基于GMM-RBF神经网络的前列腺癌诊断方法

 前列腺癌是近年来发病率上升速度最快的男性癌症，严重威胁着患者的身体健康，准确地判断癌症患者的患病情况对于节约医疗资源、提高患者满意度起着至关重要的作用。近年来，基于数据挖掘的癌症诊断方法逐渐成为疾病诊断领域的研究热点，在提高诊断准确性上显示出极大优势。        针对现有前列腺癌早期诊断方法准确性不高的问题，提出一种基于高斯混合模型改进径向基函数神经网络的前列腺癌诊断方法--GMM-RBF神经网络方法。该方法通过使用高斯混合模型对径向基函数神经网络中径向基函数的参数进行预训练，使模型避免陷入局部最优，之后采用改进的粒子群优化算法对神经网络进行训练。采用国家临床医学科学数据中心提供的数据进行前列腺癌诊断实验，将所提出的方法与径向基神经网络、分类回归树、支持向量机和逻辑回归等主流的机器学习算法进行对比，并使用准确性、特异性、敏感性和AUC值对模型的性能进行评价。        研究结果表明，与改进前的神经网络模型相比，GMM-RBF神经网络模型收敛速度更快、初始准确度更高；与其它机器学习算法相比，GMM-RBF神经网络模型在10折交叉验证中取得了较高的准确性、敏感性、特异性和AUC值。        GMM-RBF神经网络方法在模型预测精度上比传统的径向基函数神经网络模型有很大提升，能够得到更为可靠的前列腺癌诊断结果，为医疗工作者初步诊断前列腺癌和穿刺活检操作提供有效的辅助决策支持，该方法的提出对于减少患者痛苦、提高患者满意度和节约医疗资源具有实际意义。

Prostate Cancer Diagnosis Method Based on GMM-RBF Neural Network

 Prostate cancer is the fastest rising incidence of male cancer in recent years, which is a serious health threat to the patients. How to diagnose the condition of cancer patients more accurately is very important for the timely treatment and reduction of the mortality of pros-tate cancer. In recent years, cancer diagnosis based on data mining has gradually become a research focus in the field of disease diagnosis, and it has shown great advantages in improving the accuracy of diagnosis.        In order to solve the problem that the low accuracy of the existing methods for early diagnosis of prostate cancer, this paper presents a new diagnosis method called GMM-RBF neural network based on improved RBF neural network with GMM. In this method, the parameters of radial basis function in radial basis function neural network are pre-trained by using Gaussian mixture model to avoid the model getting into local optimum. Then, the improved PSO algorithm is used to train the neural network. In the experiment, the data provided by the National Clinical Medical Science Data Center is used to compare the proposed method with the other popular machine learning methods such as RBF neural network, classification and regression tree, support vector machine and logistic regression. The performance of the model is evaluated using accuracy, specificity, sensitivity, and AUC.        The experimental results show that the GMM-RBF neural network model has faster convergence rate and higher initial accuracy than the pre-improved neural network model. Compared with other machine learning algorithms, the GMM-RBF neural network model achieves a higher accuracy, sensitivity, specificity and AUCduring ten-fold cross-validation.        In this paper, the proposed GMM-RBF neural network method has a great improvement on the model prediction accuracy compared with the traditional RBF neural network model, which can provide more reliable results for the diagnosis of prostate cancer. It provides effective auxiliary decision-making support for the preliminary diagnosis of prostate cancer for medical workers and has practical significance to reduce the pain of patients, improve patient satisfaction and save medical resources.