河南农业科学 ›› 2023, Vol. 52 ›› Issue (6): 111-119.DOI: 10.15933/j.cnki.1004-3268.2023.06.012

• 园艺 • 上一篇    下一篇

基于3 种分析方法的板栗果实营养品质综合评价

魏源1,2,3,吕梦炀1,2,3,马亚特1,2,3,刘静1,2,3,4,王旋1,2,3,4 ,王东升1,2,3,4   

  1. (1.河北科技师范学院园艺科技学院,河北 秦皇岛 066000;2.板栗产业技术教育部工程研究中心,河北 秦皇岛 066000;3.河北省特色园艺种质挖掘与创新利用重点实验室,河北 秦皇岛 066000;4.河北省板栗产业协同创新中心,河北 秦皇岛 066000)
  • 收稿日期:2023-01-05 出版日期:2023-06-15 发布日期:2023-07-11
  • 通讯作者: 王东升(1987-),男,山东济南人,讲师,博士,主要从事果树育种和栽培生理研究。E-mail:wdsgfly@126.com
  • 作者简介:魏源(1998-),男,河北邯郸人,在读硕士研究生,研究方向:果树育种和栽培生理。E-mail:626385046@qq.com
  • 基金资助:
    板栗产业技术教育部工程研究中心项目(PF2023-01);河北省重点研发计划项目(21326304D);河北省2022年大学生创新创业训练计划项目(S202210798051)

Comprehensive Evaluation of the Nutritional Quality of Chestnut Based on Three Methods

WEI Yuan1,2,3,LÜ Mengyang1,2,3,MA Yate1,2,3,LIU Jing1,2,3,4,WANG Xuan1,2,3,4,WANG Dongsheng1,2,3,   

  1. (1.College of Horticulture Science and Technology,Hebei Normal University of Science and Technology,Qinhuangdao 066000,China;2.Engineering Research Center of Chestnut Industry Technology,Ministry of Education,Qinhuangdao 066000,China;3.Hebei Key Laboratory of Horticultural Germplasm Excavation and Innovative Utilization,Qinhuangdao 066000,China;4.Hebei Collaborative Innovation Center of Chestnut Industry,Qinhuangdao 066000,China)
  • Received:2023-01-05 Published:2023-06-15 Online:2023-07-11

摘要: 通过对燕山地区12个板栗品种(系)的7个品质指标进行测定,对各营养指标进行差异显著性分析,并应用隶属函数法、主成分分析法、K-均值聚类法对果实品质进行综合评价,最后比较不同分析方法对综合评价的影响,以期为选育优良板栗品种提供参考。结果表明:12个板栗品种(系)的果实单粒质量、单粒质量标准差、可溶性糖含量、淀粉含量、脂肪含量、蛋白质含量、含水量差异显著,其中,淀粉含量和脂肪含量存在极显著正相关,可溶性糖含量和含水量存在显著负相关。通过建立各指标的线性隶属函数,根据最低隶属原则进行品质排序,前5名分别为迁西42号、抚宁12号、燕宝、燕龙、兴隆1号。采用SPSS进行主成分分析,根据整体评价指标可分别提取3个主成分,累计方差贡献率为84.647%,根据3个主成分得分及相应权重累加综合计算得分,确定前5名分别为燕宝、兴隆1号、青龙47号、迁西42号、抚宁12号。通过K-均值聚类法,根据果实品质指标对12个板栗品种(系)进行聚类,确定优质板栗品种(系)为燕龙、抚宁8号、燕京8号、燕宝、青龙47号、兴隆1号、迁西42号。3种分析方法对不同板栗品种(系)的算法和评价侧重点有所不同,在3种评价方法中综合品质均优质的有燕宝、兴隆1号和迁西42号3个品种(系)。在实际生产中,使用多种方法进行板栗坚果品质的综合评价可以获得更加准确的结论。

关键词: 板栗, 综合评价, 果实品质, 主成分分析法, 隶属函数法, K-均值聚类法, 种质资源筛选

Abstract: By measuring seven quality indexes of 12 chestnut varieties(lines) in Yanshan area,the significant difference analysis of each nutritional index was conducted. The affiliation function method,principal component analysis and K⁃means clustering method were applied to make a comprehensive evaluation of fruit quality.Finally,the effects of different analysis methods on comprehensive evaluation were compared,in order to provide reference for breeding excellent chestnut varieties.The results showed that there were significant differences in single fruit weight,standard deviation of single fruit weight,soluble sugar content,starch content,fat content,protein content and water content among 12 chestnut varieties(lines).The starch content and fat content showed extremely significant positive correlation,while the soluble sugar content and water content showed significant negative correlation.By establishing the linear affiliation function of each index and ranking the quality according to the minimum membership principle,the top five varieties(lines)were Qianxi 42,Funing 12,Yanbao,Yanlong and Xinglong 1.Three principal components could be extracted according to the overall evaluation index using SPSS,and the cumulative variance contribution rate was 84. 647%. According to the scores of the three principal components and the corresponding weights,the top five varieties(lines) were Yanbao,Xinglong 1,Qinglong 47,Qianxi 42,and Funing 12.The 12 chestnut cultivars(lines)were clustered according to their fruit quality indicators by K⁃means clustering method,and the high⁃quality cultivars(lines)were determined to be Yanlong,Funing 8,Yanjing 8,Yanbao,Qinglong 47,Xinglong 1,and Qianxi 42.The three analysis methods had different algorithms and evaluation emphasis for different chestnut varieties(lines).Based on the three evaluation methods,Yanbao,Xinglong 1 and Qianxi 42 had high comprehensive quality.In actual production,a more accurate conclusion can be obtained by using multiple methods for the comprehensive evaluation of chestnut quality.

Key words: Chestnut, Comprehensive evaluation, Fruit quality, Principal component analysis method, Membership function method, K?mean clustering method, Germplasm screening

中图分类号: