加载中........
×

中医中药临床研究特点与分析(5)

2015-7-4 作者:MedSci   来源:MedSci原创 我要评论4
Tags:
分享到:

表7 用最大方差旋转后的结果
Ratation Method: Varimax
Rotated Factor Pattern

  FACTOR1FACTOR2FACTOR3FACTOR4FACTOR5FACTOR6FACTOR7
X1-0.062450.09920-0.014970.011700.02965-0.522080.69946
X20.312840.473560.382970.13580-0.303220.169460.28832
X3-0.17773-0.19200-0.11119-0.081040.191580.149630.25128
X4-0.13269-0.14422-0.15737-0.015010.69796-0.015960.04959
X5-0.17938-0.21177-0.03035-0.228570.687580.130860.12703
X6-0.02530-0.08450-0.172250.064750.76330-0.11227-0.06947
X7-0.147330.049190.01089-0.167570.85082-0.16959-0.04231
X8-0.16971-0.20456-0.21945-0.237080.74389-0.25076-0.00352
X9-0.14366-0.17139-0.13784-0.173780.87694-0.020780.00422
X10-0.10847-0.15775-0.27138-0.272710.65952-0.063810.05083
X11-0.17802-0.08509-0.135400.71244-0.054700.039050.11520
X12-0.15806-0.16017-0.126980.80888-0.03059-0.150880.12370
X13-0.13129-0.15894-0.110350.86245-0.14046-0.00676-0.01933
X14-0.15416-0.02643-0.162340.82347-0.13623-0.121970.01558
X15-0.105110.027130.077160.63256-0.09744-0.07457-0.07479
X160.04878-0.10513-0.110080.753230.00831-0.07448-0.31314
X17-0.13955-0.15942-0.111690.87826-0.10333-0.083080.03964
X18-0.14037-0.166660.053820.73416-0.16796-0.083150.15436
X19-0.13098-0.168980.662780.006340.07020-0.03612-0.11285
X20-0.02514-0.170770.757620.062390.18627-0.144200.12874
X21-0.14359-0.052790.75331-0.13224-0.018470.05489-0.24698
X22-0.15270-0.054970.65062-0.01759-0.15204-0.105230.08826
X230.077510.066720.49343-0.24906-0.30559-0.15385-0.28155
X24-0.12474-0.162540.85747-0.17044-0.00825-0.208460.03920
X25-0.14086-0.169370.79189-0.02309-0.187450.028330.04644
X26-0.17647-0.190830.77289-0.17914-0.189900.151760.09265
X27-0.13279-0.154220.86015-0.163970.00070-0.207990.05489
X280.78951-0.09797-0.09749-0.08246-0.09867-0.080530.03229
X290.79031-0.00189-0.09192-0.08787-0.111950.11394-0.06613
X300.96221-0.10807-0.10072-0.113590.12063-0.100680.01203
X310.87972-0.07479-0.07066-0.09191-0.13017-0.09542-0.08905
X320.96221-0.10807-0.10072-0.11359-0.12063-0.100680.01203
X330.79749-0.105800.01633-0.14135-0.06141-0.01773-0.15558
X340.75968-0.13054-0.116980.03794-0.025640.05090-0.04497
X350.68666-0.14848-0.13638-0.000860.07241-0.016720.02839
X360.88266-0.11142-0.09950-0.119000.01177-0.135540.06197
X370.89554-0.12100-0.12774-0.14008-0.00817-0.119310.06353
X380.96221-0.10807-0.10072-0.11359-0.12063-0.100680.01203
X390.40972-0.173950.07549-0.13283-0.198580.126200.033989
X40-0.078270.61666-0.21300-0.12472-0.28374-0.187370.18011
X410.035280.869370.04620-0.12372-0.13685-0.14034-0.15547
X42-0.120640.81353-0.06667-0.10586-0.131700.012580.15111
X43-0.116650.89402-0.10428-0.107680.01752-0.12253-0.07788
X44-0.130390.74934-0.077380.011480.04020-0.170050.06373
X45-0.115120.75658-0.127850.02541-0.07516-0.032180.07889
X46-0.124340.863390.04028-0.13699-0.14773-0.047520.00202
X470.029840.86314-0.09087-0.11635-0.00343-0.15660-0.06027
X48-0.124900.84121-0.072030.02647-0.11887-0.02607-0.01005
X49-0.121320.73472-0.14034-0.145810.00274-0.016470.14430
X50-0.031820.75018-0.11873-0.04883-0.04291-0.095440.08917
X51-0.144610.75212-0.11587-0.00182-0.044310.09531-0.06635
X52-0.179530.55048-0.03960-0.03045-0.08535-0.04549-0.14377
X53-0.04971-0.17155-0.02453-0.120580.001300.73953-0.04679
X54-0.215960.05834-0.15809-0.15368-0.015330.781530.10120
X550.03912-0.24650-0.214280.17168-0.160500.695890.24996
X56-0.19395-0.197360.24560-0.05270-0.197610.67134-0.06766
X570.07609-0.20411-0.03643-0.14941-0.043380.665420.11201
X58-0.129150.11855-0.27371-0.11133-0.318520.482710.05074
X59-0.18473-0.039340.03112-0.08266-0.105540.75723-0.17403
X60-0.06272-0.20120-0.13354-0.000950.055920.75884-0.18211
X61-0.032000.145820.094050.100140.05766-0.015340.85592
X62-0.314750.03563-0.10272-0.109120.239000.35331-0.15774
X63-0.155360.022640.438550.20841-0.225900.099970.07932
X64-0.32352-0.33074-0.41432-0.39550-0.41868-0.43633-0.22665
X65-0.32352-0.33074-0.41432-0.39550-0.41868-0.43633-0.22665
X66-0.32352-0.33074-0.41432-0.39550-0.41868-0.43633-0.22665

