聚丙烯复合材料老化数据集成学习

doi:10.12066/j.issn.1007-2861.2382

上海大学学报(自然科学版) ›› 2022, Vol. 28 ›› Issue (3): 440-450.doi: 10.12066/j.issn.1007-2861.2382

• 数据采集、数据库和数据处理 • 上一篇下一篇

聚丙烯复合材料老化数据集成学习

武星¹^,²^,⁴(), 高进¹, 丁鹏³^,⁴

1.上海大学计算机工程与科学学院, 上海 200444
2.之江实验室, 浙江杭州 311100
3.上海大学理学院, 上海 200444
4.上海大学材料基因组工程研究院材料信息与数据科学中心, 上海 200444

收稿日期:2022-03-26 出版日期:2022-06-30 发布日期:2022-05-27
通讯作者: 武星 E-mail:xingwu@shu.edu.cn
作者简介:武星 (1980—), 男, 教授, 博士生导师, 博士,研究方向为多模态数据挖掘、机器学习. E-mail: xingwu@shu.edu.cn
基金资助:
国家重点研发计划资助项目(2018YFB0704400);国家自然基金资助项目(52073168);云南省重大科技专项资助项目(202102AB080019-3);云南省重大科技专项资助项目(202002AB080001-2);之江实验室科研攻关资助项目(2021PE0AC02);上海张江国家自主创新示范区专项发展资金重大资助项目(ZJ2021-ZD-006)

Ensemble learning of polypropylene-composite aging data

WU Xing¹^,²^,⁴(), GAO Jin¹, DING Peng³^,⁴

1. School of Computer Science and Engineering, Shanghai University, Shanghai 200444, China
2. Zhejiang Laboratory, Hangzhou 311100, Zhejiang, China
3. Research Center of Nanoscience and Nanotechnology, College of Sciences, Shanghai University, Shanghai 200444, China
4. Center of Materials Informatics and Data Science, Materials Genome Institute, Shanghai University, Shanghai 200444, China

Received:2022-03-26 Online:2022-06-30 Published:2022-05-27
Contact: WU Xing E-mail:xingwu@shu.edu.cn

摘要/Abstract

摘要：

聚丙烯复合材料老化实验周期长, 且单次实验采集的数据样本少, 使用传统机器学习方法进行预测的准确度较低. 为了解决聚丙烯复合材料老化数据样本少与预测准确性低的问题, 提出了一种虚拟样本生成(virtual sample generation, VSG)的集成学习预测方法. 首先, 对聚丙烯复合材料老化数据使用高斯混合模型(Gaussian mixed model, GMM)虚拟样本生成方法平滑生成验证有效的虚拟样本; 然后, 使用生成后的数据集建立集成学习预测模型, 该模型包含随机森林(random forest, RF)、极端梯度提升(extreme gradient boosting, XGBoost)算法、轻量级梯度提升机(light gradient boosting machine, LightGBM)算法以及分类梯度提升(categorical boosting, CatBoost)算法. 实验表明: 集成学习模型的 LightGBM 算法与 CatBoost 算法性能最优, 在测试数据上均方误差为 0.001 3 与 0.000 1, 比 RF 算法与 XGBoost 算法分别高出 0.4 与 0.2. 聚丙烯复合材料老化虚拟样本生成与集成学习方法可以有效解决实验周期长、单次实验采集的数据样本少的问题, 并可取得比单一机器学习算法更优的性能.

关键词: 聚丙烯复合材料, 材料老化, 集成学习, 高斯混合模型

Abstract:

Aging experiments conducted on polypropylene composites have long durations, and a limited number of samples can be collected in a single experiment. As a result, traditional machine-learning approaches have a low prediction accuracy. To address these issues, we present an ensemble learning prediction based on virtual sample generation (VSG). To generate valid virtual samples of aging data for polypropylene composites, we first adopted the Gaussian mixed model (GMM) method and then used the generated data set to build an ensemble-learning prediction model comprising the random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and categorical boosting (CatBoost) algorithms. The LightGBM and CatBoost algorithms in the ensemble learning model demonstrate the best performance on the test data; the mean square errors are 0.001 3 and 0.000 1, respectively, which are 0.4 and 0.2 higher than those of the RF algorithm and XGBoost algorithm, respectively. This study's aging VSG and ensemble learning approach for polypropylene composites can not only successfully overcome the long experimental times and insufficient number of data samples acquired in a single experiment but outperforms a single machine-learning algorithm.

