): 742-753..
3. 俞曉英, 夏晴, 趙方宇, 等. 法醫臨床司法鑒定爭議意見及其形成原因的分析. 中國法醫學雜志, 2025, 40(1): 12-18..
4. Musslick S, Bartlett L K, Chandramouli S H, et al. Automating the practice of science: opportunities, challenges, and implications. Proc Natl Acad Sci USA, 2025, 122(5): 240-251..
5. Abdel-Aty M, Ding S. A matched case-control analysis of autonomous vs human-driven vehicle accidents. Nat Commun, 2024, 15(1): 31-49..
6. 王少華, 黃建玲, 陳艷艷, 等. 自行車駕駛員交通行為方式判定研究. 重慶交通大學學報(自然科學版), 2020, 39(6): 19-24..
7. Xu Z Y, Shen Z F, Li Y, et al. Enhanced DeepLab algorithm and adaptive loss function for high-resolution remote sensing image classification. J Remote Sens, 2022, 26(2): 406-415..
8. Hong B W. Research on pedestrian object detection model based on improved loss function. Electron Prod, 2024, 32(19): 54-56,112-120..
9. Chen L F, Wei M R. 3D point cloud classification and segmentation network based on enhanced local features. J Appl Sci, 2022, 40(2): 328-337..
10. Tak J H, Hong B W. Enhancing model agnostic meta-learning via gradient similarity loss. Electronics, 2024, 13(3): 5-35..
11. Fernando K R M, Tsokos C P. Dynamically weighted balanced loss: class imbalanced learning and confidence calibration of deep neural networks. IEEE Trans Neural Netw Learn Syst, 2022, 33(7): 2940-2951..
12. Fan S, Zhang X, Song Z. Imbalanced sample selection with deep reinforcement learning for fault diagnosis. IEEE Trans Ind Inf, 2022, 18(4): 2518-2527..
13. Li Q, Gao X, Lu H, et al. A meta-learning imbalanced classification framework via boundary enhancement strategy with Bayes imbalance impact index. Neural Netw, 2025, 185: 107126..
14. Lian Z, Li S, Huang Q, et al. UBMF: uncertainty-aware bayesian meta-learning framework for fault diagnosis with imbalanced industrial data. Knowl Based Syst, 2025, 323: 113772..
15. Qian L, Zuo Q, Li D, et al. DGMSCL: a dynamic graph mixed supervised contrastive learning approach for class imbalanced multivariate time series classification. Neural Netw, 2025, 185: 107131..
16. Li Y, Zhao X, Zeng Y, et al. Node transfer with graph contrastive learning for class-imbalanced node classification. Neural Netw, 2025, 190: 107674..
17. Cabezas M, Diez Y. An analysis of loss functions for heavily imbalanced lesion segmentation. Sensors, 2024, 24(6): 1981..
18. Johnson J M, Khoshgoftaar T M. Survey on deep learning with class imbalance. J Big Data, 2019, 6(1): 27..
19. Razali M N, Arbaiy N, Lin P C, et al. Optimizing multiclass classification using convolutional neural networks with class weights and early stop for imbalanced datasets. Electronics, 2025, 14(4): 705..
20. Yousef M, Voskergian D. TextNetTopics: text classification based word grouping as topics and topics’ scoring. Front Genet, 2022, 13: 907378..
21. He L, Zhou Y, Liu L, et al. Research and application of deep learning object detection methods for forest fire smoke recognition. Sci Rep, 2025, 15(1): 16628..
22. Carvalho M, Pinho A J, Brás S. Resampling approaches to handle class imbalance: a review from a data perspective. J Big Data, 2025, 12(1): 71..
23. Yang Z, Liu Y. A steel surface defect detection method based on improved RetinaNet. Sci Rep, 2025, 15(1): 14560..
24. Ding H, Huang N, Cui X. Leveraging GANs data augmentation for imbalanced medical image classification. Appl Soft Comput, 2024, 165: 112050..
25. Bai J, Shi W, Xiao Z, et al. Achieving better category separability for hyperspectral image classification: a spatial-spectral approach. IEEE Trans Neural Netw Learn Syst, 2023, 35(7): 9621-9635..