根据表7数据,为每个公因子选择出所支配的指标(按贡献大小排序):
F1(第1公因子):X30(内热便秘),X38(舌红),X32(饮不解渴),X31(尿黄短少),X37(脉细弦数),X36(耳鸣聋),X29(口燥咽干),X28(形弱消瘦),X33(少眠心烦),X34(五心烦热),X35(喜凉饮)。
第1公因子支配以上12项指标,称为F1质,或称为阴虚燥热质。
F2(第2公因子):X43(唇淡口和),X41(面色不华),X46(大便稀糖),X47(夜尿清长),X48(毛发易落),X42(形寒怕冷),X45(肌冷自汗),X51(舌淡胖),X44(四肢冷),X50(脉沉无力),X49(喜热饮),X40(形体白胖),X52(齿嫩印)
第2公因子支配以上13项指标,称为F2质,或称为阴虚衰冷质。
F3(第3公因子):X27(舌苔多腻),X24(口干不饮),X25(胸满昏眩),X26(脉濡或滑),X20(中脘痞满),X21(口甜粘),X19(体形肥胖),X22(身重如裹),X23(大便不实),X63(舌淡)。
第3公因子支配以上10项指标,称为F3质,或称为痰湿郁滞质。
F4(第4公因子):X17(脉沉涩缓),X13(眼眶暗黑),X14(肌肤甲错),X12(口唇紫暗),X16(痞闷作胀),X18(舌质青紫),X11(肤色晦暗),X15(丝缕斑闪)
第4公因子支配以上8项指标,称为F4质,或称为血瘀晦涩质。
F5(第5公因子):X9(脉沉有力),X7(口微干),X6(耐寒暑),X8(二便调),X4(面色红润),X5(胃纳佳),X10(舌正)。
第5公因子支配以上7项指标,称为F7质,或称为平秘调和质。
F6(第6公因子)X54(气短懒言),X60(手易麻),X59(盆腔脏器下坠感),X53(面色白光白),X55(乏力晕眩),X57(脱肛感),X56(心悸健忘),X58(动辄汗出),X62(脉细弱无力)。
第6公因子支配以上13项指标,称为F6质,或称为气虚怠惰质。
F7(第7公因子):X61(月经淡少),X1(性别),X2(年龄),X3(体壮力强);X64(皮肤斑贴试验阳性),X65(皮肤划痕试验阳性),X66(皮内试验阳性)
结合专业知识,把X61(月经淡少),X1(性别),X2(年龄),X3(体壮力强)删除。则F7质可称为敏感质。  

