Analisis Data Multivariat 1: Matriks
Hanifah Putri Hariel
2025-09-09
Matriks
Matriks adalah susunan bilangan atau objek matematika lain yang disusun dalam baris dan kolom membentuk suatu bangun persegi panjang. Matriks biasanya dituliskan dalam tanda kurung dan dinotasikan dengan huruf kapital. Elemen-elemennya disebut dengan unsur atau entri, dan diberi indeks baris dan kolom untuk menunjukkan posisinya.
Secara umum, matriks \(A\) dengan ordo \(m \times n\) (m baris dan n kolom) ditulis seperti berikut:
\[ A = \begin{bmatrix} a_{11} & a_{12} & \cdots & a_{1n} \\ a_{21} & a_{22} & \cdots & a_{2n} \\ \vdots & \vdots & \ddots & \vdots \\ a_{m1} & a_{m2} & \cdots & a_{mn} \end{bmatrix} \]
Di mana \(a_{ij}\) adalah elemen matriks pada baris ke-\(i\) dan kolom ke-\(j\).
Berikut contoh matriks dengan ukuran 2×2 dan 3×3:
- Matriks 2×2
\[ B = \begin{bmatrix} 1 & 2 \\ 3 & 4 \end{bmatrix} \]
- Matriks 3×3
\[ C = \begin{bmatrix} 5 & 6 & 7 \\ 8 & 9 & 10 \\ 11 & 12 & 13 \end{bmatrix} \]
Membuat Matriks
#Input Data
X = matrix(c(6.5,8.2,7.9,
5.4,7.0,6.7,
8.1,6.9,9.2), nrow = 3, ncol = 3); X## [,1] [,2] [,3]
## [1,] 6.5 5.4 8.1
## [2,] 8.2 7.0 6.9
## [3,] 7.9 6.7 9.2Y = matrix(c(7.3,6.8,8.5,
8.9,7.6,6.1,
9.4,8.0,7.2), nrow = 3, ncol = 3, byrow = TRUE); Y## [,1] [,2] [,3]
## [1,] 7.3 6.8 8.5
## [2,] 8.9 7.6 6.1
## [3,] 9.4 8.0 7.2#Cara Lainnya
A <- matrix(21:40, nrow=4, ncol=5) ; A## [,1] [,2] [,3] [,4] [,5]
## [1,] 21 25 29 33 37
## [2,] 22 26 30 34 38
## [3,] 23 27 31 35 39
## [4,] 24 28 32 36 40b <- c(4,7,2,8,5,9,6,3,1,10,12,11)
B <- matrix(b, nrow=3, ncol=4, byrow=TRUE) ; B## [,1] [,2] [,3] [,4]
## [1,] 4 7 2 8
## [2,] 5 9 6 3
## [3,] 1 10 12 11sel <- c(15,9,27,18)
nama_kolom <- c("C1", "C2")
nama_baris <- c("R1", "R2")
C <- matrix(sel, nrow=2, ncol=2,
byrow=TRUE, dimnames=list(nama_baris,
nama_kolom)) ; C## C1 C2
## R1 15 9
## R2 27 18Memanggil Komponen Matriks
A## [,1] [,2] [,3] [,4] [,5]
## [1,] 21 25 29 33 37
## [2,] 22 26 30 34 38
## [3,] 23 27 31 35 39
## [4,] 24 28 32 36 40A[,2] # Kolom 2## [1] 25 26 27 28A[3,] # Baris 3## [1] 23 27 31 35 39A[3,2] # Sel(3, 2)## [1] 27A[c(1,3),2] # Sel(1,2) dan sel(3,2)## [1] 25 27A[,1:3] # kolom(1,2,3)## [,1] [,2] [,3]
## [1,] 21 25 29
## [2,] 22 26 30
## [3,] 23 27 31
## [4,] 24 28 32A[2:4,] # baris(2,3,4)## [,1] [,2] [,3] [,4] [,5]
## [1,] 22 26 30 34 38
## [2,] 23 27 31 35 39
## [3,] 24 28 32 36 40Operasi Matriks
Operasi matriks adalah operasi dasar yang digunakan dalam analisis data multivariat untuk memanipulasi data secara struktural dan matematis. Operasi ini meliputi penjumlahan, pengurangan, perkalian, transpose, invers, dan determinan.
