当前位置: X-MOL 学术Comput. Electr. Eng. › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Combination of loss functions for robust breast cancer prediction
Computers & Electrical Engineering ( IF 4.0 ) Pub Date : 2020-06-01 , DOI: 10.1016/j.compeleceng.2020.106624
Hamideh Hajiabadi , Vahide Babaiyan , Davood Zabihzadeh , Moein Hajiabadi

Abstract Cancer detection can be formulated as a binary classification in a machine learning paradigm. Loss functions are a critical part of almost every machine learning algorithm. While each loss function comes up with its own advantages and disadvantages, in this paper, inspired by ensemble methods, we propose a novel objective function that is a linear combination of single losses. We then integrate the proposed objective function into an Artificial Neural Network (ANN) to diagnose breast cancer. By doing so, both the coefficients of loss functions and weight parameters of the ANN are learned jointly via backpropagation. As the patients’ data are sometimes very noisy, we evaluate our method by doing comprehensive experiments on Wisconsin Breast Cancer Diagnosis (WBCD) dataset at different noise levels. The experiments show its performance declines very slowly (from 0.97 to 0.96) compared to the peer methods with the increase of noise level.

中文翻译:

用于稳健乳腺癌预测的损失函数组合

摘要 癌症检测可以表述为机器学习范式中的二元分类。损失函数是几乎所有机器学习算法的关键部分。虽然每个损失函数都有自己的优点和缺点,但在本文中,受集成方法的启发,我们提出了一种新颖的目标函数,它是单个损失的线性组合。然后,我们将提出的目标函数集成到人工神经网络 (ANN) 中以诊断乳腺癌。通过这样做,神经网络的损失函数系数和权重参数都是通过反向传播联合学习的。由于患者的数据有时非常嘈杂,我们通过在不同噪声水平下对威斯康星州乳腺癌诊断 (WBCD) 数据集进行综合实验来评估我们的方法。
更新日期:2020-06-01
down
wechat
bug