一种基于 SC-StackingPSO-MI 的油菜籽粒表观特征筛选方法

A method of screening phenotypic features of rape seeds based on SC-StackingPSO-MI

摘要

[目的]为实现油菜籽粒关键表观特征的有效筛选，服务于油菜优良种质资源快速筛选，提出了一种基于谱聚类(spectral clustering,SC)-堆叠集成学习粒子群优化(stacking particle swarm optimization,StackingPSO)-互信息(mutual information,MI)的油菜籽粒关键表观特征筛选方法 SC-StackingPSO-MI。[方法]在确定油菜籽粒最优谱聚类的基础上，结合特征与其聚类结果的 MI 值，基于逻辑回归(logistic regression,LR)算法联合 k 最近邻(k-nearest neighbor,KNN)、随机森林(random forest,RF)和支持向量机(support vector machine,SVM)3 种基础学习器设计 Stacking 集成学习模型，用于优化 PSO 中适应度函数进行油菜籽粒关键表观特征的筛选，并基于优选特征开展油菜籽粒聚类分析、不同分类器以及不同特征选择算法下的油菜籽粒分类精度比对。[结果]二分聚类条件下，SC-StackingPSO-MI 特征筛选算法能够很好地提取出油菜籽粒表观特征中统计方差较大的特征指标，优选特征谱聚类下的 Calinski-Harabasz(CH)聚类指数从 42.89 提高到 83.90，特征聚类性能得到显著改善。基于二分聚类标签对比不同分类器下 SC-StackingPSO-MI 油菜籽粒筛选特征的整体分类测试效果，除 RF 模型外的 KNN、SVM 和 LR 等分类模型的测试精度相较于全特征组合均有不同程度提高，特征筛选结果具有较好的稳健性。与 LASSO、SVM-RFE 和 XGBoost 3 种特征选择方法对比，SC-StackingPSO-MI 优选特征组合下，基于二分聚类标签的 SVM 模型测试集精度准确率(ACC)和 Kappa 值分别为 97.22%和 93.88%，均高于 LASSO、SVM-RFE 和 XGBoost 优选特征组合下的测试精度，特征筛选方法具有较高的可靠性。[结论]本文提出的特征筛选方法在油菜籽粒关键表观特征选择中展现出较好的优越性，可为油菜籽粒以及其他作物优良种质资源筛选中的关键表观特征选取提供相关技术参考。

Abstract

[Objectives] To extract effectively the key phenologic traits of rape seeds and serve the rapid screening of excellent rape seed germplasm resources, based on spectral clustering (SC), stacking particle swarm optimization (StackingPSO) and mutual information(MI), a method of rape seeds key phenotypic trait characters screening SC-StackingPSO-MI was proposed. [Methods] After the determination of the optimal spectral clustering of rape seed grains, this paper designed the Stacking ensemble learning model based on the logistic regression(LR) algorithm combining with k-nearest neighbor(KNN), random forest(RF), and support vector machine(SVM) to evaluate the fitness function in the particle swarm optimization (PSO) for the screening of key apparent trait features of rape seed grains by combining the MI between the features and their clustering results. Furthermore, clustering analysis of rape seed grains, the comparison of the classification accuracies of rape seed grains under different classifiers and different feature selection algorithms based on the selected key features were carried out. [Results] Under the binary clustering condition, the SC-StackingPSO-MI feature selection algorithm could effectively extract the feature indicators with high statistical variance from the apparent features of rape seeds. The Calinski-Harabasz(CH) indices with the optimized features increased from 42.89 to 83.90, which resulted in a significant improvement in the clustering performance. The overall classification test performance of rape seeds screening features from SC-StackingPSO-MI under different classifiers was compared based on binary clustering labels. Compared with the full feature combination, three classification models such as KNN, SVM and LR all showed varying degrees of improvement in testing accuracy apart from RF model. That proved the feature selection results had good robustness. Compared to the feature selection methods of LASSO, SVM-RFE and XGBoost under binary clustering labels, the SVM model's test accuracy(ACC) and Kappa values were 97.22% and 93.88%, respectively, indicating the higher reliability. [Conclusions] The feature screening method proposed in this paper had demonstrated good superiority in the selection of key apparent trait features of rape seed grains, and could provide relevant technical references for the selection of key apparent trait characteristics in the screening of excellent germplasm resources of rape seed grains and other crops.

