摘要
针对老年病临床早期精准诊断效率低,老年患者的健康管理意识薄弱等问题,文中提出了一种基于模糊聚类算法的健康数据评价分析模型。在模型的设计过程中,对模糊数学理论进行研究,结合Takagi-Sugeno推理算法与神经网络在模式识别上的优越性,引入了自适应神经网络模糊推理模型。为了验证该模型的有效性,采集723家医疗机构的数据进行搜集与清洗,获取脑卒中与心力衰竭两种典型疾病的相关数据。仿真与测试结果表明,在脑卒中与心力衰竭的预测上,模型的AUC与精确度分别可以达到0.86/0.93、0.81/0.89。在预测过程中,通过数据分析提取了对两个疾病影响较大的8个特征对应的ICD编码,为疾病的风险防范提供了基础数据支持。
In view of the low efficiency of clinical accurate diagnosis of geriatric diseases and the weak awareness of health management of elderly patients,this paper proposes a health data evaluation and analysis model based on fuzzy clustering algorithm.In the process of model design,the theory of fuzzy mathematics is studied.Combining the advantages of Takagi-Sugeno reasoning algorithm and neural network in pattern recognition,the adaptive neural network fuzzy reasoning model is introduced.In order to verify the effectiveness of the model,723 medical institutions were collected for data collection and cleaning,and two typical diseases of stroke and heart failure were obtained.The simulation and test results show that the AUC and accuracy of the model can reach 0.86/0.93 and 0.81/0.89 respectively.In the process of prediction,this paper extracts eight features corresponding to ICD codes that have the greatest impact on two diseases through data analysis,providing basic data support for disease risk prevention.
作者
谈笑
TAN Xiao(Shaanxi National Defense College of Industrial Technology,Xi’an 710300,China)
出处
《电子设计工程》
2021年第3期13-17,共5页
Electronic Design Engineering
基金
陕西省职业技术教育学会项目(SGKCSZ2020-894)。
