Safety Assessment Method Based on Degree of Feature Matching and Fusion of Heterogeneous Sub-Models

doi:10.19678/j.issn.1000-3428.0068028

Abstract

Abstract:

The quality of machine learning models affects the prediction accuracy as well as the fit between the input and output results. In complex systems, when a single model is used to evaluate system security problems, the results are easily affected by data volume, data format, model structure, environmental interference, and other factors. These issues reduce the effectiveness of the model in simultaneously solving multiple problems, although being effective in solving one problem at a time, leading to unsatisfactory results. To address these issues, this study proposes a safety assessment method based on degree of feature matching and heterogeneous sub-model fusion. First, the datasets are divided into different sizes based on the output values of the sampled data to construct sub-models. Second, the weights of each sub-model are obtained by calculating the matching degree of each new data sample with these sub-models. Finally, the final multi-model fusion result is obtained by fusing the sub-outputs of all the sub-models based on their respective weights. The proposed safety assessment method is applied to the mining dataset of the Xiaoyun Coal Mine in Jining City, Shandong Province. The experimental results show that when the sub-model is constructed at a ratio of 330/70, the Root Mean Square Error (RMSE) of this proposed method is reduced by 15.13%, 51.67%, and 12.46% compared with the diversified single model, small-sample single model, and traditional multi-model, respectively. The proposed method fully integrates the effective information provided by each sub-model, reduces and disperses the prediction error of a single model, and improves the prediction accuracy and generalization ability of the model.

Key words: degree of feature matching, heterogeneous sub-models, single model, multi-model fusion, safety assessment

摘要：

机器学习模型的好坏影响预测精度、输入与输出结果的拟合情况。在复杂系统中, 使用单一模型评估系统安全性问题时容易受数据量、数据格式、模型结构以及环境干扰等因素影响, 使得这个模型在解决某个问题的能力上比较出色, 而在解决其他问题时, 结果却不尽如人意。针对上述问题, 提出一种基于特征匹配度和异类子模型融合的安全性评估方法。首先, 按照采样数据的输出值划分不同规模的数据集并构建子模型; 其次, 通过计算每个新数据对于这些子模型的匹配度, 进而得到每个子模型的权重; 最后, 根据权重大小融合所有子模型的子输出得到最终的多模型融合结果。所提方法对山东省济宁市霄云煤矿采掘数据集进行研究, 实验结果表明, 该方法与多样本单模型、少样本单模型和传统多模型方法相比, 在以330/70的比例来构建子模型的情况下均方根误差(RMSE)分别降低了15.13%、51.67%和12.46%, 该方法充分集成各子模型所能提供的有效信息, 减少和分散单一模型的预测误差, 以提高模型的预测精度和泛化能力。

关键词: 特征匹配度, 异类子模型, 单模型, 多模型融合, 安全性评估

Xiaobin XU, Yunshuo ZHANG, Fan SHI, Leilei CHANG, Zhigang TAO. Safety Assessment Method Based on Degree of Feature Matching and Fusion of Heterogeneous Sub-Models[J]. Computer Engineering, 2024, 50(8): 113-122.

徐晓滨, 张云硕, 施凡, 常雷雷, 陶志刚. 基于特征匹配度与异类子模型融合的安全性评估方法[J]. 计算机工程, 2024, 50(8): 113-122.

/ Recommend / Download Citations

URL: https://www.ecice06.com/EN/10.19678/j.issn.1000-3428.0068028

https://www.ecice06.com/EN/Y2024/V50/I8/113

Figures/Tables 11

Fig.1 Overall framework of the multi-model fusion method

Fig.2 Xiaoyun coal mine 2301 track tunnel roadway

Fig.3 Output values and errors of Model_all on the test datasets

Fig.4 Output values of the Model_more and Model_less models on the training datasets