26. Prasad A, Mohammad Alenazy W, Ahmad N, et al. Optimizing IoT intrusion detection with cosine similarity based dataset balancing and hybrid deep learning. Sci Rep, 2025, 15(1): 13039..
27. Huang J, Wu W, Hu D, et al. Distributed swarm learning for communication-efficient capsule endoscopy classification on solving Non-iid data. Alex Eng J, 2025, 128: 92-99..
28. Zhao W, Li W, Li Y, et al. Constructing balanced training samples: a new perspective on long-tailed classification. IEEE Trans Multimedia, 2025, 16(28): 72-79..
29. LeCun Y, Bengio Y, Hiton G. Deep learning. Nature, 2015, 524(7553): 436-444..
30. Li J, Yuan Y, Luo X. Learning error refinement in stochastic gradient descent-based latent factor analysis via diversified pid controllers. IEEE Trans Emerg Top Comput Intell, 2025, 41(6): 236-264..
31. Lin M, Yang H, Xu X, et al. Momentum-accelerated and biased unconstrained non-negative latent factor model for handling high-dimensional and incomplete data. ACM Trans Knowl Discov Data, 2025, 19(3): 1-25..
32. Xiong N, Ding L, Du S S. How over-parameterization slows down gradient descent in matrix sensing: the curses of symmetry and initialization. arXiv, 2023: 2310.01769..
33. Zhu W, Chen X, Wu W B. Online covariance matrix estimation in stochastic gradient descent. J Am Stat Assoc, 2023, 118(541): 393-404..
34. Luo X, Chen J, Yuan Y, et al. Pseudo gradient-adjusted particle swarm optimization for accurate adaptive latent factor analysis. IEEE Trans Syst Man Cybern Syst, 2024, 54(4): 2213-2226..
35. Meng F, Wang Z, Zhang M. Pissa: principal singular values and singular vectors adaptation of large language models. Adv Neural Inf Process Syst, 2024, 37: 121038-121072..
36. Zhou L, Ran G, Tan H, et al. Asymmetric gradient penalty based on power exponential function for imbalanced data classification. Complex Intell Syst, 2024, 10(1): 1333-1348..
37. Geiersbach C, Scarinci T. Stochastic proximal gradient methods for nonconvex problems in Hilbert spaces. Comput Optim Appl, 2021, 78(3): 705-740..
38. Shen S, Hong T, Shen Y, et al. RMAAC: joint Markov games and robust multiagent actor-critic for explainable malware defense in social IoT. IEEE Trans Dependable Secure Comput, 2025, 22(6): 7091-7106..
39. Lu S, Yang H. Convergence analysis of log-linear learning of games with irrationality. IEEE Trans Syst Man Cybern Syst, 2025, 55(9): 5013-5017..
40. Iannelli A. Adaptive control mechanisms in gradient descent algorithms//2025 IEEE 64th Conference on Decision and Control, Rio de Janeiro: IEEE, 2025: 1664-1669..
41. Xie Z, Tian W, Li R, et al. Horse-MinHash: high-performance and secure jaccard similarity estimation for cloud storage. IEEE Trans Dependable Secure Comput, 2025, 22(6): 6685-6697..
42. Wang Z, Xu Q, Yang Z, et al. Top-K pairwise ranking: bridging the gap among ranking-based measures for multi-label classification. Int J Comput Vis, 2025, 133(1): 211-253..
43. Yang J, Lu Y, Zhang Z, et al. A deep learning method coupling a channel attention mechanism and weighted dice loss function for water extraction in the Yellow River Basin. Water, 2025, 17(4): 511..

Previous Article
Gait recognition based on feature-level fusion of motion posture and surface electromyography
Next Article
Research progress on the neuromodulation targets in stroke rehabilitation

Journal of Biomedical Engineering

Research on the classification method of unbalanced data for the assessment of traffic accident disability level

Abstract Full text Figures/Tables Video References Cited by

Previous Article

Next Article

Format

Content