融合分数阶微分与PIMP-RF算法的集成学习模型预测成熟期苹果可溶性固形物含量

doi:10.3964/j.issn.1000-0593(2023)10-3059-08

摘要
参考文献
相关文章 (15)

全文: PDF (5406 KB)
输出: BibTeX | EndNote (RIS)

摘要：可溶性固形物含量（SSC）是反映苹果品质和成熟度的重要生理指标，能够用于苹果品质分析和成熟度预测。以新疆阿克苏冰糖心红富士苹果为研究对象，从果实膨大定形期至完熟期，以等间隔周期3 d采摘样本，测其380~1 100 nm的可见/近红外光谱和SSC，共552个样本。然后融合分数阶微分（FD）及置换重要性-随机森林（PIMP-RF）算法，构建成熟期苹果SSC预测的集成学习模型。结果表明，基于PLS模型优选的分数阶微分阶次为0阶、0.4阶、1.1阶和1.6阶，且通过PIMP-RF算法进行特征重要性和可解释性分析结果显示，利用可见/近红外光谱预测成熟期苹果SSC的关键波长主要为可见光波段，这为今后研发新疆冰糖心红富士苹果的快速无损检测设备提供参考；基于分数阶微分技术和PIMP-RF算法构建的成熟期苹果SSC集成学习模型具有很好的预测能力，其训练集的相关系数r等于0.989 2，平均绝对误差MAE等于0.241 2，均方根误差RMSE等于0.309 1，平均绝对百分误差等于0.018 3；测试集的相关系数r等于0.903 8，平均绝对误差MAE等于0.549 9，均方根误差RMSE等于0.740 8，平均绝对百分误差等于0.043 4，相比于FD0-PIMP-RF、FD0.4-PIMP-RF、FD1.1-PIMP-RF和FD1.6-PIMP-RF模型，集成学习模型为最优。故而，集成分数阶微分技术与PIMP-RF算法，结合可见近红外光谱技术可有效地实现成熟期苹果的可溶性固形物含量预测。

关键词：可见/近红外光谱；分数阶微分；置换重要性-随机森林；K近邻（KNN）回归；可溶性固形物含量

Abstract：Soluble solids content (SSC) is an important physiological indicator of apple quality and maturation, and can be used for predicting the quality and maturity of apples. In this paper, 552 samples were collected at equal intervals of 3 d from the fruit swelling and setting stage to the complete mature stage, and the SSC was determined by collecting visible/near-infrared spectra from 380 to 1100 nm, and fused with fractional differential (FD) technique and replacement importance-random forest (Permutation Importance-Random Forest, PIMP-RF) algorithm to construct an ensemble learning model for SSC prediction in apple during maturing period. The results showed that the fractional differential orders of the PLS model were 0, 0.4, 1.1, and 1.6, and the results of feature importance and interpretability analysis by the PIMP-RF algorithm showed that the key wavelengths for predicting the SSC of maturity apples using visible/near-infrared spectroscopy were mainly in the visible band, which provided a theoretical basis for the future development of a rapid nondestructive detection device for Xinjiang Red Fuji apples. The ensemble learning model of apple ripening SSC constructed based on fractional differential technique and PIMP-RF algorithm has good prediction ability, with the correlation coefficient r equal to 0.989 2, mean absolute error MAE equal to 0.241 2, root mean square error RMSE equal to 0.309 1 and mean absolute percentage error equal to 0.018 3 in the training set. The correlation coefficient r of the test set is equal to 0.903 8, the mean absolute error MAE is equal to 0.549 9, the root mean square error RMSE is equal to 0.740 8, and the mean absolute percentage error is equal to 0.043 4, compared to the FD0-PIMP-RF, FD0.4-PIMP-RF, FD1.1-PIMP-RF, and FD1.6-PIMP-RF models, the ensemble learning model is optimal. Therefore, the integrated fractional order differentiation technique and PIMP-RF algorithm, combined with visible/near-infrared spectroscopy, can successfully and effectively predict the soluble solids content of apples during maturing period.