关键词

关键表观特征选择 / 油菜籽粒 / 粒子群优化 / 互信息 / 集成学习

Key words

selection of key epigenetic trait characteristics / rape seeds / particle swarm optimization / mutual information / ensemble learning

参考文献

[1] 张智猛， 慈敦伟， 丁红， 等. 花生种子大小和形状对出苗和幼苗建成的影响[J]. 花生学报， 2014， 43(1): 16-23.

[2] Zhang ZM, Ci DW, Ding H, et al. Effects of peanut seed size and shape on seedling emergence and establishment[J]. Journal of Peanut Science, 2014, 43(1): 16-23(in Chinese with English abstract).

[3] 杨蜀秦. 农作物籽粒的图像处理和识别方法研究[D]. 杨凌:西北农林科技大学， 2012.

[4] Yang SQ. The research of image processing and identification method for crop's kernels [D]. Yangling: Northwest A&F University, 2012 (in Chinese with English abstract).

[5] 刘伟， 刘长虹， 郑磊. 基于支持向量机的多光谱成像稻谷品种鉴别[J]. 农业工程学报， 2014， 30(10): 145-151.

[6] Liu W, Liu CH, Zheng L. Discrimination in varieties of rice seeds with multispectral imaging using support vector machine[J]. Transactions of the Chinese Society of Agricultural Engineering, 2014, 30(10): 145-151(in Chinese with English abstract).

[7] 倪浩然， 王润， 胡昊， 等. 基于多光谱成像技术鉴别不同紫花苜蓿品种的种子[J]. 草地学报， 2023， 31(10): 2985-2991.

[8] Ni HR, Wang R, Hu H, et al. Seed identification of alfalfa varieties based on multispectral imaging[J]. Acta Agrestia Sinica, 2023, 31(10): 2985-2991(in Chinese with English abstract).

[9] Fu XZ, Bai MJ, Xu YW, et al. Cultivar identification of oat(Avena sativa L.)seed via multispectral imaging analysis[J]. Frontiers in Plant Science, 2023, 14: 1113535.

[10] Zhang SH, Zeng HG, Ji W, et al. Non-destructive testing of alfalfa seed vigor based on multispectral imaging technology[J]. Sensors, 2022, 22 (7): 2760.

[11] 冯继克， 郑颖， 李平， 等. 基于特征选择的小麦籽粒品种识别研究[J]. 中国农机化学报， 2022， 43(7): 116-123.

[12] Feng JK, Zheng Y, Li P, et al. Research on the identification of wheat grain varieties based on feature selection[J]. Journal of Chinese Agricultural Mechanization, 2022, 43(7): 116-123(in Chinese with English abstract).

[13] 黄琼， 杨红云， 万颖. 基于特征数据的水稻种子分类识别方法[J]. 江苏农业学报， 2021， 37(1): 8-15.

[14] Huang Q, Yang HY, Wan Y. Classification and recognition method of rice seeds based on feature data[J]. Jiangsu Journal of Agricultural Sciences, 2021, 37(1): 8-15(in Chinese with English abstract).

[15] 孟惜， 王克俭， 韩宪忠. 基于改进 BP 网络的小麦品种识别[J]. 贵州农业科学， 2017， 45(10): 156-160.

[16] Meng X, Wang KJ, Han XZ. Classification of wheat varieties by improved BP neural network[J]. Guizhou Agricultural Sciences, 2017, 45(10): 156-160(in Chinese with English abstract).

[17] von Luxburg U. A tutorial on spectral clustering[J]. Statistics and Computing, 2007, 17(4): 395-416.

