Author(s)

Manli Liu ¹, Wenhan Fu ^1,2, Junxiang Li ^1,2, Qiuyang Wang ¹

Affiliation(s)

¹ Business School, University of Shanghai for Science and Technology, Shanghai, 200093, China
² School of Intelligent Emergency Management, University of Shanghai for Science and Technology, Shanghai, 200093, China

Corresponding Author

Wenhan Fu

ABSTRACT

To address the challenges of complex quality evolution mechanisms and the difficulty in characterizing prediction uncertainty in multi-stage manufacturing processes, this study proposes a deep neural network-based multi-stage quality prediction method using MC Dropout for approximate Bayesian inference. First, systematic data preprocessing is performed, with SMOTE used for data augmentation and stage-wise PCA for dimensionality reduction. Subsequently, the Dropout mechanism is incorporated into the deep neural network, and MC Dropout is employed to jointly model the predictive mean and uncertainty. A five-fold cross-validation strategy is adopted for model training and evaluation, and the effectiveness of the proposed method is validated using real-world TFT-LCD manufacturing data. Experimental results demonstrate that the proposed method achieves superior prediction accuracy while providing uncertainty estimates for multi-stage quality prediction.

KEYWORDS

Multi-stage quality prediction; Deep neural network; MC Dropout; Bayesian inference; Uncertainty modeling

CITE THIS PAPER

Manli Liu, Wenhan Fu, Junxiang Li, Qiuyang Wang. A Bayesian Deep Neural Network with MC Dropout for Multi-stage Quality Prediction. Industrial Engineering and Innovation Management (2026). Vol. 9, No. 1, 105-112. DOI: http://dx.doi.org/10.23977/ieim.2026.090112.

REFERENCES

[1] Zhao Xiaoqiang, Liu Yongyong, Hui Yongyong, et al. Quality prediction model for batch processes based on improved temporal convolutional network and multi-head self-attention mechanism [J]. Computer Applications, 2025,45(07):2245-2252.
[2] Ma Chao, Weng Zhiyi, He Fei. Quality prediction modeling of piezoelectric ceramic sintering process based on ensemble learning [J]. Computer Integrated Manufacturing Systems, 2025,31(01):147-157. 
[3] Wanda Z, Yanchao Y, Jun T, et al. A method for the spatiotemporal correlation prediction of the quality of multiple operational processes based on S-GGRU [J]. Advanced Engineering Informatics, 2023,58.
[4] Xinyi X, Clarke W, Carter H, et al. Quality prediction and control in wire arc additive manufacturing via novel machine learning framework [J]. Micromachines, 2022,13(1):137.
[5] Qu D, Liang W, Zhang Y, et al. Research on machining quality prediction method based on machining error transfer network and grey neural network [J]. Journal of Manufacturing and Materials Processing, 2024,8(5):203.
[6] Guo Ziyu, Wang Ning, Wang Jianjun, et al. Research on interpretable product quality prediction based on probabilistic distribution attention mechanism [J/OL]. Chinese Journal of Management Science, 2026. 
[7] Ding Long, Fu Wenhan, He Yuchun, et al. A wafer manufacturing cycle time prediction method based on interpretable deep learning [J/OL]. Industrial Engineering Journal, 2026. 
[8] Yang Jianxin, Lan Xiaoping, Yao Zhiqiang, et al. Multi-stage quality prediction method based on LSSVM optimized by coyote optimization algorithm [J]. Manufacturing Automation, 2021,43(12):105-109.
[9] Mowbray M, Kay H, Kay S, et al. Probabilistic machine learning based soft-sensors for product quality prediction in batch processes [J]. Chemometrics and Intelligent Laboratory Systems, 2022,228:162-168.
[10] Wang Chengtang, Wang Hao, Qin Weimin. Collapse probability evaluation of deep foundation pits in subway stations based on polymorphic fuzzy Bayesian networks [J]. Rock and Soil Mechanics, 2020,41(05):1670-1679+1689.
[11] Li Baoluo, Xu Shiyun, Li Zonghan, et al. Interval prediction method of critical short-circuit ratio for renewable energy multi-station systems based on Bayesian deep learning [J]. Proceedings of the CSEE, 2024,44(14):5451-5463. 
[12] Zhu Saihua, He Zitong, Feng Ye, et al. Prediction method of cigarette sensory evaluation indicators based on Bayesian regularized neural network with feature fusion [J]. Computer Integrated Manufacturing Systems, 2026,32(02):628-640. 
[13] Gal Y, Ghahramani Z. Dropout as a Bayesian approximation: Representing model uncertainty in deep learning [C]//International Conference on Machine Learning. PMLR, 2016:1050-1059.

Sponsors, Associates, and Links

---

• Information Systems and Economics
  
  
• Accounting, Auditing and Finance
  
  
• Tourism Management and Technology Economy
  
  
• Journal of Computational and Financial Econometrics
  
  
• Financial Engineering and Risk Management
  
  
• Accounting and Corporate Management
  
  
• Social Security and Administration Management
  
  
• Population, Resources & Environmental Economics
  
  
• Statistics & Quantitative Economics
  
  
• Agricultural & Forestry Economics and Management
  
  
• Social Medicine and Health Management
  
  
• Land Resource Management
  
  
• Information, Library and Archival Science
  
  
• Journal of Human Resource Development
  
  
• Manufacturing and Service Operations Management
  
  
• Operational Research and Cybernetics