上海交通大学学报（医学版）

上海交通大学学报（医学版）

• 论著（基础研究） • 上一篇    下一篇

基于神经网络的左心室后壁厚度参考值与地理因素关系的研究

岑敏仪1，葛 淼1，王聪霞2，何进伟1，杨绍芳1，姜吉琳1，许金辉1，张 雯1，刘新蕾1   

  1. 陕西师范大学 旅游与环境学院健康地理研究所， 西安 710062； 2.西安交通大学 医学院第二附属医院心内科， 西安 710004
  • 出版日期:2014-12-28 发布日期:2014-12-30
  • 通讯作者: 葛 淼, 电子信箱: gemiao@snnu.edu.cn。
  • 作者简介:岑敏仪（1990—）， 女， 硕士生; 电子信箱: minyicen@126.com。
  • 基金资助:

    国家自然科学基金（40971060）

Relationship between reference value of left ventricular posterior wall thickness and geographical factors based on neural network

CEN Min-yi1, GE Miao1, WANG Cong-xia2, HE Jin-wei1, YANG Shao-fang1, JIANG Ji-lin1, XU Jin-hui1, ZHANG Wen1, LIU Xin-lei1   

  1. 1.Institute of Health and Geography, College of Tourism and Environment, Shanxi Normal University, Xi'an 710062, China; 2.Department of Cardiovasology, the Second Affiliated Hospital, Medical School of Xi'an Jiao Tong University, Xi'an 710004, China
  • Online:2014-12-28 Published:2014-12-30
  • Supported by:

    National Natural Science Foundation of China, 40971060

PDF (PC)

1135

摘要：

目的 分析我国中老年人左心室后壁厚度参考值与地理因素之间的关系，为制定其统一标准提供科学依据。方法 收集67个市县3 543例健康中老年人左心室后壁厚度参考值，运用相关分析方法研究其与地势、气候、土壤有关的共18项地理因素之间的关系。选取与左心室后壁厚度参考值有相关性的地理因素进行多元线性回归和BP神经网络建模。借用反距离加权插值法拟合出关于我国中老年人左心室后壁厚度参考值的地理分布图。结果 我国中老年人左心室后壁厚度参考值与经度、海拔、表土砂粒百分率、表土粉粒百分率、表土总可交换量以及表土碱度6项地理因素呈显著的相关关系。经比较分析，BP神经网络比多元回归分析模型具有更好的模拟和预测性能。我国中老年人左心室后壁厚度参考值分布图呈现西高东低的分布特征。结论 若已知某地各项地理指标的值，则可通过构建神经网络模型或绘制地理分布图得出该区的中老年人左心室后壁厚度参考值。

关键词: 心功能, 左心室后壁厚度, 参考值, 地理要素, BP神经网络

Abstract:

Objective To analyze the relationship between the reference value of left ventricular posterior wall thickness (LVPW) of Chinese middle-aged and elderly people and geographical factors and to provide scientific evidences for establishing the uniform standard. Methods Reference values of LVPW of 3 543 Chinese healthy middle-aged and elderly people from 67 cities were collected. The correlation analysis method was adopted to investigate the relationship between the reference value and 18 geographical factors, including terrain, climate, soil, etc. Geographical factors that significantly correlated with the reference value were selected for conducting the multiple linear regression and BP neural network modeling. The spatial distribution map of the reference value of LVPW of Chinese middle-aged and elderly people was fitted by the inverse distance weight method. Results The reference value of LVPW of Chinese middle-aged and elderly people was significantly correlated with the longitude, altitude, percentage of sand in topsoil, percentage of silt in topsoil, total exchangeable capacity of topsoil, and alkalinity of topsoil. The simulation and prediction performance of BP neural network model was better than that of the multiple linear regression model. The spatial distribution map of the reference value of LVPW of Chinese middle-aged and elderly people showed a distribution feature of higher in the west and lower in the east. Conclusion If the geographical factors of a certain area are known, the reference value of LVPW of Chinese middle-aged and elderly people can be obtained by establishing the neural network model or plotting the spatial distribution map.

Key words: cardiac function, left ventricular posterior wall thickness, reference value, geographical factors, BP neural network