表8 计算7个公因子的标准化得分系数
Standardized Scoring Coefficients
 FACTOR1FACTOR2FACTOR3FACTOR4FACTOR5FACTOR6FACTOR7
X1-0.62441-0.39193-0.16605-0.07860-0.84623-0.982051.85482
X21.231270.711020.117760.061321.557341.99064-2.03620
X30.912490.828130.526080.349731.474541.59763-0.92471
X4-3.97518-3.27768-1.69669-1.25040-5.75278-6.914213.95627
X5-0.85080-0.51002-0.13200-0.07635-1.02393-1.267011.65172
X6-0.33484-0.41768-0.37535-0.31123-0.38044-0.77212-0.29572
X70.952760.783170.454610.206831.542381.37576-1.53847
X8-0.77344-0.57658-0.18458-0.13527-0.94863-1.278141.17617
X93.967483.240431.904111.592556.015266.96562-3.15370
X101.267931.150640.578740.393112.119882.26594-1.02029
X11-3.32310-1.695760.069750.40352-4.19032-4.117777.40201
X125.470313.913911.718951.256497.716328.72161-8.05129
X13-2.21089-1.163630.128500.39703-2.77714-2.532574.58972
X14-0.92464-0.403610.139960.39489-1.18498-0.824772.08080
X15-0.52647-0.34468-0.080080.11195-0.72222-0.811501.07185
X16-0.300280.052340.254840.41762-0.25466-0.429101.57615
X17-0.04248-1.11949-1.82219-1.55982-0.75730-2.52068-3.88253
X181.713681.106740.434730.413742.321302.62774-2.60750
X19-2.67849-1.54052-0.39994-0.28518-3.58156-3.975414.11853
X201.073790.501020.20673-0.007221.246621.14722-2.44193
X21-0.98830-0.70783-0.20085-0.16764-1.34682-1.225151.47833
X220.813500.538340.178910.061221.040320.95247-0.74271
X231.286070.920330.543470.268531.758892.03840-2.11053
X24-3.44563-3.15401-1.93133-1.65100-5.26907-6.355091.84840
X25-1.34159-0.71455-0.07544-0.05688-1.79204-2.222022.96678
X260.01616-0.051980.244180.134110.062771.100880.56180
X275.884754.925933.335572.293939.019559.68944-5.50277
X284.085012.679511.088030.686235.570376.49398-6.25926
X291.933991.371740.691590.411752.584013.28347-3.40459
X304.593214.110073.582792.850197.041767.60477-0.42632
X310.600600.450910.169220.083000.851960.73313-0.73509
X320.000000.000000.000000.000000.000000.000000.00000
X33-1.23921-0.448340.330970.27859-1.48648-1.098513.03795
X343.049721.847820.453160.249173.913153.94647-5.19096
X353.012861.972000.784100.578994.091244.56040-4.09893
X36-9.46459-7.51720-4.58508-3.26003-14.20568-15.8329811.00858
X37-2.75959-2.29804-1.23416-0.92745-4.20354-4.932732.73161
X380.000000.000000.000000.000000.000000.000000.00000
X39-3.32481-2.25865-0.85539-0.56961-4.73914-5.599995.33527
X40-0.248990.03146-0.03541-0.01914-0.26338-0.190960.77200
X41-2.57853-1.70620-0.96367-0.64739-3.80328-4.825433.28267
X421.752301.516410.810610.481792.614902.95074-2.44996
X430.36418-0.20871-0.75346-0.589820.08419-0.35486-2.47613
X44-0.36384-0.032470.258320.22869-0.32690-0.076660.62848
X450.749641.409591.296821.057491.599532.375081.60736
X46-0.123860.135530.034830.02531-0.20477-0.579810.90787
X470.000000.000000.000000.000000.000000.000000.00000
X480.000000.000000.000000.000000.000000.000000.00000
X490.000000.000000.000000.000000.000000.000000.00000
X500.000000.000000.000000.000000.000000.000000.00000
X510.000000.000000.000000.000000.000000.000000.00000
X520.138050.148310.031850.006880.254640.29915-0.32968
X530.267280.11814-0.00997-0.046480.341620.69663-0.43735
X540.000000.000000.000000.000000.000000.000000.00000
X550.000000.000000.000000.000000.000000.000000.00000
X560.000000.000000.000000.000000.000000.000000.00000
X570.000000.000000.000000.000000.000000.000000.00000
X580.000000.000000.000000.000000.000000.000000.00000
X590.000000.000000.000000.000000.000000.000000.00000
X600.000000.000000.000000.000000.000000.000000.00000
X610.000000.000000.000000.000000.000000.000000.00000
X620.000000.000000.000000.000000.000000.000000.00000
X630.000000.000000.000000.000000.000000.000000.00000
X640.000000.000000.000000.000000.000000.000000.00000
X650.000000.000000.000000.000000.000000.000000.00000
X660.000000.000000.000000.000000.000000.000000.00000
先计算各指标的标准化值:X' =(X – X )/S,作为各公因子标准化得分系数。
根据表8,可写出7个公因子的标准化得分公式:
F1 = -0.62441X '1+1.23127X'2+……+ 0.00000X'66
……………
F7 = 1.85482X,1-2.03620X,2+……+ 0.00000X,66
计算因子得分的用途:把任何一人的各项指标分别代入F1—F7公因子标准化得分公式,哪个公因子的标准化得分之代数和最高,就可以诊断为该公因子(该体质),从而建立了体质的计量诊断与鉴别诊断函数。