Key words: polypropylene composites, material aging, ensemble learning, Gaussian mixture model

中图分类号:

TB 33

武星, 高进, 丁鹏. 聚丙烯复合材料老化数据集成学习[J]. 上海大学学报(自然科学版), 2022, 28(3): 440-450.

WU Xing, GAO Jin, DING Peng. Ensemble learning of polypropylene-composite aging data[J]. Journal of Shanghai University（Natural Science Edition）, 2022, 28(3): 440-450.

图/表 8

图1

图2

图3

表1

表2

表3

图4

表4

参考文献 25

[1]	Oladele I O, Oladejo M O, Adediran A A, et al. Influence of designated properties on the characteristics of dombeya buettneri fiber/graphite hybrid reinforced polypropylene composites[J]. Scientific Reports, 2020, 10(1): 1-13. doi: 10.1038/s41598-019-56847-4
[2]	李宏岩, 周琳霞. 聚丙烯/竹纤维复合材料的制备及力学和抗老化性能研究[J]. 塑料科技, 2021, 49(7): 43-46.
[3]	Nishiwaki M, Fujiwara H. Highly accurate prediction of material optical properties based on density functional theory[J]. Computational Materials Science, 2020, 172: 109315. doi: 10.1016/j.commatsci.2019.109315
[4]	He X L, Lei Z, Han Z, et al. Virtual sample generation method and its application in reforming data modeling[J]. Petroleum Processing and Petrochemicals, 2021, 52(6): 92-95.
[5]	Han M, Wang Z, Zhang X. An approach to data acquisition for urban building energy modeling using a gaussian mixture model and expectation-maximization algorithm[J]. Buildings, 2021, 11(1): 30-48. doi: 10.3390/buildings11010030
[6]	Delon J, Desolneux A. A Wasserstein-type distance in the space of Gaussian mixture models[J]. SIAM Journal on Imaging Sciences, 2020, 13(2): 936-970. doi: 10.1137/19M1301047
[7]	Kopp M, Pevný T, Holeñ M. Anomaly explanation with random forests[J]. Expert Systems with Applications, 2020, 149: 113187-113202. doi: 10.1016/j.eswa.2020.113187
[8]	Aldrich C. Process variable importance analysis by use of random forests in a shapley regression framework[J]. Minerals, 2020, 10(5): 420-436. doi: 10.3390/min10050420
[9]	Daneshvar D, Behnood A. Estimation of the dynamic modulus of asphalt concretes using random forests algorithm[J]. International Journal of Pavement Engineering, 2022, 23(2): 250-260. doi: 10.1080/10298436.2020.1741587
[10]	Sagi O, Rokach L. Approximating XGBoost with an interpretable decision tree[J]. Information Sciences, 2021, 572: 522-542. doi: 10.1016/j.ins.2021.05.055
[11]	Yan J, Xu Y, Cheng Q, et al. LightGBM: Accelerated genomically designed crop breeding through ensemble learning[J]. Genome Biology, 2021, 22(1): 1-24. doi: 10.1186/s13059-020-02207-9
[12]	Hancock J T, Khoshgoftaar T M. CatBoost for big data: an interdisciplinary review[J]. Journal of Big Data, 2020, 7(1): 1-45. doi: 10.1186/s40537-019-0278-0
[13]	de Rooij M, Weeda W. Cross-validation: A method every psychologist should know[J]. Advances in Methods and Practices in Psychological Science, 2020, 3(2): 248-263. doi: 10.1177/2515245919898466
[14]	Cai S, Zhao L, Ban Y, et al. GAN-based image-to-friction generation for tactile simulation of fabric material[J]. Computers & Graphics, 2021, 102: 460-473. doi: 10.1016/j.cag.2021.09.007
[15]	Ali M A, Guan Q, Umer R, et al. Deep learning based semantic segmentation of $\mu $CT images for creating digital material twins of fibrous reinforcements[J]. Composites Part A: Applied Science and Manufacturing, 2020, 139: 106131-106137. doi: 10.1016/j.compositesa.2020.106131
[16]	Li L, Damarla S K, Wang Y, et al. A Gaussian mixture model based virtual sample generation approach for small datasets in industrial processes[J]. Information Sciences, 2021, 581: 262-277. doi: 10.1016/j.ins.2021.09.014
[17]	Zhang Z, Mansouri Tehrani A, Oliynyk A O, et al. Finding the next superhard material through ensemble learning[J]. Advanced Materials, 2021, 33(5): 2005112-2005119. doi: 10.1002/adma.202005112
[18]	Talukdar S, Ghose B, Salam R, et al. Flood susceptibility modeling in Teesta River basin, Bangladesh using novel ensembles of bagging algorithms[J]. Stochastic Environmental Research and Risk Assessment, 2020, 34(12): 2277-2300. doi: 10.1007/s00477-020-01862-5
[19]	Liu K, Hu X, Zhou H, et al. Feature Analyses and Modeling of Lithium-Ion Battery Manufacturing Based on Random Forest Classification[J]. IEEE/ASME Transactions on Mechatronics, 2021, 26(6): 2944-2955. doi: 10.1109/TMECH.2020.3049046
[20]	Gao X, Wang L, Yao L. Porosity prediction of ceramic matrix composites based on random forest[C]// IOP Conference Series: Materials Science and Engineering. IOP Publishing: Information Technology, 2020: 052115-052121.
[21]	Khan M A, Memon S A, Farooq F, et al. Compressive strength of fly-ash-based geopolymer concrete by gene expression programming and random forest[J]. Advances in Civil Engineering, 2021, 2021: 1-17.
[22]	Ebrahimy H, Feizizadeh B, Salmani S, et al. A comparative study of land subsidence susceptibility mapping of Tasuj plane, Iran, using boosted regression tree, random forest and classification and regression tree methods[J]. Environmental Earth Sciences, 2020, 79(10): 1-12. doi: 10.1007/s12665-019-8746-6
[23]	Song K, Yan F, Ding T, et al. A steel property optimization model based on the XGBoost algorithm and improved PSO[J]. Computational Materials Science, 2020, 174: 109472-109484. doi: 10.1016/j.commatsci.2019.109472
[24]	Zhao Y, Fu C, Fu L, et al. Data-driven machine learning models for the quick and accurate prediction of Tg and Td of OLED materials[J]Materials Chemistry, 2021, 22: 1-30.
[25]	Bhamare D K, Saikia P, Rathod M K, et al. A machine learning and deep learning based approach to predict the thermal performance of phase change material integrated building envelope[J]. Building and Environment, 2021, 199: 107927-107938. doi: 10.1016/j.buildenv.2021.107927

