系统仿真学报 ›› 2021, Vol. 33 ›› Issue (9): 2279-2288.doi: 10.16182/j.issn1004731x.joss.20-0421

• 国民经济仿真 • 上一篇    

基于降维技术的烟叶质量可视分析方法

田东1,2, 单桂华1,2, 迟学斌1,2, 张艳玲3, 冯伟华3, 王建伟3, 王爱国3, 王锐3,*   

  1. 1.中国科学院 计算机网络信息中心,北京 100190;
    2.中国科学院大学,北京 100190;
    3.中国烟草总公司郑州烟草研究院,郑州 450001
  • 收稿日期:2020-06-29 修回日期:2020-08-31 出版日期:2021-09-18 发布日期:2021-09-17
  • 通讯作者: 王锐(1992-),女,硕士,工程师,研究方向为数据挖掘。E-mail: aboutstefanie@163.com
  • 作者简介:田东(1983-),男,硕士,高工,研究方向为数据可视化。E-mail: tiandong@cnic.cn
  • 基金资助:
    中国烟草总公司烟草科研大数据重大专项(110201901029(SJ-08),110201901025(SJ-04))

Visual Analysis Method of Tobacco Quality Data Based on Dimension Reduction

Tian Dong1,2, Shan Guihua1,2, Chi Xuebin1,2, Zhang Yanling3, Feng Weihua3, Wang Jianwei3, Wang Aiguo3, Wang Rui3,*   

  1. 1. Computer Network Information Center, Chinese Academy of Sciences, Beijing 100190, China;
    2. University of Chinese Academy of Sciences, Beijing 100190, China;
    3. Zhengzhou Tobacco Research Institute, China National Tobacco Corporation, Zhengzhou 450001, China
  • Received:2020-06-29 Revised:2020-08-31 Online:2021-09-18 Published:2021-09-17

摘要: 为了满足烟叶选材过程中跨区域匹配烟叶的需求,研发融合了降维和相关性分析方法的烟叶质量数据可视分析方法。通过基于香型区分类的降维算法、对比算法和针对烟叶质量数据的可视化交互方法,为研究人员提供了对烟叶质量数据进行探索式空间划分和相关性分析的可视化分析手段。全国烟叶质量数据分析案例和专家论证,证明该方法可以较好实现烟叶质量数据分析。

关键词: 烟叶质量, 降维, 相关性, 可视化, 空间划分

Abstract: In order to meet the requirements of tobacco leaf matching across regions in tobacco material selection, a visual analysis method of tobacco leaf quality data that incorporating dimension reduction and correlation analysis methods is developed. Through the dimension reduction algorithm, the comparison algorithm and the visual interaction method based on the classification of aroma area for tobacco leaf quality data, a visual analysis method for exploring space division and correlation analysis of tobacco leaf quality data is provided. National tobacco leaf quality data analysis cases and expert demonstrations show that the method can carry out the tobacco leaf quality data analysis well.

Key words: tobacco quality, dimension reduction, correlation, visualization, spatial division

中图分类号: