Perbandingan Menggunakan Bootstrap dan Jackknife

Berikut merupakan hasil output R

Memanggil data
> data=read.csv("D://TUGAS/world95.csv", sep=",")

> data

Country Urban life.expf life.expm literacy pop.increase babymort

1 Afghanistan 18 44 45 29 2.8 168.0

2 Argentina 86 75 68 95 1.3 25.6

3 Armenia 68 75 68 98 1.4 27.0

4 Australia 85 80 74 100 1.4 7.3

5 Austria 58 79 73 99 0.2 6.7

6 Azerbaijan 54 75 67 98 1.4 35.0

7 Bahrain 83 74 71 77 2.4 25.0

8 Bangladesh 16 53 53 35 2.4 106.0

9 Barbados 45 78 73 99 0.2 20.3

10 Belarus 65 76 66 99 0.3 19.0

11 Belgium 96 79 73 99 0.2 7.2

12 Bolivia 51 64 59 78 2.7 75.0

13 Bosnia 36 78 72 86 0.7 12.7

14 Botswana 25 66 60 72 2.7 39.3

15 Brazil 75 67 57 81 1.3 66.0

16 Bulgaria 68 75 69 93 -0.2 12.0

17 Burkina Faso 15 50 47 18 2.8 118.0

18 Burundi 5 50 46 50 2.3 105.0

19 Cambodia 12 52 50 35 2.9 112.0

20 Cameroon 40 58 55 54 2.9 77.0

21 Canada 77 81 74 97 0.7 6.8

22 Cent. Afri.R 47 44 41 27 2.4 137.0

23 Chile 85 78 71 93 1.7 14.6

24 China 26 69 67 78 1.1 52.0

25 Colombia 70 75 69 87 2.0 28.0

26 Costa Rica 47 79 76 93 2.3 11.0

27 Croatia 51 77 70 97 -0.1 8.7

28 Cuba 74 78 74 94 1.0 10.2

29 Czech Rep. 50 77 69 90 0.2 9.3

30 Denmark 85 79 73 99 0.1 6.6

31 Domincan R. 60 70 66 83 1.8 51.5

32 Ecuador 56 73 67 88 2.0 39.0

33 Egypt 44 63 60 48 2.0 76.4

34 El Salvador 44 69 64 73 2.0 41.0

35 Estonia 72 76 67 99 0.5 19.0

36 Ethiopia 12 54 51 24 3.1 110.0

37 Finland 60 80 72 100 0.3 5.3

38 France 73 82 74 99 0.5 6.7

39 Gabon 46 58 52 61 1.5 94.0

40 Gambia 23 52 48 27 3.1 124.0

41 Georgia 56 76 69 99 0.8 23.0

42 Germany 85 79 73 99 0.4 6.5

43 Greece 63 80 75 93 0.8 8.2

44 Guatemala 39 67 62 55 2.6 57.0

45 Haiti 29 47 43 53 1.6 109.0

46 Honduras 44 70 65 73 2.7 45.0

47 Hong Kong 94 80 75 77 -0.1 5.8

48 Hungary 64 76 67 99 -0.3 12.5

49 Iceland 91 81 76 100 1.1 4.0

50 India 26 59 58 52 1.9 79.0

51 Indonesia 29 65 61 77 1.6 68.0

52 Iran 57 67 65 54 3.5 60.0

53 Iraq 72 68 65 60 3.7 67.0

54 Ireland 57 78 73 98 0.3 7.4

55 Israel 92 80 76 92 2.2 8.6

56 Italy 69 81 74 97 0.2 7.6

57 Japan 77 82 76 99 0.3 4.4

58 Jordan 68 74 70 80 3.3 34.0

59 Kenya 24 55 51 69 3.1 74.0

60 Kuwait 96 78 73 73 5.2 12.5

61 Latvia 71 75 64 99 0.5 21.5

62 Lebanon 84 71 67 80 2.0 39.5

63 Liberia 45 57 54 40 3.3 113.0

64 Libya 82 65 62 64 3.7 63.0

65 Lithuania 69 77 68 99 0.3 17.0

66 Malaysia 43 72 66 78 2.3 25.6

67 Mexico 73 77 69 87 1.9 35.0

68 Morocco 46 70 66 50 2.1 50.0

69 N. Korea 60 73 67 99 1.8 27.7

70 Netherlands 89 81 75 99 0.6 6.3

71 New Zealand 84 80 73 99 0.6 8.9

72 Nicaragua 60 67 61 57 2.7 52.5

73 Nigeria 35 57 54 51 3.1 75.0

74 Norway 75 81 74 99 0.4 6.3

75 Oman 11 70 66 NA 3.5 36.7

76 Pakistan 32 58 57 35 2.8 101.0

77 Panama 53 78 71 88 1.9 16.5

78 Paraguay 48 75 72 90 2.7 25.2

79 Peru 70 67 63 85 2.0 54.0

80 Philippines 43 68 63 90 1.9 51.0

81 Poland 62 77 69 99 0.3 13.8

82 Portugal 34 78 71 85 0.4 9.2

83 Romania 54 75 69 96 0.1 20.3

84 Russia 74 74 64 99 0.2 27.0

85 Rwanda 6 46 43 50 2.8 117.0

86 S. Korea 72 74 68 96 1.0 21.7

87 Saudi Arabia 77 70 66 62 3.2 52.0

88 Senegal 40 58 55 38 3.1 76.0

