摘 要：利用机器学习辅助提高医生诊断效率是常用的方法。用机器学习方法进行乳腺癌检测，常由于乳腺癌数据的不平衡而出现问题。为了解决这一问题，研究在多层感知机的基础上嵌入监督对比学习进行乳腺癌检测，该方式通过数据增广，弥补不平衡数据的影响，同时利用同一类特征距离拉近，反之拉远的性质，增强特征表示效果，提高诊断准确率。实验结果证明，与现有的算法相比在准确率等方面优于其他算法，这证明了该算法的有效性。
关键词：监督对比学习；乳腺癌检测；多层感知机；不平衡数据
DOI:10.19850/j.cnki.2096-4706.2023.02.019
中图分类号：TP181 文献标识码：A 文章编号：2096-4706（2023）02-0079-05
Breast Cancer Detection Algorithm Based on Supervised Contrastive Learning
LI Xin
(Taiyuan Normal University, Jinzhong 030619, China)
Abstract: Using Machine learning is a common method to assist doctors with improving the diagnosis efficiency. The use of machine learning methods for breast cancer detection often causes problems due to the imbalance of breast cancer data. In order to solve this problem, supervised contrastive learning is embedded on the basis of multi-layer perceptron for breast cancer detection. This method makes up for the influence of unbalanced data through data augmentation, and makes use of the nature of the distance narrowing of features of the same class, and the nature of the distance stretching on the contrary, to enhance the feature representation effect and improve the diagnostic accuracy. The experimental results show that compared with the existing algorithms, the algorithm is superior to other algorithms in terms of accuracy and other aspects, which proves the effectiveness of the algorithm.
Keywords: supervised contrastive learning; breast cancer detection; multi-layer perceptron; unbalanced data
参考文献：
[1] BRAY F，FERLAY J，SOERJOMATARAM I，et al.Global Cancer Statistics 2018：Globocan Estimates of incidence and Mortality Worldwide for 36 Cancers in 185 Countries [J].CA Cancer J Clin， 2018，68（6）：394-424.
[2] HASTIE T，TIBSHIRANI R，FRIEDMAN J.The Elements of Statistical Learning [M].Springer Series in Statistics.Berlin：Springer， 2009.
[3] ISA I S，SAAD Z，OMAR S，et al.Suitable MLP Network Activation Functions for Breast Cancer and Thyroid Disease Detection [C]//2010 Second International Conference on Computational Intelligence，Modelling and Simulation.Bali：IEEE，2010：39-44.
[4] CHEN T，KORNBLITH S，NOROUZI M，et al.A Simple Framework for Contrastive Learning of Visual Representations [J/ OL].arXiv：2002.05709 [cs.LG].[2022-08-03].https://arxiv.org/ abs/2002.05709v2.
[5] TIAN Y L，SUN C，POOLE B，et al.What makes for good views for Contrastive Learning? [J/OL].arXiv：2005.10243 [cs.CV]. [2022-08-06].https://arxiv.org/abs/2005.10243.
[6] HE K，FAN H，WU Y，et al.Momentum Contrast for Unsupervised Visual Representation Learning [J].Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition，2020：9726-9735.
[7] HU X M，ZHANG C W，XU Y S，et al.SelfORE：SelfSupervised Relational Feature Learning for Open Relation Extraction [J/OL].arXiv：2004.02438 [cs.CL].[2022-08-09].https://arxiv.org/ abs/2004.02438.
[8] KHOSLA P，TETERWAK P，WANG C，et al.Supervised Contrastive Learning [J/OL].arXiv：2004.11362 [cs.LG].[2022-08-07].https://arxiv.org/abs/2004.11362.
[9] YANG S，WANG S D，WANG Y Q，et al.MB-Supcon： Microbiome-based Predictive Models via Supervised Contrastive Learning [J].Journal of Molecular Biology，2022，434（15）：167693.
[10] WANG J，YANG C C，YAN L H，et al.Speech Enhancement Algorithm of Binary Mask Estimation Based on a Priori SNR Constraints [C]//2018 Asia-Pacific Signal and Information Processing Association Annual Summit and Conference （APSIPA ASC）.Honolulu：IEEE，2018：937-943.
[11] 王成，刘亚峰，王新成，等 . 分类器的分类性能评价指标 [J]. 电子设计工程，2011，19（8）：13-15+21.
作者简介：栗鑫（1995—），男，汉族，山西长治人，硕士研究生在读，研究方向：机器学习。