• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Canadian Conference on Artificial Intelligence >Detection and Diagnosis of Breast Cancer Using a Bayesian Approach
【24h】

Detection and Diagnosis of Breast Cancer Using a Bayesian Approach

机译：贝叶斯方法检测和诊断乳腺癌

获取原文
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

This paper aims to build a reliable decision support detection system to help physicians give more accurate risk profiles for breast cancer patients. In this quantitative study, we use Bayesian network to uncover hidden insights from the Wisconsin Breast Cancer Diagnostic dataset. A Bayesian network was learned from the data and conditional probability queries were performed. Lastly, a Bayesian network classifier was built. We found diagnosis was conditionally dependent upon two features: worst concave points and worst radius. Both the highest probability for malignant cancer and lowest probability for benign diagnosis were detected by Very High, (0.161, 0.291] for worst concave points and Very High, (18.8, 36] μm for worst radius. The highest probability for benign diagnosis and lowest probability for malignant cancer were detected by Low, [0, 0.0649] for worst concave points and Low, [7.93, 13] μm for worst radius. Our proposed Bayesian network classifier had: 96.31% accuracy, 92.92% sensitivity, 98.32% specificity, 97.04% positive predictive value, and 95.90% negative predictive value, making it a robust model in terms of accuracy, sensitivity, and specificity.
机译：本文旨在建立一个可靠的决策支持检测系统，以帮助医生为乳腺癌患者提供更准确的风险概况。在这项定量研究中，我们使用贝叶斯网络从威斯康星州乳腺癌诊断数据集中发现隐藏的见解。从数据中学习了贝叶斯网络，并进行了条件概率查询。最后，建立了贝叶斯网络分类器。我们发现诊断有条件地取决于两个特征：最差的凹点和最差的半径。通过极高（最坏凹点为（0.161，0.291]）和极高（最差半径为（18.8，36]μm），可以检测到恶性肿瘤的最高概率和最低良性诊断率。低，最坏凹点的发生率低，[0，0.0649]，最半径，低，[7.93，13]μm，检测出恶性肿瘤的可能性。我们提出的贝叶斯网络分类器具有：96.31％的准确性，92.92％的敏感性，98.32％的特异性， 97.04％的阳性预测值和95.90％的阴性预测值，使其成为在准确性，敏感性和特异性方面的可靠模型。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号