采用动态相关度权重的特征选择算法

doi:10.3778/j.issn.1002-8331.2306-0230

摘要/Abstract

摘要： 基于互信息的特征选择算法在考虑候选特征提供的新分类信息时，通常忽略了候选特征的加入会使得已选特征和类标签的相关性发生变化而带来额外的新增信息量，以及在计算冗余信息时采用累加求和的形式可能导致低估候选特征的冗余程度。针对以上问题，提出动态相关度权重的定义，以更全面地考虑候选特征带来的新信息量成分；提出改进冗余项的定义，采用取最大值和归一化策略，以解决传统算法存在的低估冗余问题；在此基础上提出一种采用动态相关度权重的特征选择算法（feature selection using dynamic relevance weight，FSDRW）。选取五种当前主流的基于互信息的过滤式特征选择算法进行对比实验，在来自加州大学尔湾分校UCI和亚利桑那州立大学ASU的机器学习测试数据集上的实验表明，所提出的算法在分类准确率及综合性能方面具有较好的表现。最后将所提出算法应用于广西某水库工程的微震、爆破信号识别中，算法选取出的特征用于微震信号识别可达到98.86%的分类准确率，验证了算法在实际应用中的有效性。

关键词: 特征选择, 互信息, 信息熵, 动态相关度权重

Abstract: When considering the new classification information provided by the candidate features, the features selection algorithm based on mutual information usually ignores that the addition of the candidate features will cause the change of the correlation between the selected features and class labels, which will bring additional new information; in addition, when calculating redundant information, using the form of cumulative summation may lead to the underestimation of the redundancy degree of candidate features. In view of the above problems, the definition of dynamic relevance weight is proposed to more comprehensively consider the new information components brought by candidate features. The improved definition of redundant items is proposed, and the maximum and normalization strategy are adopted to solve the problem of underestimating redundancy. On this basis, the feature selection using dynamic relevance weight (FSDRW) is proposed. Five current mainstream filter-based feature selection algorithms based on mutual information are selected for comparative experiments. Experiments on machine learning test datasets from UCI (University of California Irvine) and ASU (Arizona State University) show that the proposed algorithm works well in classification accuracy and comprehensive performance. Finally, the proposed algorithm is applied to the recognition of microseismic and blasting signals in a reservoir project in Guangxi. The selected features of the algorithm are used for microseismic signal recognition, achieving a classification accuracy of 98.86%, verifying the effectiveness of the algorithm in practical applications.

Key words: feature selection, mutual information, information entropy, dynamic relevance weight

许华杰, 刘冠霆, 张品, 秦远卓. 采用动态相关度权重的特征选择算法[J]. 计算机工程与应用, 2024, 60(4): 89-98.

XU Huajie, LIU Guanting, ZHANG Pin, QIN Yuanzhuo. Feature Selection Algorithm Using Dynamic Relevance Weight[J]. Computer Engineering and Applications, 2024, 60(4): 89-98.