聚丙烯复合材料老化数据集成学习

Ensemble learning of polypropylene-composite aging data

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 25

相关文章 6

编辑推荐

Metrics

本文评价

[1]	武星, 胡明涛, 丁鹏. 陶瓷涂层材料多模态数据表征学习[J]. 上海大学学报(自然科学版), 2022, 28(3): 492-503.
[2]	徐燕, 胡红青, 刘茜, 张玉凤, 丁广太, 张惠然. 面向复合材料带隙预测的两段式集成学习模型构建[J]. 上海大学学报(自然科学版), 2022, 28(3): 504-511.
[3]	侯丽敏, 张伟涛, 施丹, 刘焕成. 基于鼾声的声学特征估计睡眠呼吸暂停指数[J]. 上海大学学报(自然科学版), 2019, 25(4): 435-444.
[4]	侯丽敏缪炜. 基于韵律信息的语种辨识系统[J]. 上海大学学报(自然科学版), 2009, 15(5): 441-444.
[5]	刘悦;吴耿锋;丁智国;. 基于特征加权的神经网络集成及其应用[J]. 上海大学学报(自然科学版), 2007, 13(5): 623-628 .
[6]	李国正;李丹. 集成学习中特征选择技术[J]. 上海大学学报(自然科学版), 2007, 13(5): 598-604 .