Rumus Dasar
- Penjumlahan: \(C = A + B\)
- Pengurangan: \(C = A - B\)
- Perkalian matriks: \(C = A \times
B\)
- Transpose: \(A^T\)
- Invers: \(A^{-1}\) jika \(\det(A) \neq 0\)
- Determinan: \(\det(A)\)
Contoh Implementasi
#Penjumlahan
X + Y## [,1] [,2] [,3]
## [1,] 13.8 12.2 16.6
## [2,] 17.1 14.6 13.0
## [3,] 17.3 14.7 16.4#Pengurangan
X - Y## [,1] [,2] [,3]
## [1,] -0.8 -1.4 -0.4
## [2,] -0.7 -0.6 0.8
## [3,] -1.5 -1.3 2.0Y - X## [,1] [,2] [,3]
## [1,] 0.8 1.4 0.4
## [2,] 0.7 0.6 -0.8
## [3,] 1.5 1.3 -2.0#Perkalian
X %*% Y## [,1] [,2] [,3]
## [1,] 171.65 150.04 146.51
## [2,] 187.02 164.16 162.08
## [3,] 203.78 178.24 174.26Y %*% X## [,1] [,2] [,3]
## [1,] 170.36 143.97 184.25
## [2,] 168.36 142.13 180.65
## [3,] 183.58 155.00 197.58#Perkalian antar elemen
X * Y## [,1] [,2] [,3]
## [1,] 47.45 36.72 68.85
## [2,] 72.98 53.20 42.09
## [3,] 74.26 53.60 66.24#Transpose
transX = t(X); transX## [,1] [,2] [,3]
## [1,] 6.5 8.2 7.9
## [2,] 5.4 7.0 6.7
## [3,] 8.1 6.9 9.2transY = t(Y); transY## [,1] [,2] [,3]
## [1,] 7.3 8.9 9.4
## [2,] 6.8 7.6 8.0
## [3,] 8.5 6.1 7.2#Invers
inv_X = solve(X); inv_X## [,1] [,2] [,3]
## [1,] 8.3848639 2.118136 -8.9709275
## [2,] -9.6585141 -1.933549 9.9538533
## [3,] -0.1661283 -0.410706 0.5629903inv_Y = solve(Y); inv_Y## [,1] [,2] [,3]
## [1,] -1.27147766 -4.089347 4.965636
## [2,] 1.44759450 5.871993 -6.683849
## [3,] 0.05154639 -1.185567 1.082474#Determinan
det(X)## [1] 2.167det(Y)## [1] -4.656Interpretasi
Operasi ini memungkinkan analisis dan manipulasi struktur data numerik dalam format matriks, yang penting untuk langkah-langkah analisis multivariat berikutnya.
Nilai Eigen dan Vektor Eigen
Eigenvalue dan eigenvector mengungkap informasi penting mengenai variasi dan struktur pada matriks data, berguna dalam reduksi dimensi dan analisis komponen utama.
Rumus Dasar
\(A v = \lambda v\)
Contoh Implementasi
eigX = eigen(X); eigX## eigen() decomposition
## $values
## [1] 22.0140019 0.4816027 0.2043953
##
## $vectors
## [,1] [,2] [,3]
## [1,] 0.5268942 0.5218595 0.61121520
## [2,] 0.5752916 -0.8360039 -0.78978922
## [3,] 0.6256373 0.1695881 0.05146821eigY = eigen(Y); eigY## eigen() decomposition
## $values
## [1] 23.230397 -1.286223 0.155826
##
## $vectors
## [,1] [,2] [,3]
## [1,] -0.5633011 -0.7710948 0.5522886
## [2,] -0.5584126 0.5270180 -0.8126545
## [3,] -0.6089887 0.3573023 0.1859299#Nilai Eigen
eigvalX = eigX$values; eigvalX## [1] 22.0140019 0.4816027 0.2043953eigvalY = eigY$values; eigvalY## [1] 23.230397 -1.286223 0.155826#Vektor Eigen
eigvecX = eigX$vectors; eigvecX## [,1] [,2] [,3]
## [1,] 0.5268942 0.5218595 0.61121520
## [2,] 0.5752916 -0.8360039 -0.78978922
## [3,] 0.6256373 0.1695881 0.05146821eigvecY = eigY$vectors; eigvecY## [,1] [,2] [,3]
## [1,] -0.5633011 -0.7710948 0.5522886
## [2,] -0.5584126 0.5270180 -0.8126545
## [3,] -0.6089887 0.3573023 0.1859299Interpretasi
Nilai eigen menandakan besaran variasi yang dijelaskan, sedangkan vektor eigen adalah arah data utama yang membantu mengeksplor struktur internal variabel.
Dekomposisi Singular Value (SVD)
SVD memecah matriks ke dalam tiga matriks dasar untuk mengkaji struktur data lebih lanjut, penting dalam pengurangan dimensi dan rekonstruksi data.
Rumus Dasar
\(A = U \Sigma V^T\)
Contoh Implementasi
library(MASS)
A = matrix(c(5,-3,6,2,-4,8,-2,5,-1,7,3,9), 4, 3, byrow=TRUE)
svd_result = svd(A)
#Singular Values
svd_result$d## [1] 16.07076 7.41936 3.11187#Matriks U
svd_result$u## [,1] [,2] [,3]
## [1,] -0.5046975 0.2278362 -0.3742460
## [2,] -0.5178195 0.4138180 0.7413297
## [3,] 0.1646416 -0.6063789 0.5337354
## [4,] -0.6708477 -0.6396483 -0.1596770#Matriks V
svd_result$v ## [,1] [,2] [,3]
## [1,] -0.5341591 -0.17494276 -0.8270847
## [2,] 0.1490928 -0.98251336 0.1115295
## [3,] -0.8321330 -0.06373793 0.5509011Interpretasi
Singular values mengukur energi atau informasi yang dikandung oleh masing-masing komponen, membantu dalam pemodelan dan analisis data kompleks.