[18] 李帅， 常锦才， 李吕牧之， 等. 基于差分隐私保护的 Stacking 集成聚类算法研究[J]. 计算机工程与科学， 2022， 44(8): 1402-1408.

[19] Li S, Chang JC, Lilü MZ, et al. A stacking ensemble clustering algorithm based on differential privacy protection[J]. Computer Engineering and Science, 2022, 44(8): 1402-1408(in Chinese with English abstract).

[20] Kennedy J, Eberhart R. Particle swarm optimization[C] // Proceedings of ICNN'95 - International Conference on Neural Networks. November 27 - December 1, 1995, Perth, WA, Australia. IEEE, 1995: 1942-1948.

[21] Yao H, Fu BL, Zhang Y, et al. Combination of hyperspectral and quad-polarization SAR images to classify marsh vegetation using stacking ensemble learning algorithm[J]. Remote Sensing, 2022, 14(21): 5478.

[22] 张婧， 曹峰， 董毓莹， 等. 基于互信息和遗传算法的特征选择算法[J]. 山西大学学报(自然科学版)， 2024， 47(1): 1-8.

[23] Zhang J, Cao F, Dong YY, et al. Feature selection algorithm based on mutual information and genetic algorithm[J]. Journal of Shanxi University (Natural Science Edition), 2024, 47(1): 1-8(in Chinese with English abstract).

[24] 王天一. 基于互信息和多目标粒子群优化的特征选择方法研究[D]. 镇江:江苏大学， 2022.

[25] Wang TY. A study of feature selection method based on mutual information and multi-objective particle swarm optimization [ D]. Zhenjiang: Jiangsu University, 2022(in Chinese with English abstract).

[26] 王学武， 魏建斌， 周昕， 等. 一种基于超体积指标的多目标进化算法[J]. 华东理工大学学报(自然科学版)， 2020， 46(6): 780-791.

[27] Wang XW, Wei JB, Zhou X, et al. Hypervolume-based multi-objective evolutionary algorithm[J]. Journal of East China University of Science and Technology, 2020, 46(6): 780-791(in Chinese with English abstract).

[28] 康岩松，臧顺来. 基于多种策略的改进粒子群优化算法[J]. 东北大学学报(自然科学版)，2023，44(8):1089-1097.

[29] Kang YS, Zang SL. Improved particle swarm optimization algorithm based on multiple strategies[J]. Journal of Northeastern University(Natural Science), 2023, 44(8):1089-1097(in Chinese with English abstract).

[30] 薛汉军， 孙利军， 梁升平， 等. 甘蓝型油菜粒色与含油量的研究[J]. 西北农业学报， 2010， 19(1): 81-85.

[31] Xue HJ, Sun LJ, Liang SP, et al. Research of the color and oil content of the seed in Brassica napus L.[J]. Acta Agriculturae Boreali-Occidentalis Sinica, 2010, 19(1): 81-85(in Chinese with English abstract).

[32] 陈浩. 不同地域气候条件对油菜籽粒颜色及萌发活力的影响研究[D]. 武汉:华中农业大学， 2022.

[33] Chen H. The influence of different regional climatic conditions on grain color and germination vigor of rape seed [ D]. Wuhan: Huazhong Agricultural University, 2022(in Chinese with English abstract).

[34] 马元花， 汪乐印， 张祯鑫， 等. 融合光谱和纹理特征的玉米产量预测研究[J]. 西北农业学报， 2024， 33(10): 1827-1838.

[35] Ma YH, Wang LY, Zhang ZX, et al. Remote sensing estimation for maize yield integrating spectral and texture features[J]. Acta Agriculturae Boreali-Occidentalis Sinica, 2024, 33(10): 1827-1838(in Chinese with English abstract).

基金资助

重庆市市级财政科研项目(cqass2023sjczhx003)

重庆市自然科学基金项目(CSTB2024NSCQ-MSX0804)

农业农村部农业监测预警技术重点实验室开放课题(KLAMEWT202402)