人工神经生长细胞结构网络在医疗诊断的应用  

Applications of Growing Cell Structures of Artificial Neural Network for Medical Diagnosis

作　　者：刘淑英[1] 程国建[2] 彭方[1] 

机构地区：[1]咸阳师范学院计算机系,陕西咸阳712000 [2]西安石油大学计算机学院,陕西西安710065

出　　处：《计算机技术与发展》2009年第5期231-234,共4页Computer Technology and Development

基　　金：国家自然科学基金资助项目(40572082);咸阳师范学院院内基金资助项目(07XSYK261)

摘　　要：神经网络在医疗智能诊断、细胞图像识别、信息处理等方面的应用效果显著,具有广阔的发展前景。近年来人们对自身健康的重视,人工神经网络在医疗诊断方面应用的研究,正逐渐成为人们关注的焦点。与常用的BP神经网络相比,人工神经网络生长细胞结构(GCS)网络可通过降维映射用于高维数据的聚类和可视化,其网络生成与网络结构和大小都由输入的数据加以确定。研究中采用GCS网络对患者的乳腺癌症细胞的病变类别做出了较为精确的诊断,同时进行了特性分析。并结合实际问题,对某医院的癌症测试指标数据库中699例患者数据进行试验。结果表明:该模型是一种操作简便、易于实现、性能良好的有效模型。Neural networks are significant in the application of medical diagnosis, cell image recognition, information processing and so on, with broad prospects for development. As people in recent years focus the importance of their health, artificial neural network applications in medical diagnostic research is gradually becoming the focus of attention. In comparison with BP model, growing cell structures of artifi- cisl neural network can be used in high- dimensional mapping data clustering and visualization with drop- dimensional mapping, and generation network is determined by the network structure and size of the input data. In this paper, the growing cell structures of artificial neural network is proposed for the processing parameter, it makes a more accurate diagnosis to breast cancer cells of the type of lesion and analyzes the characteristics on the network. Simulation on the testing data about the mammary cancer of the 699 sufferer shows that this algorithm is practicable and effective for actual reservoir modeling.

关 键 词：生长细胞结构 聚类 可视化 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]