融合小波变换和非齐次泊松过程的开源软件故障预测模型
Open Source Software Fault Prediction Model Integrating Wavelet Transform and Non-Homogeneous Poisson Process
DOI: 10.12677/aam.2025.148385, PDF,    国家自然科学基金支持
作者: 顾子欣, 李柳英, 肖雪珊：南宁师范大学数学与统计学院，广西 南宁；杨剑锋*：南宁师范大学数学与统计学院，广西 南宁；南宁师范大学广西应用数学中心，广西 南宁
关键词: 开源软件；小波变换；非齐次泊松过程；故障预测；Open Source Software； Wavelet Transform； Non-Homogeneous Poisson Process； Failure Prediction
摘要: 开源软件凭借其开放源代码、社区协作开发等特性，拥有极为广泛的应用。但开源软件故障频发，传统基于非齐次泊松过程(NHPP)的预测方法在处理开源软件故障数据时，由于故障数据具有非平稳性，难以准确捕捉数据中的复杂特征和规律。为了提高软件故障预测效果，本文提出了一种融合小波变换和非齐次泊松过程的开源软件故障预测模型。利用小波变换提取故障数据多尺度特征，提取故障数据的低频趋势与高频波动分量，来克服传统模型缺乏多尺度分析的缺陷。结合Delayed S-Shaped (DSS)可叠加可靠性模型进行故障预测。以Tomcat 9开源软件故障数据为例，通过极大似然估计法对模型进行参数估计，用提出的模型与DSS模型进行性能对比。结果显示，本文提出的模型故障拟合效果优于DSS模型。
Abstract: Open source software has gained extensive applications due to its open-source code and community-driven development. However, frequent failures in such software pose challenges for traditional Non-Homogeneous Poisson Process (NHPP)-based prediction methods, which struggle to capture complex patterns in non-stationary failure data. To address this, this paper proposes a hybrid model integrating Wavelet Transform (WT) and NHPP for open source software failure prediction. The WT decomposes failure data into multi-scale components (low-frequency trends and high-frequency fluctuations) to overcome the limitations of single-scale NHPP models. By combining this with the Delayed S-Shaped (DSS) reliability model, the hybrid approach enables more accurate forecasting. Using Tomcat 9 failure data as a case study, parameters are estimated via maximum likelihood estimation, and performance is compared against the standalone DSS model. The results show that the failure fitting effect of the model proposed in this paper is better than that of the DSS model.
文章引用：顾子欣, 杨剑锋, 李柳英, 肖雪珊. 融合小波变换和非齐次泊松过程的开源软件故障预测模型[J]. 应用数学进展, 2025, 14(8): 227-238. https://doi.org/10.12677/aam.2025.148385

参考文献

[1] 陈静, 杨剑锋, 王喜宾, 等. NHPP类开源软件可靠性增长模型的极大似然估计[J]. 广西大学学报(自然科学版), 2022, 47(1): 174-184.
[2] Das, S., Kundu, D. and Dewanji, A. (2022) Software Reliability Modeling Based on NHPP for Error Occurrence in Each Fault with Periodic Debugging Schedule. Communications in Statistics—Theory and Methods, 51, 4890-4902. [Google Scholar] [CrossRef] 
[3] 吕勋. 基于NHPP的软件可靠性模型研究[D]: [硕士学位论文]. 哈尔滨: 哈尔滨工程大学, 2017.
[4] 周波. 基于NHPP软件可靠性模型的预测研究及实现[D]: [硕士学位论文]. 成都: 电子科技大学, 2016.
[5] Kim, Y.S., Song, K.Y. and Chang, I.H. (2023) Prediction and Comparative Analysis of Software Reliability Model Based on NHPP and Deep Learning. Applied Sciences, 13, Article 6730. [Google Scholar] [CrossRef] 
[6] Aggarwal, A., Kumar, S. and Gupta, R. (2024) Testing Coverage Based NHPP Software Reliability Growth Modeling with Testing Effort and Change-Point. International Journal of System Assurance Engineering and Management, 15, 5157-5166. [Google Scholar] [CrossRef] 
[7] Lin, J.C. and Zhu, A.M. (2022) CNN-Wavelet-Transform-Based Model for Solar Photovoltaic Power Prediction. 19th IEEE International Computer Conference on Wavelet Active Media Technology and Information Processing, Chengdu, 16-18 December 2022, 1-5.
[8] Serhal, H., Abdallah, N., Marion, J., Chauvet, P., Oueidat, M. and Humeau-Heurtier, A. (2022) Wavelet Transformation Approaches for Prediction of Atrial Fibrillation. 2022 30th European Signal Processing Conference (EUSIPCO), Belgrade, 29 August-2 September 2022, 1188-1192. [Google Scholar] [CrossRef] 
[9] Cheng, R., Liu, Z. and Zhang, M. (2022) An Improved Traffic Flow Prediction Model: Spatial-Temporal Network Based on Wavelet and LSTM. 2022 16th IEEE International Conference on Signal Processing (ICSP), Beijing, 21-24 October 2022, 276-280. [Google Scholar] [CrossRef] 
[10] Wang, Y.T., Liu, J., Li, R., Suo, X.Y. and Lu, E.H. (2022) Medium and Long-Term Precipitation Prediction Using Wavelet Decomposition-Prediction-Reconstruction Model. Water Resources Management, 36, 971-987. [Google Scholar] [CrossRef] 
[11] Yuan, M.X., Wei, S.K., Sun, M. and Zhao, J.D. (2022) Wavelet Decomposition and Seq2Seq Hybrid Models for Water Quality Prediction. Water Resources, 49, 743-752. [Google Scholar] [CrossRef] 
[12] Feng, Q.M. and Qian, S.P. (2022) Research on the Prediction of Short-Term Wind Power Based on Wavelet Neural Networks. Energy Reports, 8, 553-559. [Google Scholar] [CrossRef] 
[13] Yu, C.J., Li, Y.L., Chen, Q., Lai, X.P. and Zhao, L.Y. (2022) Matrix-Based Wavelet Transformation Embedded in Recurrent Neural Networks for Wind Speed Prediction. Applied Energy, 324, Article 119692. [Google Scholar] [CrossRef] 
[14] Nasser, A. and Simon, V. (2022) Wavelet‐Attention-Based Traffic Prediction for Smart Cities. IET Smart Cities, 4, 3-16. [Google Scholar] [CrossRef] 
[15] Samani, S., Vadiati, M., Delkash, M. and Bonakdari, H. (2023) A Hybrid Wavelet-Machine Learning Model for Qanat Water Flow Prediction. Acta Geophysica, 71, 1895-1913. [Google Scholar] [CrossRef] 
[16] Dincer, N.G., Yalcin, M.O. and Guneri, O.I. (2022) A Hybrid Time Series Prediction Model Based on Fuzzy Time Series and Maximal Overlap Discrete Wavelet Transform. Gazi University Journal of Science, 35, 1152-1169. [Google Scholar] [CrossRef] 

为你推荐