基于标签特定特征和共同特征的标记分布学习

doi:10.6040/j.issn.1672-3961.0.2025.132

摘要/Abstract

摘要： 标记分布学习(label distribution learning, LDL)是一种新颖的机器学习范式,重点关注特定实例的标签描述度。现有的LDL模型都建立在一个简单的特征空间上,其中所有特征都由所有类标签共享。然而,这种传统的数据表示策略倾向于选择对所有标签都可区分的特征,而忽略对每个类标签相关和有区别的标签特定特征。本研究提出基于标签特定特征和共同特征的标记分布学习方法。首先,通过计算标签相关性矩阵对标签组对,针对每一组标签对,按照描述度划分所有实例,在每个类别上进行聚类,构建每个标签对的特定特征;其次,通过对所有实例聚类分析,构建标签的共同特征。最后,通过查询聚类结果进行训练和测试。在多个真实标记分布数据集上进行的综合试验验证本研究方法的优越性以及标签特定特征的有效性。

关键词: 标记分布学习, 标签特定特征, 标签共同特征, 标签相关性, 聚类

Abstract: Label distribution learning(LDL)is an innovative machine learning paradigm that focuses on the label descriptiveness of specific instances. Existing LDL models were built on a simple feature space where all features were shared by all class labels. However, this traditional data representation strategy tended to select features that were distinguishable across all labels while overlooking label-specific features that were relevant and distinctive for each class label. In this paper, we proposed a label distribution learning method based on label-specific features and common features. First, we calculated the label correlation matrix for label pairs, and for each label pair, all instances were partitioned based on descriptiveness, clustering was performed for each class, and label pair-specific features were constructed. Next, clustering analysis was performed on all instances to construct common features for the labels. Finally, training and testing were conducted by querying the clustering results. Extensive experiments on multiple real label distribution datasets validated the superiority of the method and the effectiveness of label-specific features.

Key words: label distribution learning, label-specific feature, label common features, label correlation, clustering

中图分类号:

TP181

孔舒,陈琳琳,何强,王恒友. 基于标签特定特征和共同特征的标记分布学习[J]. 山东大学学报 (工学版), 2026, 56(5): 52-61.

Kong Shu, Chen Linlin, He Qiang, Wang Hengyou. Label distribution learning based on label-specific features and common features[J]. Journal of Shandong University(Engineering Science), 2026, 56(5): 52-61.

参考文献

[1] Zhang M L, Wu L. Lift: multi-label learning with label-specific features[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(1): 107-120.
[2] 张海翔, 李培培, 胡学钢. 类不平衡的公共和标签特定特征多标签分类[J]. 计算机技术与发展, 2024, 34(2): 46-52. Zhang Haixiang, Li Peipei, Hu Xuegang. Class imbalance multi-label classification with common and label specific features[J]. Computer Technology and Development, 2024, 34(2): 46-52.
[3] Ren T T, Jia X Y, Li W W, et al. Label distribution learning with label-specific features[C] //Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence. New York, USA: International Joint Conferences on Artificial Intelligence Organization, 2019: 3318-3324.
[4] 王进, 梁晨, 孙开伟, 等. 基于标签相关性的标签特定特征多标签学习[J]. 江苏大学学报(自然科学版), 2023, 44(5): 554-563. Wang Jin, Liang Chen, Sun Kaiwei, et al. Multi-label learning with label-specific features via label correlations[J]. Journal of Jiangsu University(Natural Science Edition), 2023, 44(5): 554-563.
[5] Qian W B, Ye Q Z, Li Y H, et al. Label distribution feature selection with feature weights fusion and local label correlations[J]. Knowledge-Based Systems, 2022, 256: 109778.
[6] 石荣毅. 基于标签特定特征和偏标签的多标签学习[D]. 舟山: 浙江海洋大学, 2025: 19-26. Shi Rongyi. Multi-label learning with label-specific features and partial labels[D]. Zhoushan: Zhejiang Ocean University, 2025: 19-26.
[7] Li G L, Zhang H R, Min F, et al. Two-stage label distribution learning with label-independent prediction based on label-specific features[J]. Knowledge-Based Systems, 2023, 267: 110426.
[8] Geng X, Smith-Miles K, Zhou Z H. Facial age estimation by learning from label distributions[C] //Proceedings of the 24th AAAI Conference on Artificial Intelligence. Atlanta, USA: AAAI Press, 2010: 451-456.
[9] Geng X. Label distribution learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2016, 28(7): 1734-1748.
[10] Liu S P, Samonte M J C. The advancement in multi-label and label distribution learning[C] //2024 International Conference on Electronics and Devices, Computational Science. Marseille, France: IEEE, 2024: 747-751.
[11] Shu W H, Xia Q, Qian W B. Label distribution feature selection based on label-specific features[J]. Applied Intelligence, 2024, 54(19): 9195-9212.
[12] Zhang Z Y, Wang J, Geng X. Label distribution learning by utilizing common and label-specific feature fusion space[J]. International Journal of Machine Learning and Cybernetics, 2025, 16(3): 1545-1558.
[13] Huang J, Qin F, Zheng X, et al. Improving multi-label classification with missing labels by learning label-specific features[J]. Information Sciences, 2019, 492: 124-146.
[14] Ma J H, Chow T W S, Zhang H J. Semantic-gap-oriented feature selection and classifier construction inmultilabel learning[J]. IEEE Transactions on Cybernetics, 2022, 52(1): 101-115.
[15] Li J L, Li P P, Hu X G, et al. Learning common and label-specific features for multi-Label classification with correlation information[J]. Pattern Recognition, 2022, 121: 108259.
[16] Zhang J, Wu H R, Jiang M, et al. Group-preserving label-specific feature selection for multi-label learning[J]. Expert Systems with Applications, 2023, 213: 118861.
[17] Hang J Y, Zhang M L. Collaborative learning of label semantics and deep label-specific features for multi-labelclassification[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022, 44(12): 9860-9871.
[18] Zhang M L, Fang J P, Wang Y B.BiLabel-specific features for multi-label classification[J]. ACM Transactions on Knowledge Discovery from Data, 2021, 16(1): 1-23.
[19] 马坤, 刘筱云, 李乐平, 等. 用于意图识别的自适应多标签信息学习模型[J]. 山东大学学报(工学版), 2024, 54(1): 45-51. Ma Kun, Liu Xiaoyun, Li Leping, et al. Adaptive label information learning for intention detection[J]. Journal of Shandong University(Engineering Science), 2024, 54(1): 45-51.
[20] 唐杰烽, 张佳, 龙锦益. 基于全局冗余最小的快速多标签特征选择方法[J]. 山东大学学报(工学版), 2025, 55(6): 21-34. Tang Jiefeng, Zhang Jia, Long Jinyi. Fast multi-label feature selection method based on global redundancy minimization[J]. Journal of Shandong University(Engineering Science), 2025, 55(6): 21-34.
[21] Klein Y, Efrati E, Sharon E. Shaping of elastic sheets by prescription of non-euclidean metrics[J]. Science, 2007, 315(5815): 1116-1120.
[22] Geng X, Luo L R. Multilabel ranking with inconsistent rankers[C] //2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, USA: IEEE, 2014: 3742-3747.
[23] Boutell M R, Luo J B, Shen X P, et al. Learning multi-label scene classification[J]. Pattern Recognition, 2004, 37(9): 1757-1771.
[24] Yin L J, Wei X Z, Sun Y, et al. A 3D facial expression database for facial behavior research[C] //7th International Conference on Automatic Face and Gesture Recognition. LosAlamitos, USA: IEEE, 2006: 211-216.
[25] Wang J, Feng F,Lv J H, et al. Residual k-nearest neighbors label distribution learning[J]. Pattern Recognition, 2025, 158: 111006.
[26] Friedman M. A comparison of alternative tests of significance for the problem of m rankings[J]. The Annals of Mathematical Statistics, 1940, 11(1): 86-92.
[27] Dunn O J. Multiple comparisons among means[J]. Journal of the American Statistical Association, 1961, 56(293): 52-64.
[28] Demšar J. Statistical comparisons of classifiers over multiple data sets[J]. Journal of Machine Learning Research, 2006, 7: 1-30.