Key words：Visible/near-infrared spectrum; Fractional differential; Permutation importance-random forest; K-nearest neighbors (KNN) regression; Soluble solids content

收稿日期: 2022-03-09 修订日期: 2022-11-28

中图分类号:

O657.3

基金资助: 国家自然科学基金项目（61367001），新疆维吾尔自治区教育厅面上重点项目（XJEDU2020I009），新疆维吾尔自治区科技厅面上基金项目(2019D01A52)，2022年度新疆农业大学大学生创新项目资助

通讯作者: 郭俊先 E-mail: junxianguo@163.com

作者简介: 黄华，1981年生，新疆农业大学数理学院副教授 e-mail: 18899511151@163.com

引用本文:

黄华，刘亚，库尔班古丽·都力昆，曾繁琳，玛依热·麦麦提，阿瓦古丽·麦麦提，买地努尔汗·艾则孜，郭俊先. 融合分数阶微分与PIMP-RF算法的集成学习模型预测成熟期苹果可溶性固形物含量[J]. 光谱学与光谱分析, 2023, 43(10): 3059-3066.
HUANG Hua, LIU Ya, KUERBANGULI·Dulikun, ZENG Fan-lin, MAYIRAN·Maimaiti, AWAGULI·Maimaiti, MAIDINUERHAN·Aizezi, GUO Jun-xian. Ensemble Learning Model Incorporating Fractional Differential and PIMP-RF Algorithm to Predict Soluble Solids Content of Apples During Maturing Period. SPECTROSCOPY AND SPECTRAL ANALYSIS, 2023, 43(10): 3059-3066.

链接本文:

https://www.gpxygpfx.com/CN/10.3964/j.issn.1000-0593(2023)10-3059-08 或 https://www.gpxygpfx.com/CN/Y2023/V43/I10/3059

[1] Pavlina D D, Georgios P. Scientia Horticulturae，2011, 129(4): 752.
[2] Wang H, Peng J, Xie B, et al. Sensors, 2015, 15(5): 11889.
[3] Wang L, Sun D W, Pu H B, et al. Food Science and Nutrition, 2017, 57(7): 1524.
[4] ZHAO Jie-wen, ZHANG Hai-dong, LIU Mu-hua（赵杰文, 张海东, 刘木华）. Transactions of the Chinese Society of Agricultural Engineering（农业工程学报）, 2005, 21(3): 162.
[5] YUAN Lei-ming, GAO Hai-ning, LÜ Song, et al（袁雷明, 高海宁, 吕松, 等）. Journal of Food Safety and Quality（食品安全质量检测报）, 2012, 3(5): 448.
[6] Fan G, Zha J, Du R, et al. Journal of Food Engineering, 2009, 93(4): 416.
[7] LIU Yan-de, WANG Jun-zheng, JIANG Xiao-gang, et al（刘燕德, 王军政, 姜小刚, 等）. Spectroscopy and Spectral Analysis（光谱学与光谱分析）, 2021, 41(7): 2064.
[8] Tian X, Li J, Wang Q, et al. Food Chemistry, 2018, 239(15): 1055.
[9] Fan S, Huang W, Li J, et al. Biosystem Engineering，2016, 43(45): 9.
[10] Guo Z, Huang W, Peng Y, et al. Postharvest Biology & Technology, 2016, 115: 81.
[11] MENG Qing-long, SHANG Jing, HUANG Ren-shuai, et al（孟庆龙, 尚静, 黄人帅, 等）. Storage and Process（保鲜与加工）, 2020, 20(5): 185.
[12] Bai Y, Xiong Y, Huang J, et al. Postharvest Biology and Technology, 2019, 156: 110943.
[13] ZHANG He-dong, WU Jing-zhu, HAN Ping, et al（张鹤冬, 吴静珠, 韩平, 等）. Journal of Food Safety and Quality（食品安全质量检测学报）, 2019, 10(1): 209.
[14] Zhang Dongyan, Xu Yunfei, Huang Wenqian, et al. Infrared Physics & Technology, 2019, 98: 297.
[15] RAO Li-bo, CHEN Xiao-yan, PANG Tao（饶利波, 陈晓燕, 庞涛）. Chinese Journal of Luminescence（发光学报）, 2019, 40(3): 389.
[16] Huang Y, Wang J, Li N, et al. Pattern Recognition Letters, 2021, 151: 76.
[17] Xia Y, Fan S, Li J, et al. Chemometrics and Intelligent Laboratory Systems, 2020, 201(15): 104017.
[18] Tian X, Fan S, Li J, et al. Biosystems Engineering, 2020, 197: 64.
[19] Ahyeong L, Jaeseung S, Balgeum K, et al. Journal of Food Engineering, 2022, 321: 110945.
[20] Lengauer T. Bioinformatics, 2010, 26(10): 1340.