89 Singapore 100 79 73 88 1.2 5.7

90 Somalia 24 55 54 24 3.2 126.0

91 South Africa 49 68 62 76 2.6 47.1

92 Spain 78 81 74 95 0.3 6.9

93 Sweden 84 81 75 99 0.5 5.7

94 Switzerland 62 82 75 99 0.7 6.2

95 Syria 50 68 65 64 3.7 43.0

96 Taiwan 71 78 72 91 0.9 5.1

97 Tanzania 21 45 41 46 2.5 110.0

98 Thailand 22 72 65 93 1.4 37.0

99 Turkey 61 73 69 81 2.0 49.0

100 U.Arab Em. 81 74 70 68 4.8 22.0

101 UK 89 80 74 99 0.2 7.2

102 USA 75 79 73 97 1.0 8.1

103 Uganda 11 43 41 48 2.4 112.0

104 Ukraine 67 75 65 97 0.1 20.7

105 Uruguay 89 77 71 96 0.8 17.0

106 Uzbekistan 41 72 65 97 2.1 53.0

107 Venezuela 91 76 70 88 2.2 28.0

108 Vietnam 20 68 63 88 1.8 46.0

109 Zambia 42 45 44 73 2.8 85.0

gdp.cap birth.rt death.rt fertility category

1 205 53.0 22 6.9 C

2 3408 20.0 9 2.8 C

3 5000 23.0 6 3.2 B

4 16848 15.0 8 1.9 A

5 18396 12.0 11 1.5 A

6 3000 23.0 7 2.8 C

7 7875 29.0 4 4.0 B

8 202 35.0 11 4.7 C

9 6950 16.0 8 1.8 B

10 6500 13.0 11 1.9 B

11 17912 12.0 11 1.7 A

12 730 34.0 9 4.2 C

13 3098 14.0 6 3.0 C

14 2677 32.0 8 5.1 C

15 2354 21.0 9 2.7 C

16 3831 13.0 12 1.8 C

17 357 47.0 18 6.9 C

18 208 44.0 21 6.8 C

19 260 45.0 16 5.8 C

20 993 41.0 12 5.7 C

21 19904 14.0 8 1.8 A

22 457 44.0 21 5.4 C

23 2591 23.0 6 2.5 C

24 377 21.0 7 1.8 C

25 1538 24.0 6 2.5 C

26 2031 26.0 4 3.1 C

27 5487 11.0 11 1.7 B

28 1382 17.0 7 1.9 C

29 7311 13.0 11 1.8 B

30 18277 12.0 12 1.7 A

31 1034 25.0 6 2.8 C

32 1085 26.0 6 3.1 C

33 748 29.0 9 3.8 C

34 1078 33.0 7 3.8 C

35 6000 14.0 12 2.0 B

36 122 45.0 14 6.8 C

37 15877 13.0 10 1.8 A

38 18944 13.0 9 1.8 A

39 4283 28.0 14 4.0 C

40 351 46.0 16 6.3 C

41 4500 16.0 9 2.2 C

42 17539 11.0 11 1.5 A

43 8060 10.0 10 1.5 B

44 1342 35.0 8 4.8 C

45 383 40.0 19 5.9 C

46 1030 35.0 6 4.9 C

47 14641 13.0 6 1.4 B

48 5249 12.0 13 1.8 B

49 17241 16.0 7 2.1 A

50 275 29.0 10 4.5 C

51 681 24.0 9 2.8 C

52 1500 42.0 8 6.3 C

53 1955 44.0 7 6.7 C

54 12170 14.0 9 2.0 B

55 13066 21.0 7 2.8 B

56 17500 11.0 10 1.3 A

57 19860 11.0 7 1.6 A

58 1157 39.0 5 5.6 C

59 323 42.0 11 5.9 C

60 6818 28.0 2 4.0 B

61 7400 14.0 12 2.0 B

62 1429 27.0 7 3.4 C

63 409 43.0 12 6.8 C

64 5910 45.0 8 6.4 B

65 6710 15.0 10 2.0 B

66 2995 29.0 5 3.5 C

67 3604 28.0 5 3.2 C

68 1062 29.0 6 3.8 C

69 1000 24.0 6 2.4 C

70 17245 13.0 9 1.6 A

71 14381 16.0 8 2.0 B

72 447 35.0 7 4.3 C

73 282 44.0 12 6.4 C

74 17755 13.0 10 2.0 A

75 7467 40.0 5 6.5 B

76 406 42.0 10 6.4 C

77 2397 25.0 5 2.9 C

78 1500 33.0 5 4.3 C

79 1107 26.0 7 3.1 C

80 867 27.0 7 3.4 C

81 4429 14.0 10 1.9 C

82 9000 12.0 10 1.5 B

83 2702 14.0 10 1.8 C

84 6680 13.0 11 1.8 B

85 292 49.0 21 8.2 C

86 6627 16.0 6 1.7 B

87 6651 38.0 6 6.7 B

88 744 43.0 12 6.1 C

89 14990 16.0 6 1.9 B

90 2126 46.0 13 7.3 C

91 3128 34.0 8 4.4 C

92 13047 11.0 9 1.4 B

93 16900 14.0 11 2.1 A

94 22384 12.0 9 1.6 A

95 2436 44.0 6 6.7 C

96 7055 15.6 6 6.0 B

97 263 46.0 19 6.2 C

98 1800 19.0 6 2.1 C

99 3721 26.0 6 3.2 C

100 14193 28.0 3 4.5 B

101 15974 13.0 11 1.8 A

102 23474 15.0 9 2.1 A

103 325 49.0 24 6.8 C

104 2340 12.0 13 1.8 C

105 3131 17.0 10 2.4 C

106 1350 30.0 7 3.7 C