相关文章 15

[1]	代洪娜,李忠锐,冯海霞,赵欢欢,范宏贤. 基于ETC收费数据的特大城市高速公路通勤流识别方法研究[J]. 山东大学学报 (工学版), 2026, 56(5): 40-51.
[2]	李晓辉,刘小飞,孙炜桐,赵毅,董媛,靳引利. 基于车辆与无人机协同的巡检任务分配与路径规划算法[J]. 山东大学学报 (工学版), 2025, 55(5): 101-109.
[3]	陈素根,赵志忠. 融合局部截断距离及小簇合并的密度峰值聚类[J]. 山东大学学报 (工学版), 2025, 55(2): 58-70.
[4]	王梅,宋凯文,刘勇,王志宝,万达. DMKK-means——一种深度多核K-means聚类算法[J]. 山东大学学报 (工学版), 2024, 54(6): 1-7.
[5]	王丽娟,徐晓,丁世飞. 面向密度峰值聚类的高效相似度度量[J]. 山东大学学报 (工学版), 2024, 54(3): 12-21.
[6]	张鑫,费可可. 基于log鲁棒核岭回归的子空间聚类算法[J]. 山东大学学报 (工学版), 2023, 53(6): 26-34.
[7]	李兆彬,叶军,周浩岩,卢岚,谢立. 变异萤火虫优化的粗糙K-均值聚类算法[J]. 山东大学学报 (工学版), 2023, 53(4): 74-82.
[8]	侯延琛,赵金东. 任意形状聚类的SPK-means算法[J]. 山东大学学报 (工学版), 2023, 53(2): 87-92.
[9]	卢建云,张蔚,李林. 一种基于动态局部密度和聚类结构的聚类算法[J]. 山东大学学报 (工学版), 2022, 52(2): 118-127.
[10]	程业超,刘惊雷. 自适应图正则的单步子空间聚类[J]. 山东大学学报 (工学版), 2022, 52(2): 57-66.
[11]	孟银凤,杨佳宇,曹付元. 函数型数据的分裂转移式层次聚类算法[J]. 山东大学学报 (工学版), 2022, 52(1): 19-27.
[12]	朱恒东, 马盈仓, 代雪珍. 自适应半监督邻域聚类算法[J]. 山东大学学报 (工学版), 2021, 51(4): 24-34.
[13]	解子奇,王立宏,李嫚. 块对角子空间聚类中成对约束的主动式学习[J]. 山东大学学报 (工学版), 2021, 51(2): 65-73.
[14]	朱昌明,岳闻,王盼红,沈震宇,周日贵. 主动三支聚类下的全局和局部多视角多标签学习算法[J]. 山东大学学报 (工学版), 2021, 51(2): 34-46.
[15]	李蓓,赵松,谢志佳,牛萌. 电动汽车虚拟储能可用容量建模[J]. 山东大学学报 (工学版), 2020, 50(6): 101-111.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed