R in Action, Second Edition Data analysis and graphics with R

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

the .rar document contains the book, code, Bonus chapter 23, and Errata

R in Action Data analysis and graphics with R
Second Edition
by Robert I. Kabacoff

R in Action, Second Edition presents both the R language and the examples that make it so useful for business developers. Focusing on practical solutions, the book offers a crash course in statistics and covers elegant methods for dealing with messy and incomplete data that are difficult to analyze using traditional methods. You'll also master R's extensive graphical capabilities for exploring and presenting data visually. And this expanded second edition includes new chapters on time series analysis, cluster analysis, and classification methodologies, including decision trees, random forests, and support vector machines.

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏8 回帖

关键词：Graphics Analysis Edition Analysi GRAPHIC difficult practical business document examples

1. Listing 8.1. Simple linear regression
   
2. 
   
3. > fit <- lm(weight ~ height, data=women)
   
4. > summary(fit)
   
5. 
   
6. Call:
   
7. lm(formula=weight ~ height, data=women)
   
8. 
   
9. Residuals:
   
10.    Min     1Q Median     3Q    Max
    
11. -1.733 -1.133 -0.383  0.742  3.117
    
12. 
    
13. Coefficients:
    
14.             Estimate Std. Error t value Pr(>|t|)
    
15. (Intercept) -87.5167     5.9369   -14.7  1.7e-09 ***
    
16. height        3.4500     0.0911    37.9  1.1e-14 ***
    
17. ---
    
18. Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 '' 1
    
19. 
    
20. Residual standard error: 1.53 on 13 degrees of freedom
    
21. Multiple R-squared: 0.991,      Adjusted R-squared: 0.99
    
22. F-statistic: 1.43e+03 on 1 and 13 DF,  p-value: 1.09e-14
    
23. 
    
24. > women$weight
    
25. 
    
26. [1] 115 117 120 123 126 129 132 135 139 142 146 150 154 159 164
    
27. 
    
28. > fitted(fit)
    
29. 
    
30.      1      2      3      4      5      6      7      8      9
    
31. 112.58 116.03 119.48 122.93 126.38 129.83 133.28 136.73 140.18
    
32.     10     11     12     13     14     15
    
33. 143.63 147.08 150.53 153.98 157.43 160.88
    
34. 
    
35. > residuals(fit)
    
36. 
    
37.     1     2     3     4     5     6     7     8     9    10    11
    
38. 2.42  0.97  0.52  0.07 -0.38 -0.83 -1.28 -1.73 -1.18 -1.63 -1.08
    
39.    12    13    14    15
    
40. -0.53  0.02  1.57  3.12
    
41. 
    
42. > plot(women$height,women$weight,
    
43.        xlab="Height (in inches)",
    
44.        ylab="Weight (in pounds)")
    
45. > abline(fit)

复制代码

1. Listing 9.2. Tukey HSD pairwise group comparisons
   
2. 
   
3. > TukeyHSD(fit)
   
4.   Tukey multiple comparisons of means
   
5.     95% family-wise confidence level
   
6. 
   
7. Fit: aov(formula = response ~ trt)
   
8. 
   
9. $trt
   
10.                diff    lwr   upr p adj
    
11. 2times-1time   3.44 -0.658  7.54 0.138
    
12. 4times-1time   6.59  2.492 10.69 0.000
    
13. drugD-1time    9.58  5.478 13.68 0.000
    
14. drugE-1time   15.17 11.064 19.27 0.000
    
15. 4times-2times  3.15 -0.951  7.25 0.205
    
16. drugD-2times   6.14  2.035 10.24 0.001
    
17. drugE-2times  11.72  7.621 15.82 0.000
    
18. drugD-4times   2.99 -1.115  7.09 0.251
    
19. drugE-4times   8.57  4.471 12.67 0.000
    
20. drugE-drugD    5.59  1.485  9.69 0.003
    
21. 
    
22. > par(las=2)
    
23. > par(mar=c(5,8,4,2))
    
24. > plot(TukeyHSD(fit))

复制代码

1. Listing 9.3. One-way ANCOVA
   
2. 
   
3. > data(litter, package="multcomp")
   
4. > attach(litter)
   
5. > table(dose)
   
6. dose
   
7.   0   5  50 500
   
8. 20  19  18  17
   
9. > aggregate(weight, by=list(dose), FUN=mean)
   
10.   Group.1    x
    
11. 1       0 32.3
    
12. 2       5 29.3
    
13. 3      50 29.9
    
14. 4     500 29.6
    
15. > fit <- aov(weight ~ gesttime + dose)
    
16. > summary(fit)
    
17.             Df  Sum Sq Mean Sq F value   Pr(>F)
    
18. gesttime     1  134.30  134.30  8.0493 0.005971 **
    
19. dose         3  137.12   45.71  2.7394 0.049883 *
    
20. Residuals   69 1151.27   16.69
    
21. ---
    
22. Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

复制代码

1. Listing 9.4. Multiple comparisons employing user-supplied contrasts
   
2. 
   
3. > library(multcomp)
   
4. > contrast <- rbind("no drug vs. drug" = c(3, -1, -1, -1))
   
5. > summary(glht(fit, linfct=mcp(dose=contrast)))
   
6. 
   
7. Multiple Comparisons of Means: User-defined Contrasts
   
8. 
   
9. Fit: aov(formula = weight ~ gesttime + dose)
   
10. 
    
11. Linear Hypotheses:
    
12.                       Estimate Std. Error t value Pr(>|t|)
    
13. no drug vs. drug == 0    8.284      3.209   2.581   0.0120 *
    
14. ---
    
15. Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
    
16. The contrast c(3, -1, -1, -1) specifies a comparison of the first group with the average of the other three. The hypothesis is tested with a t statistic (2.581 in this case), which is significant at the p < .05 level. Therefore, you can conclude that the no-drug group has a higher birth weight than drug conditions. Other contrasts can be added to the rbind() function (see help(glht) for details).

复制代码

1. Listing 10.1. Sample sizes for detecting significant effects in a one-way ANOVA
   
2. 
   
3. library(pwr)
   
4. es <- seq(.1, .5, .01)
   
5. nes <- length(es)
   
6. 
   
7. samsize <- NULL
   
8. for (i in 1:nes){
   
9.     result <- pwr.anova.test(k=5, f=es[i], sig.level=.05, power=.9)
   
10.     samsize[i] <- ceiling(result$n)
    
11. }
    
12. 
    
13. plot(samsize,es, type="l", lwd=2, col="red",
    
14.      ylab="Effect Size",
    
15.      xlab="Sample Size (per cell)",
    
16.      main="One Way ANOVA with Power=.90 and Alpha=.05")
    
17. Graphs such as these can help you estimate the impact of various conditions on your experimental design. For example, there appears to be little bang for the buck in increasing the sample size above 200 observations per group. We’ll look at another plotting example in the next section.

复制代码

SleepyTom 发表于 2016-5-8 00:28
the .rar document contains the book, code, Bonus chapter 23, and Errata

R in Action Data analysis ...

谢谢分享

thanks ...