在进行因子分析时我们总是希望:
  • 保留的公因子个数q远小于原始指标个数m,一般按以下原则来确定:①若特征值λi≥1,则保留其对应的公因子;②若前k个公因子累积贡献率达到一特定的数量(一般认为达到70%以上为宜),则保留前k个公因子,使m个原始指标的总基本上能被所保留的公因子解释。
  • 各共性方差hi2接近于1,即各原始指标Xi的约大部分能由所保留的公因子解释。
  • 各原始指标在同一公因子Fj上的因子载荷的绝对值|aij|(I=1,2,3,…,m,即竖读因子载荷阵)之间的判别应尽可能大,使得公因子Fj的意义主要由一个或几个|aij|值大的原始指标所表达。
当然,必须同时结合专业上的考虑,不然结果可能无法用专业作出合理的解释(可参考本文作者著作《中医临床研究设计与SAS编程统计分析》)。

讨论  

1、本章设计的例题仅是虚构的47人的数据,只为演示方法用,其结论仅供参考。    

2、选择的指标应该是真正反映体质的,得到的公因子才是体质。如果指标既反映体质又反映证侯,则得到的公因子是体质和证侯的混合物。在提出体质的概念假设阶段,就包括体质与证侯的区别点。

参考文献

  1. 王家良主编。临床流行病学,第二版。上海:上海科学技术出版社,2001。
  2. 金丕焕、苏炳华、贺佳主编。医用SAS统计分析。上海:复旦大学出版社,2000
  3. 孙振球主编。医学统计学。北京:人民卫生出版社,2002,338—340页。
  4. 孙爱琴:医用统计基本方法。摘自:顾婉先、张永祥主编。高等医药院校选用教材《预防医学概论》(第二版),上海:上海科学技术出版社,1996年,231页。
  5. 方积乾主编。医学统计学与电脑实验。上海科学技术出版社,第二版,2001;第203页程序11-1.
  6. 杨建伯。样本含量计算。摘自中国医学百科全书。上海:上海科学技术出版社,1985,51-52页。
  7. 胡良平编。Windows SAS 6.12 and 8.0 实用统计分析教程。北京:军事医学科学出版社,2001,531-532页。



小提示:78%用户已下载梅斯医学APP,更方便阅读和交流,请扫描二维码直接下载APP

版权声明:本文系梅斯MedSci原创编译整理,未经本网站授权不得转载和使用。如需获取授权,请点击
分享到:

只有APP中用户,且经认证才能发表评论!马上下载

hlycom3356

希望以后可以实现中西医的更好结合

2015-8-6 10:39:00 回复

hlycom3356

中药的确有很多价值

2015-8-6 10:39:00 回复

hlycom3356

很不错哦的

2015-8-6 10:38:00 回复

huaxipanxing

看看

(来自:梅斯医学APP)

2015-7-8 12:37:00 回复

web对话
live chat

扫码领取IF曲线

IF连续增长的期刊