面向乳腺癌辅助诊断的改进支持向量机方法

根据针吸细胞学方法影像中提取的特征值, 设计了一种改进的支持向量机分类方法, 并应用于乳腺癌的辅助诊断。通过对几种常用核函数的对比分析, 所建立的新核函数在诊断中具有很好的综合性能。使用实际临床数据分析显示, 该方法比模因佩雷托(memetic Pareto artificial neural network, MPANN)与一种改进型人工神经网络(evolutionary artificial neural network, EANN)方法在乳腺癌辅助诊断中具有更好的效果, 可以为医疗机构对该疾病的诊断提供有力的决策支持。

Improved method for computer-aided diagnosis of breast cancerbased on support vector machines

According to features from a digitized image of a fine needle aspirate, this paper proposed an evolutionary classification method based on support vector machines for the disease diagnosis. Through the contrastive analysis using some common kernel functions, it showed experimentally that the new created kernel function has better integrative capability than original kernel functions. Compared with MPANN and EANN approach, this method has more effective in computer-aided diagnosis of breast cancer using the same clinical data, which can support the medical domain efficiently.