PB实验设计结果分析
PB实验：我做的十一因子（其中有两个空白项）十二次实验，结果分析中除了三个因子显著外，两个空白项也显著， 如果将两个空白项设置为误差项，可以得出三个显著因子，但是校正拟合度与预测拟合度相差较大，请教一下这是什么原因呢，是实验误差太大了吗？
空白项显著，应该是误差较大。校正拟合度与预测拟合度相差较大，说明拟合的情况不理想。楼主还是分析原因，控制误差。然后重新进行试验。
回 1楼(细履平沙) 的帖子
嗯，看实验结果，也觉得过程中出现误差较大，太感谢了，准备重新做一次实验，再看看结果如何。 楼主别忘了回来反馈一下结果吆。 一 结果分析1Use your mouse to right click on individual cells for definitions.
Response 1 Viable cells
ANOVA for selected factorial model
Analysis of variance table
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 241.98 6 40.33 39.74 0.0005 significant
A-Ethanolamine 55.77 1 55.77 54.96 0.0007
F-Hydrocortisone 51.46 1 51.46 50.71 0.0008
H-Lipoic aciid 12.55 1 12.55 12.36 0.0170
J-Lipid mixture 84.75 1 84.75 83.52 0.0003
K-Dummy 1 13.85 1 13.85 13.65 0.0141
L-Dummy 2 23.60 1 23.60 23.26 0.0048
Residual 5.07 5 1.01
Cor Total 247.05 11
The Model F-value of 39.74 implies the model is significant.There is only
a 0.05% chance that a "Model F-Value" this large could occur due to noise.
Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, F, H, J, K, L are significant model terms.
Values greater than 0.1000 indicate the model terms are not significant.
If there are many insignificant model terms (not counting those required to support hierarchy),
model reduction may improve your model.
Std. Dev. 1.01 R-Squared 0.9795
Mean 13.66 Adj R-Squared 0.9548
C.V. % 7.38 Pred R-Squared 0.8817
PRESS 29.22 Adeq Precision 18.730
The "Pred R-Squared" of 0.8817 is in reasonable agreement with the "Adj R-Squared" of 0.9548.
"Adeq Precision" measures the signal to noise ratio.A ratio greater than 4 is desirable.Your
ratio of 18.730 indicates an adequate signal.This model can be used to navigate the design space.
Coefficient Standard 95% CI 95% CI
Factor Estimate df Error Low High VIF
Intercept 13.66 1 0.29 12.91 14.40
A-Ethanolamine -2.16 1 0.29 -2.90 -1.41 1.00
F-Hydrocortisone -2.07 1 0.29 -2.82 -1.32 1.00
H-Lipoic aciid -1.02 1 0.29 -1.77 -0.27 1.00
J-Lipid mixture 2.66 1 0.29 1.91 3.41 1.00
K-Dummy 1 -1.07 1 0.29 -1.82 -0.33 1.00
L-Dummy 2 1.40 1 0.29 0.65 2.15 1.00
Final Equation in Terms of Coded Factors:
Viable cells =
+13.66
-2.16 * A
-2.07 * F
-1.02 * H
+2.66 * J
-1.07 * K
+1.40 * L
Final Equation in Terms of Actual Factors:
Viable cells =
+16.24750
-1.72467 * Ethanolamine
-8.28333 * Hydrocortisone
-4.09000 * Lipoic aciid
+5.31500 * Lipid mixture
-1.07417 * Dummy 1
+1.40250 * Dummy 2
The Diagnostics Case Statistics Report has been moved to the Diagnostics Node.
In the Diagnostics Node, Select Case Statistics from the View Menu.
Proceed to Diagnostic Plots (the next icon in progression).Be sure to look at the:
1) Normal probability plot of the studentized residuals to check for normality of residuals.
2) Studentized residuals versus predicted values to check for constant error.
3) Externally Studentized Residuals to look for outliers, i.e., influential values.
4) Box-Cox plot for power transformations.
If all the model statistics and diagnostic plots are OK, finish up with the Model Graphs icon.
分析：虚拟项也为显著因素
R-Squared 0.9795
Adj R-Squared 0.9548
Pred R-Squared 0.8817
二 将两个虚拟项设置为误差项，得到结果分析2
Use your mouse to right click on individual cells for definitions.
Response 1 Viable cells
ANOVA for selected factorial model
Analysis of variance table
Sum of Mean F p-value
Source Squares df Square Value Prob > F
Model 191.98 3 63.99 9.30 0.0055 significant
A-Ethanolamine 55.77 1 55.77 8.10 0.0216
F-Hydrocortisone 51.46 1 51.46 7.48 0.0257
J-Lipid mixture 84.75 1 84.75 12.31 0.0080
Residual 55.07 8 6.88
Cor Total 247.05 11
The Model F-value of 9.30 implies the model is significant.There is only
a 0.55% chance that a "Model F-Value" this large could occur due to noise.
Values of "Prob > F" less than 0.0500 indicate model terms are significant.
In this case A, F, J are significant model terms.
Values greater than 0.1000 indicate the model terms are not significant.
If there are many insignificant model terms (not counting those required to support hierarchy),
model reduction may improve your model.
Std. Dev. 2.62 R-Squared 0.7771
Mean 13.66 Adj R-Squared 0.6935
C.V. % 19.21 Pred R-Squared 0.4985
PRESS 123.91 Adeq Precision 9.089
The "Pred R-Squared" of 0.4985 is in reasonable agreement with the "Adj R-Squared" of 0.6935.
"Adeq Precision" measures the signal to noise ratio.A ratio greater than 4 is desirable.Your
ratio of 9.089 indicates an adequate signal.This model can be used to navigate the design space.
Coefficient Standard 95% CI 95% CI
Factor Estimate df Error Low High VIF
Intercept 13.66 1 0.76 11.91 15.40
A-Ethanolamine -2.16 1 0.76 -3.90 -0.41 1.00
F-Hydrocortisone -2.07 1 0.76 -3.82 -0.32 1.00
J-Lipid mixture 2.66 1 0.76 0.91 4.40 1.00
Final Equation in Terms of Coded Factors:
Viable cells =
+13.66
-2.16 * A
-2.07 * F
+2.66 * J
Final Equation in Terms of Actual Factors:
Viable cells =
+15.22500
-1.72467 * Ethanolamine
-8.28333 * Hydrocortisone
+5.31500 * Lipid mixture
The Diagnostics Case Statistics Report has been moved to the Diagnostics Node.
In the Diagnostics Node, Select Case Statistics from the View Menu.
Proceed to Diagnostic Plots (the next icon in progression).Be sure to look at the:
1) Normal probability plot of the studentized residuals to check for normality of residuals.
2) Studentized residuals versus predicted values to check for constant error.
3) Externally Studentized Residuals to look for outliers, i.e., influential values.
4) Box-Cox plot for power transformations.
If all the model statistics and diagnostic plots are OK, finish up with the Model Graphs icon.
分析：得到三个显著因素，但是
R-Squared 0.7771
Adj R-Squared 0.6935
Pred R-Squared 0.4985
请大家帮忙分析一下原因，请问第二个分析结果可用吗？
这个实验是不是存在很大的误差啊？
回 1楼(细履平沙) 的帖子
您好，请教从单因素到PB的问题：是选取单因素中最好的因素做PB 吗？例如，碳源最好的淀粉，氮源是酵母膏。如果是这样的话，做PB实验，里面的空白因素不就很多了吗？谢谢
页:
[1]