中文摘要：

在乳腺肿瘤识别优化的研究中,传统的识别方法容易漏诊。为提高乳腺肿瘤识别准确率,提出基于粒子群优化（Particle Swarm Optimization,PSO）参数的支持向量机（Support Vector Machine,SVM）辅助诊断方法。首先采用PSO选择最佳的SVM惩罚系数c,核函数参数γ;然后,利用最佳参数c和γ训练SVM;再利用PSO-SVM实现乳腺肿瘤分类识别,进而实现辅助诊断。将PSO-SVM乳腺肿瘤识别方法的仿真结果与LVQ神经网络识别方法、BP神经网络识别方法的结果做比对分析,表明PSO-SVM具有较高的识别准确率和较低的假阴性率。PSO-SVM乳腺肿瘤辅助诊断,可以提供决策支持,辅助医生尽可能地减少和避免采用传统的细针穿刺细胞病理学检查方法诊断乳腺肿瘤时的漏诊、误诊情况,具有非常重要的价值和意义。

英文摘要：

In order to improve the accuracy of breast tumor recognition, this paper presents a breast tumor com- puter-aided diagnosis model based on PSO-SVM. Firstly, Particle Swarm Optimization （PSO） is used to select the best coefficients of Support Vector Machine （SVM）, including ‘cost＇ and ‘γ’ in kernel function. Secondly, SVM is trained with ‘cost＇ and ‘γ＇ which are the best parameters selected by PSO. Finally, the classification and recog- nition of breast tumors are implemented. Furthermore, computer-aided breast tumor diagnosis is realized. Compared with the results of LVQ and BP neural network breast tumor diagnosis methods, recognition accuracy of PSO-SVM method is higher, meanwhile, false negative rate is lower. As the method of breast tumors computer-aided diagnosis based on PSO-SVM is realized, it can provide decision support and assist doctors to minimize and avoid missing and faulty diagnostic cases when fine needle aspiration cytopathology is used, which is the traditional examination method, to diagnose breast tumors.