CPE 619 One Factor Experiments
-
Upload
cleo-phelps -
Category
Documents
-
view
62 -
download
0
description
Transcript of CPE 619 One Factor Experiments
![Page 1: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/1.jpg)
CPE 619One Factor Experiments
Aleksandar Milenković
The LaCASA Laboratory
Electrical and Computer Engineering Department
The University of Alabama in Huntsville
http://www.ece.uah.edu/~milenka
http://www.ece.uah.edu/~lacasa
![Page 2: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/2.jpg)
2
Overview
Computation of Effects Estimating Experimental Errors Allocation of Variation ANOVA Table and F-Test Visual Diagnostic Tests Confidence Intervals For Effects Unequal Sample Sizes
![Page 3: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/3.jpg)
3
One Factor Experiments
Used to compare alternatives of a single categorical variable
For example, several processors, several caching schemes
![Page 4: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/4.jpg)
4
Computation of Effects
![Page 5: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/5.jpg)
5
Computation of Effects (Cont)
![Page 6: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/6.jpg)
6
Example 20.1: Code Size Comparison
Entries in a row are unrelated
(Otherwise, need a two factor analysis)
![Page 7: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/7.jpg)
7
Example 20.1 Code Size (cont’d)
![Page 8: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/8.jpg)
8
Example 20.1: Interpretation
Average processor requires 187.7 bytes of storage The effects of the processors R, V, and Z are
-13.3, -24.5, and 37.7, respectively. That is, R requires 13.3 bytes less than an average processor V requires 24.5 bytes less than an average processor,
and Z requires 37.7 bytes more than an average
processor.
![Page 9: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/9.jpg)
9
Estimating Experimental Errors
Estimated response for jth alternative:
Error:
Sum of squared errors (SSE):
![Page 10: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/10.jpg)
10
Example 20.2
![Page 11: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/11.jpg)
11
Allocation of Variation
![Page 12: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/12.jpg)
12
Allocation of Variation (cont’d)
Total variation of y (SST):
![Page 13: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/13.jpg)
13
Example 20.3
![Page 14: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/14.jpg)
14
Example 20.3 (cont’d)
89.6% of variation in code size is due to experimental errors (programmer differences)
Is 10.4% statistically significant?
![Page 15: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/15.jpg)
15
Analysis of Variance (ANOVA)
Importance Significance Important Explains a high percent of variation Significance
High contribution to the variation compared to that by errors
Degree of freedom = Number of independent values required to compute
Note that the degrees of freedom also add up.
![Page 16: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/16.jpg)
16
F-Test
Purpose: to check if SSA is significantly greater than SSE Errors are normally distributed
SSE and SSA have chi-square distributions The ratio (SSA/)/(SSE/e) has an F distribution
where A=a-1 = degrees of freedom for SSA
e=a(r-1) = degrees of freedom for SSE
Computed ratio > F[1- ; A, e]
SSA is significantly higher than SSE.
SSA/A is called mean square of A or (MSA)
Similary, MSE=SSE/e
![Page 17: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/17.jpg)
17
ANOVA Table for One Factor Experiments
![Page 18: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/18.jpg)
18
Example 20.4: Code Size Comparison
Computed F-value < F from Table The variation in the code sizes is mostly due to
experimental errors and not because of any significant difference among the processors
![Page 19: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/19.jpg)
19
Visual Diagnostic Tests
Assumptions:
1. Factors effects are additive
2. Errors are additive
3. Errors are independent of factor levels
4. Errors are normally distributed
5. Errors have the same variance for all factor levels
Tests: Residuals versus predicted response:
No trend Independence
Scale of errors << Scale of response
Ignore visible trends Normal quantilte-quantile plot linear Normality
![Page 20: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/20.jpg)
20
Example 20.5
Horizontal and vertical scales similar Residuals are not small Variation due to factors is small compared to the unexplained variation
No visible trend in the spread Q-Q plot is S-shaped shorter tails than normal
![Page 21: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/21.jpg)
21
Confidence Intervals For Effects
Estimates are random variables
For the confidence intervals, use t values at a(r-1) degrees of freedom.
Mean responses: Contrasts hj j: Use for 1-2
![Page 22: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/22.jpg)
22
Example 20.6: Code Size Comparison
![Page 23: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/23.jpg)
23
Example 20.6 (cont’d)
For 90% confidence, t[0.95; 12]= 1.782 90% confidence intervals:
The code size on an average processor is significantly different from zero
Processor effects are not significant
![Page 24: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/24.jpg)
24
Example 20.6 (cont’d)
Using h1=1, h2=-1, h3=0, ( hj=0):
CI includes zero one isn't superior to other
![Page 25: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/25.jpg)
25
Example 20.6 (cont’d)
Similarly,
Any one processor is not superior to another
![Page 26: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/26.jpg)
26
Unequal Sample Sizes
By definition:
Here, rj is the number of observations at jth level.
N =total number of observations:
![Page 27: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/27.jpg)
27
Parameter Estimation
![Page 28: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/28.jpg)
28
Analysis of Variance
![Page 29: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/29.jpg)
29
Example 20.7: Code Size Comparison
All means are obtained by dividing by the number of observations added
The column effects are 2.15, 13.75, and -21.92
![Page 30: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/30.jpg)
30
Example 20.6: Analysis of Variance
![Page 31: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/31.jpg)
31
Example 20.6 ANOVA (cont’d)
Sums of Squares:
Degrees of Freedom:
![Page 32: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/32.jpg)
32
Example 20.6 ANOVA Table
Conclusion: Variation due processors is insignificant as compared to that due to modeling errors
![Page 33: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/33.jpg)
33
Example 20.6 Standard Dev. of Effects
Consider the effect of processor Z: Since,
Error in 3 = Errors in terms on the right hand side:
eij's are normally distributed 3 is normal with
![Page 34: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/34.jpg)
34
Summary
Model for One factor experiments:
Computation of effects Allocation of variation, degrees of freedom ANOVA table Standard deviation of errors Confidence intervals for effects and contracts Model assumptions and visual tests
![Page 35: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/35.jpg)
35
Exercise 20.1
For a single factor design, suppose we want to write an expression for j in terms of yij's:
What are the values of a..j's? From the above expression, the error in j is seen to be:
Assuming errors eij are normally distributed with zero mean and variance e
2, write an expression for variance of ej
. Verify that your answer matches that
in Table 20.5.
![Page 36: CPE 619 One Factor Experiments](https://reader035.fdocuments.net/reader035/viewer/2022081501/56812c34550346895d90b93a/html5/thumbnails/36.jpg)
36
An Example
Analyze the following one factor experiment:
1. Compute the effects2. Prepare ANOVA table3. Compute confidence intervals for effects and interpret
4. Compute Confidence interval for 1-3
5. Show graphs for visual tests and interpret