遗传算法优化BP神经网络的专家自动诊断模型

Automatic Diagnosis Expert Model Based on GA-BP Neural Network

随着计算机信息化和人工智能的发展迅速,专家自动诊断疾病系统成为各界关注的焦点。卵巢癌是严重威胁妇女健康的恶性肿瘤之一,而且目前无有效的筛选方法和特异的诊断方法。BP人工神经网络因其具有分布式信息存储方式、自适应能力、强大的容错性和非线性处理能力,能有效地对疾病进行筛查和诊断。该文采用遗传算法来优化改进BP算法得到GA-BP算法,通过遗传和变异操作对BP神经网络的初始权值和阈值进行优化,不断更新选择,使得网络的系统总误差趋于最小,构建出卵巢癌诊断模型。采用Matlab2013、VC++编程和统计软件SPSS.18等工具来实现专家自动诊断的人工智能模型,并通过计算机仿真和预测进行检验。

With the development of computer information and artificial intelligence rapidly, automatic disease diagnosis expert system has become the focus of public attention. Ovarian cancer is the malignant tumor of the serious threat to women＇s health,diagnosis methods and currently no screening and specific and effective.BP artificial neural network because of its distributed information storage mode and adaptive ability, strong fault tolerance and nonlinear processing ability,is able to effectively carry out screening and diagnosis of diseases.This paper uses the improved BP algorithm are optimized by genetic algorithm GA--BP algorithm,through the heredity and mutation of BP neural network to optimize the initial weights and threshold and constantly updated to choose,make network system tends to the minimum total error,constructed the model of ovarian cancer diagnosis.With Matlab2013,vc ＋ ＋ programming and statistical software SPSS.18 automatic diagnostic tools for experts such as artificial intelligence model,and through the computer simulation and prediction.