107 2829 26.0 5 3.1 C

108 230 27.0 8 3.3 C

109 573 46.0 18 6.7 C

Mengambil variabel data.gdp.cap dengan data frame

> data1=data.frame(data$gdp.cap)

> data1

data.gdp.cap

1 205

2 3408

3 5000

4 16848

5 18396

6 3000

7 7875

8 202

9 6950

10 6500

11 17912

12 730

13 3098

14 2677

15 2354

16 3831

17 357

18 208

19 260
20 993

21 19904

22 457

23 2591

24 377

25 1538

26 2031

27 5487

28 1382

29 7311

30 18277

31 1034

32 1085

33 748

34 1078

35 6000

36 122

37 15877

38 18944

39 4283

40 351

41 4500

42 17539

43 8060

44 1342

45 383

46 1030

47 14641

48 5249

49 17241

50 275
51 681

52 1500

53 1955

54 12170

55 13066

56 17500

57 19860

58 1157

59 323

60 6818

61 7400

62 1429

63 409

64 5910

65 6710

66 2995

67 3604

68 1062

69 1000

70 17245

71 14381

72 447

73 282

74 17755

75 7467

76 406

77 2397

78 1500

79 1107

80 867

81 4429
82 9000

83 2702

84 6680

85 292

86 6627

87 6651

88 744

89 14990

90 2126

91 3128

92 13047

93 16900

94 22384

95 2436

96 7055

97 263

98 1800

99 3721

100 14193

101 15974

102 23474

103 325

104 2340

105 3131

106 1350

107 2829

108 230

109 573

Menghitung mean Jackknife dan biasnya dengan fungsi Jackknife

> source("D:\\TUGAS\\function_jackknife.txt")

> Jackknife(data1$data.gdp.cap)
$meanjackknife

[1] 5859.982

$bias

[1] 0

Menghitung mean Bootstrap dan biasnya dengan fungsi bootstrap

> source("D:\\TUGAS\\bootsrap.txt")

Jumlah Replikasi 1000

> bootstrap(data1$data.gdp.cap,5,1000)

$mean_bootstrap

[1] 5829.8

$se_bootstrap

[1] 2902.548

$bias

[1] 30.18145

Jumlah Replikasi 109 sesuai banyak data

> bootstrap(data1$data.gdp.cap,5,109)

$mean_bootstrap

[1] 6260.903

$se_bootstrap

[1] 2813.693

$bias

[1] -400.9211

Kesimpulan :
Dengan menggunakan Jackknife bias lebih kecil daripada menggunakan bootstrap dengan
nilai bias 0 sedangkan menggunakan bootstrap dengan replikasi 1000 nilai biasnya 30.18145
Sehingga menggunakan Jacknife lebih baik daripada menggunakan Bootstrap. Selain itu juga
dapat dibandingkan apabila menggunakan bootstrap dengan replikasi yang besar misal 1000
dengan replikasi 109 sebanyak jumlah data nilai biasnya lebih kecil dengan replikasi 1000,
maka replikasi lebih banyak pada bootstrap maka lebih baik dan bias dari bootstrap
tergantung dari jumlah data yang di resample dan replikasinya. Mean dengan menggunakan
jackknife dan menggunakan bootstrap dengan replikasi besar hampir sama yaitu 5859.982
dengan menggunakan jackknife dan 5829.8 dengan menggunakan bootstrap

Fungsi Jackknife
Jackknife=function(data1)
{
n=length(data1)
meansample=matrix(nrow=n,ncol=1)

for(i in 1:n)
{
sample=data1[-i]
meansample[i]=mean(sample)
}
meanjackknife=mean(meansample)
bias=mean(data1)-meanjackknife

list(meanjackknife=meanjackknife, bias=bias)
}
Fungsi Bootstrap
bootstrap=function(data,m,R)
{
n=length(data)
meansample=matrix(ncol=1,nrow=R)

for (i in 1:R)
{
sample=sample(data,m)
meansample[i]=mean(sample)
}
mean_bootstrap=mean(meansample)
hist(meansample)
se_bootstrap=sd(meansample)
bias=mean(data)-mean_bootstrap

list(mean_bootstrap=mean_bootstrap, se_bootstrap=se_bootstrap, bias=bias)

}