...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Journal of Theoretical and Applied Information Technology >AN ELM PREDICTIVE MODEL FOR RISK ASSESSMENTS OF CVD IN IMPAIRED GLUCOSE TOLERANCE (IGT) PATIENTS VIA GENPCNN AND SLFNS ALGORITHM
【24h】

AN ELM PREDICTIVE MODEL FOR RISK ASSESSMENTS OF CVD IN IMPAIRED GLUCOSE TOLERANCE (IGT) PATIENTS VIA GENPCNN AND SLFNS ALGORITHM

机译：通过GenPCNN和SLFNS算法对糖耐量减低（IGT）患者的CVD风险评估的ELM预测模型

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

摘要

Medical diagnosis systems play a vital role in medical practice and are used for diagnosis and treatment by several medical practitioners. Diagnosing the risk factors of pre-diabetic (IGT) cases is quite difficult. There is a big challenge to improve the diagnosis system to recognize the risks factors of impaired glucose tolerance regarding to cardiac vascular disease. In this paper, ELM classifier is combined with the hybrid of genetic algorithm and pulse coupled neural network (GENPCNN). Especially, a Single-hidden layer feed forward neural networks are suitable for solving the complex classification problem. The datasets we collected from health care centre having 270 instances of pre-diabetic, Diabetic and non-diabetic data each was having 28 attributes. A combination of genetic algorithm based neural networks to select the features from the dataset. So, it will be reduced to 14 attributes. The best population of the GA will be passed as input for the PCNN. The features extracted from the GENPCNN are passed to ELM classifier SLFNs in which the hidden nodes are chosen randomly and logically determines the output weight. First, dataset is preprocessed in order to remove the noisy data, missing values or irrelevant values and also from ‘curse of dimensionality’ which have to make suitable for training. This algorithm tends to provide good generation performance and extremely fast learning speed. The classification accuracy obtained using this approach is 94%. The obtained results have shown very promising outcomes for the prediction of risk factors of CVD in impaired glucose tolerance and impaired fasting glucose.
机译：医学诊断系统在医学实践中起着至关重要的作用，并被一些执业医生用于诊断和治疗。诊断糖尿病前期（IGT）病例的危险因素非常困难。改善诊断系统以识别关于心血管疾病的糖耐量受损的风险因素面临着巨大挑战。本文将ELM分类器与遗传算法和脉冲耦合神经网络（GENPCNN）混合在一起。特别地，单隐藏层前馈神经网络适合解决复杂的分类问题。我们从医疗中心收集的数据集有270个糖尿病前，糖尿病和非糖尿病数据实例，每个数据集都有28个属性。基于遗传算法的神经网络的组合，可从数据集中选择特征。因此，它将减少为14个属性。遗传算法的最佳种群将作为PCNN的输入传递。从GENPCNN中提取的特征将传递到ELM分类器SLFN，在其中，随机选择隐藏节点并在逻辑上确定输出权重。首先，对数据集进行预处理，以去除嘈杂的数据，缺失值或不相关的值，并从必须适合训练的“维数诅咒”中删​​除。该算法倾向于提供良好的生成性能和极快的学习速度。使用此方法获得的分类精度为94％。获得的结果已显示出非常有希望的结果，用于预测糖耐量降低和空腹血糖受损的CVD危险因素。

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