Fig.5 Weighting of the test datasets for both models

Fig.6 Output values and errors of multi-model fusion on test datasets

References 29

1	陈良臣, 傅德印. 面向小样本数据的机器学习方法研究综述. 计算机工程, 2022, 48(11): 1- 13. URL
	CHEN L C, FU D Y. Survey on machine learning methods for small sample data. Computer Engineering, 2022, 48(11): 1- 13. URL
2	姚乐. 面向大规模数据的工业过程分布式并行建模及应用[D]. 杭州: 浙江大学, 2019.
	YAO L. Distributed parallel modeling and application for industrial processes with large-scale data[D]. Hangzhou: Zhejiang University, 2019. (in Chinese)
3	张景莲, 彭艳兵. 基于特征融合的恶意代码分类研究. 计算机工程, 2019, 45(8): 281-286, 295. URL
	ZHANG J L, PENG Y B. Research on malware code classification based on features fusion. Computer Engineering, 2019, 45(8): 281-286, 295. URL
4	邹冲, 蔡敦波, 赵娜, 等. 基于SVM-LeNet模型融合的行人检测算法. 计算机工程, 2017, 43(5): 169- 173. doi: 10.3969/j.issn.1000-3428.2017.05.027
	ZOU C, CAI D B, ZHAO N, et al. Pedestrian detection algorithm based on SVM-LeNet model fusion. Computer Engineering, 2017, 43(5): 169- 173. doi: 10.3969/j.issn.1000-3428.2017.05.027
5	王扬, 黄亚楼, 卢敏, 等. 直接优化性能指标的多排序模型融合方法. 计算机学报, 2014, 37(8): 1658- 1668. URL
	WANG Y, HUANG Y L, LU M, et al. Multiple rank aggregation based on directly optimizing performance measure. Chinese Journal of Computers, 2014, 37(8): 1658- 1668. URL
6	陈强, 黄丹丹, 李彬, 等. 一种高度冲突证据混合分步合成算法. 计算机工程, 2015, 41(10): 302- 308. doi: 10.3969/j.issn.1000-3428.2015.10.057
	CHEN Q, HUANG D D, LI B, et al. A mixed step by step combination algorithm for highly conflict evidence. Computer Engineering, 2015, 41(10): 302- 308. doi: 10.3969/j.issn.1000-3428.2015.10.057
7	周倩. 求解大规模机器学习问题的优化方法[D]. 贵阳: 贵州大学, 2018.
	ZHOU Q. Optimization method for solving large-scale machine learning problems[D]. Guiyang: Guizhou University, 2018. (in Chinese)
8	王华军. 0/1损失支持向量机优化模型与算法[D]. 北京: 北京交通大学, 2022.
	WANG H J. Optimization models and algorithms for 0/1 loss support vector machine[D]. Beijing: Beijing Jiaotong University, 2022. (in Chinese)
9	李莉, 任振康, 石可欣. 代价敏感的Boosting软件缺陷预测方法. 计算机工程, 2022, 48(3): 175- 180. URL
	LI L, REN Z K, SHI K X. Cost sensitive Boosting software defect prediction method. Computer Engineering, 2022, 48(3): 175- 180. URL
10	李磊磊. 基于随机森林算法的IVF-ET治疗结局预测模型建立及预测特征分析研究[D]. 长春: 吉林大学, 2022.
	LI L L. Research on the establishment of prediction models of IVF-ET treatment outcomes and analysis of prediction characteristics based on random forest algorithm[D]. Changchun: Jilin University, 2022. (in Chinese)
11	犹梦洁. 基于文本挖掘的煤矿安全风险识别与评价研究[D]. 北京: 中国矿业大学(北京), 2023.
	YOU M J. Research on safety risk identification and assessment of coal mine based on text mining[D]. Beijing: China Universityof Mining & Technology(Beijing), 2023. (in Chinese)
12	王兆刚. 基于时序数据挖掘的煤矿安全趋势性知识发现研究[D]. 北京: 中国矿业大学(北京), 2020.
	WANG Z G. Research on trend knowledge discovery of coal mine safety based on times series data mining[D]. Beijing: China University of Mining & Technology(Beijing), 2020. (in Chinese)
13	SIDORENKO A A, IVANOV V V, SIDORENKO S A. Numerical simulation of rock massif stress state at normal fault at underground longwall coal mining. International Journal of Civil Engineering and Technology, 2019, 10(1): 844- 851.
14	MAŁKOWSKI P, NIEDBALSKI Z, SOJKA W. The assessment of the optimal time window for prediction of seismic hazard for longwall coal mining: the case study. Acta Geophysica, 2021, 69(2): 691- 699. doi: 10.1007/s11600-021-00541-5
15	ABIODUN O I, JANTAN A, OMOLARA A E, et al. State-of-the-art in artificial neural network applications: a survey. Heliyon, 2018, 4(11): e00938. doi: 10.1016/j.heliyon.2018.e00938
