• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Computational methods in science and engineering: advances in computational science >High Performance Medical Classifiers
【24h】

High Performance Medical Classifiers

机译：高性能医疗分类器

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

In this paper, parallelism methodologies for the mapping of machine learning algorithms derived rules on both software and hardware are investigated. Feeding the input of these algorithms with patient diseases data, medical diagnostic decision trees and their corresponding rules are outputted. These rules can be mapped on multithreaded object oriented programs and hardware chips. The programs can simulate the working of the chips and can exhibit the inherent parallelism of the chips design. The circuit of a chip can consist of many blocks, which are operating concurrently for various parts of the whole circuit. Threads and inter-thread communication can be used to simulate the blocks of the chips and the combination of block output signals. The chips and the corresponding parallel programs constitute medical classifiers, which can classify new patient instances. Measures taken from the patients can be fed both into chips and parallel programs and can be recognized according to the classification rules incorporated in the chips and the programs design. The chips and the programs constitute medical decision support systems and can be incorporated into portable micro devices, assisting physicians in their everyday diagnostic practice.
机译：在本文中，研究了将机器学习算法派生规则映射到软件和硬件上的并行方法。向这些算法的输入提供患者疾病数据，医疗诊断决策树及其相应规则。这些规则可以映射到面向多线程的面向对象的程序和硬件芯片上。这些程序可以模拟芯片的工作，并可以展现芯片设计固有的并行性。芯片的电路可以由许多模块组成，这些模块针对整个电路的各个部分同时运行。线程和线程间通信可用于模拟芯片块以及块输出信号的组合。芯片和相应的并行程序构成了医学分类器，可以对新的患者实例进行分类。从患者身上采取的措施既可以输入芯片，也可以并行程序，并且可以根据芯片中所包含的分类规则和程序设计进行识别。这些芯片和程序构成了医疗决策支持系统，可以集成到便携式微型设备中，从而帮助医生进行日常诊断。

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