参考文献

[1] 梁云辉, 甘舰文, 陈艳, 等. 基于对偶流形重排序的无监督特征选择算法[J]. 计算机科学, 2023, 50(11): 1-10.
LIANG Y H, GAN J W, CHEN Y, et al. Unsupervised feature selection algorithm based on dual manifold re-ranking[J]. Computer Science, 2023, 50(11): 1-10.
[2] 张俐, 陈小波. 动态加权条件互信息的特征选择算法[J]. 电子与信息学报, 2021, 43(10): 3028-3034.
ZHANG L, CHEN X B. Feature selection algorithm for dynamically weighted conditional mutual information[J]. Journal of Electronics & Information Technology, 2021, 43(10): 3028-3034.
[3] BENNASAR M, HICKS Y, SETCHI R. Feature selection using joint mutual information maximisation[J]. Expert Systems with Applications, 2015, 42(22): 8520-8532.
[4] YIN K X, XIE A F, ZHAI J R, et al. Dynamic interaction-based feature selection algorithm for maximal relevance minimal redundancy[J]. Applied Intelligence, 2022, 53(4): 1-17.
[5] 姜文煊, 段友祥, 孙歧峰. 基于交互信息的混合特征选择算法[J]. 应用科学学报, 2021, 39(4): 545-558.
JIANG W X, DUAN Y X, SUN Q F. Hybrid feature selection algorithm based on mutual information[J]. Journal of Applied Sciences, 2021, 39(4): 545-558.
[6] GAO W F, HU L, ZHANG P. Feature redundancy term variation for mutual information-based feature selection[J]. Applied Intelligence, 2020, 50(4): 1272-1288.
[7] ZHAO S J, WANG M C, MA S L, et al. A feature selection method via relevant-redundant weight[J]. Expert Systems with Applications, 2022, 207: 117923.
[8] MA X A, XU H, JU C. Class-specific feature selection via maximal dynamic correlation change and minimal redundancy[J]. Expert Systems with Applications, 2023, 229: 120455.
[9] YIN K, ZHAI J, XIE A, et al. Feature selection using max dynamic relevancy and min redundancy[J]. Pattern Analysis and Applications, 2023, 26(2): 631-643.
[10] ZHANG P, GAO W F, HU J C, et al. A conditional-weight joint relevance metric for feature relevancy term[J]. Engineering Applications of Artificial Intelligence, 2021, 106(4): 104481.
[11] ZHANG L. A feature selection method using conditional correlation dispersion and redundancy analysis[J]. Neural Processing Letters, 2023, 55(6): 7175-7209.
[12] ZHANG P, GAO W F. Feature selection considering uncertainty change ratio of the class label[J]. Applied Soft Computing, 2020, 95(4): 106537.
[13] 陈永波, 李巧勤, 刘勇国. 基于动态相关性的特征选择算法[J]. 计算机应用, 2022, 42(1): 109-114.
CHEN Y B, LI Q Q, LIU Y G. Dynamic relevance based feature selection algorithm[J]. Journal of Computer Applications, 2022, 42(1): 109-114.
[14] LI J D, CHENG K W, WANG S H, et al. Feature selection: a data perspective[J]. ACM Computing Surveys, 2017, 50(6): 1-45.
[15] 陈泽, 丁琳琳, 罗浩, 等. 基于改进小波分解和ELM的矿山微震事件识别方法[J]. 煤炭学报, 2020, 45(S2): 637-648.
CHEN Z, DING L L, LUO H, et al. Mine microseismic events classification based on improved wavelet decomposition and ELM[J]. Journal of China Coal Society, 2020, 45(S2): 637-648.
[16] 樊鑫, 程建远, 王云宏, 等. 基于小波散射分解变换的煤矿微震信号智能识别[J]. 煤炭学报, 2022, 47(7): 2722-2731.
FAN X, CHENG J Y, WANG Y H, et al. Intelligent recognition of coal mine microseismic signal based on wavelet scattering decomposition transform[J]. Journal of China Coal Society, 2022, 47(7): 2722-2731.
[17] 吴义文, 程铁栋, 易其文, 等. 基于经验小波变换的矿山微震信号识别研究[J]. 矿业安全与环保, 2020, 47(4): 39-44.
WU Y W, CHENG T D, YI Q W, et al. Research on mine microseismic signal recognition based on empirical wavelet transform[J]. Mining Safety & Environmental Protection, 2020, 47(4): 39-44.
[18] 程铁栋, 易其文, 吴义文, 等. 改进EWT_MPE模型在矿山微震信号特征提取中的应用[J]. 振动与冲击, 2021, 40(9): 92-101.
CHENG T D, YI Q W, WU Y W, et al. Application of improved EWT_MPE model in feature extraction of mine micro-seismic signals[J]. Journal of Vibration and Shock, 2021, 40(9): 92-101.
[19] 孟娟, 张家声, 李亚南. 基于改进EWT和LogitBoost集成分类器的地震事件分类识别算法[J]. 地震工程学报, 2022, 44(5): 1233-1242.
MENG J, ZHANG J S, LI Y N. Classification and recognition algorithm for earthquake events based on the improved EWT and LogitBoost ensemble classifier[J]. China Earthquake Engineering Journal, 2022, 44(5): 1233-1242.