16	周末, 金敏. 多算法多模型与在线第二次学习结合的短期电力负荷预测方法. 计算机应用, 2017, 37(11): 3317- 3322. URL
	ZHOU M, JIN M. Short-term power load forecasting method combining with multi-algorithm & multi-model and online second learning. Journal of Computer Applications, 2017, 37(11): 3317- 3322. URL
17	徐晓健, 严新平, 盛晨兴, 等. 基于证据推理规则的船舶柴油机磨损类型辨识研究. 摩擦学学报, 2017, 37(6): 814- 822. URL
	XU X J, YAN X P, SHENG C X, et al. Identification on wear mode for marine diesel engine based on evidential reasoning rule. Tribology, 2017, 37(6): 814- 822. URL
18	徐雪松, 舒俭. 多模型数据集的免疫鲁棒回归分析. 计算机应用, 2014, 34(8): 2285- 2290. URL
	XU X S, SHU J. Immune robust regression analysis for data set of multiple models. Journal of Computer Applications, 2014, 34(8): 2285- 2290. URL
19	HASTIE T, ROSSET S, ZHU J, et al. Multi-class AdaBoost. Statistics and Its Interface, 2009, 2(3): 349- 360.
20	WANG Q, HE M C, YANG J, et al. Study of a no-pillar mining technique with automatically formed gob-side entry retaining for longwall mining in coal mines. International Journal of Rock Mechanics and Mining Sciences, 2018, 110, 1- 8.
21	CHANG L L, ZHANG L M, FU C, et al. Transparent digital twin for output control using belief rule base. IEEE Transactions on Cybernetics, 2022, 52(10): 10364- 10378.
22	WANG Y, CONG J L, WANG P, et al. A data-fusion approach for speed estimation and location calibration of a metro train based on low-cost sensors in smartphones. IEEE Sensors Journal, 2019, 19(22): 10744- 10752.
23	王丽红. 煤矿开采技术与掘进支护技术探讨. 能源与节能, 2023,(8): 136- 138. URL
	WANG L H. Coal mining technology and excavation support technology. Energy and Energy Conservation, 2023,(8): 136- 138. URL
24	OUCHEN S, STEINHART H, BENBOUZID M, et al. Robust DPC-SVM control strategy for shunt active power filter based on H∞ regulators. International Journal of Electrical Power & Energy Systems, 2020, 117, 105699.
25	LI D W, XIE Q, GONG X X, et al. Automatic defect detection of metro tunnel surfaces using a vision-based inspection system. Advanced Engineering Informatics, 2021, 47, 101206.
26	ONYEKWENA C C, XUE Q, LI Q, et al. Support vector machine regression to predict gas diffusion coefficient of biochar-amended soil. Applied Soft Computing, 2022, 127, 109345.
27	朱品光. 基于随机森林回归算法的堆石坝爆破块度预测研究[D]. 天津: 天津大学, 2019.
	ZHU P G. Study on blasting fragmentation prediction of rock-fill dam based on random forest regression algorithm[D]. Tianjin: Tianjin University, 2019. (in Chinese)
28	雷维嘉, 刘玥岑. LDS多址接入系统中一种低复杂度的分组多用户检测算法. 中国科学: 信息科学, 2019, 49(6): 783- 798. URL
	LEI W J, LIU Y C. A low complexity grouping multi-user detection algorithm for the LDS multiple access system. Scientia Sinica (Informationis), 2019, 49(6): 783- 798. URL
29	孟嘉慧. 多元LDPC码编译码算法复杂度优化研究[D]. 哈尔滨: 哈尔滨工程大学, 2019.
	MENG J H. Research on complexity optimization of non-binary LDPC encoding and decoding algorithm[D]. Harbin: Harbin Engineering University, 2019. (in Chinese)

[1]	Qiuchen YU, Ruohua ZHOU, Qingsheng YUAN. Multi-model Fusion Speech Wake-up Word Detection Method Based on Ghost-SE-Res2Net [J]. Computer Engineering, 2024, 50(3): 52-59.
[2]	Liangshan SHAO, Songze ZHAO. Fractional Imputation Algorithm for Incomplete Data Based on Multi-Model Fusion [J]. Computer Engineering, 2023, 49(9): 79-88, 98.
[3]	QUE Hua-kun, YANG Jing-feng, XIAO Yong, DANG San-lei. Information Security Level Assessment Based on Integrated Platform [J]. Computer Engineering, 2013, 39(10): 133-137,142.

Please choose a citation manager

Content to export