Matriks Jarak
Matriks jarak digunakan untuk mengukur seberapa jauh atau dekat dua objek data dalam ruang multivariat, yang sangat berperan dalam teknik clustering dan segmentasi data.
Jenis Jarak dan Rumus
Jarak Euclidean
\(d = \sqrt{\sum_i (x_i - y_i)^2}\) Jarak lurus (garis terpendek) antara dua titik di ruang 𝑝-dimensi [jarak lurus standar].
Jarak Chebyshev
\(d = \max_i |x_i - y_i|\)
jarak ditentukan oleh selisih terbesar. Jarak maksimum di antara
perbedaan koordinat. Fokus pada dimensi dengan selisih terbesar.
Jarak Manhattan
\(d = \sum_i |x_i - y_i|\)
Jumlah perbedaan absolut antar koordinat, seperti berjalan di jalan kota
berbentuk grid [jarak berbasis grid (jumlah selisih)].
JarakMahalanobis
\(d = \sqrt{(x - y)^T S^{-1} (x -
y)}\)
Jarak antar titik yang mempertimbangkan skala (varians) dan korelasi
antar variabel.
Contoh Implementasi
library(factoextra)## Warning: package 'factoextra' was built under R version 4.5.1## Loading required package: ggplot2## Welcome! Want to learn more? See two factoextra-related books at https://goo.gl/ve3WBalibrary(StatMatch)## Warning: package 'StatMatch' was built under R version 4.5.1## Loading required package: proxy##
## Attaching package: 'proxy'## The following objects are masked from 'package:stats':
##
## as.dist, dist## The following object is masked from 'package:base':
##
## as.matrix## Loading required package: survey## Warning: package 'survey' was built under R version 4.5.1## Loading required package: grid## Loading required package: Matrix## Loading required package: survival##
## Attaching package: 'survey'## The following object is masked from 'package:graphics':
##
## dotchart## Loading required package: lpSolve## Loading required package: dplyr##
## Attaching package: 'dplyr'## The following object is masked from 'package:MASS':
##
## select## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionset.seed(321)
ss <- sample(1:50, 15)
df <- USArrests[ss, ]
df.scaled <- scale(df); df.scaled## Murder Assault UrbanPop Rape
## Wyoming -0.3721741 -0.02296746 -0.3418930 -0.62039386
## Illinois 0.4221896 1.02244775 1.2520675 0.62633064
## Mississippi 1.6799322 1.14124493 -1.4507350 -0.39776448
## Kansas -0.5486994 -0.56943449 0.0739228 -0.26418686
## New York 0.5766492 1.08184634 1.4599754 0.93801176
## Kentucky 0.2677300 -0.64071280 -0.8963140 -0.51650015
## Oklahoma -0.4163054 -0.14176464 0.2125281 0.03265231
## Hawaii -0.7031590 -1.38913505 1.2520675 0.06233622
## Missouri 0.1132704 0.17898775 0.3511333 1.24969289
## New Mexico 0.6428462 1.45011760 0.3511333 1.82852926
## Louisiana 1.5254725 1.02244775 0.0739228 0.35917539
## South Dakota -1.0341439 -0.91394632 -1.3814324 -1.03596869
## Iowa -1.3871944 -1.27033787 -0.5498008 -1.25859806
## North Dakota -1.6961136 -1.40101477 -1.4507350 -1.85227639
## Texas 0.9296998 0.45222127 1.0441596 0.84896001
## attr(,"scaled:center")
## Murder Assault UrbanPop Rape
## 8.486667 162.933333 64.933333 19.780000
## attr(,"scaled:scale")
## Murder Assault UrbanPop Rape
## 4.531929 84.177081 14.429467 6.737655#Jarak Euclidean
dist.eucl = dist(df.scaled, method = "euclidean"); dist.eucl## Wyoming Illinois Mississippi Kansas New York Kentucky
## Illinois 2.4122476
## Mississippi 2.6164146 3.1543527
## Kansas 0.7934567 2.3786048 3.1993198
## New York 2.7921742 0.4095812 3.3878156 2.7128511
## Kentucky 1.0532156 2.9515362 2.3433244 1.2948587 3.2757206
## Oklahoma 0.8659748 1.8685718 2.9986711 0.5547563 2.2043102 1.4993175
## Hawaii 2.2322175 2.7203365 4.4270510 1.4800030 2.9246694 2.5403456
## Missouri 2.0625111 1.4167282 3.0563398 1.8349434 1.5351057 2.3176129
## New Mexico 3.1109091 1.5775154 3.0617092 3.1551035 1.4705638 3.4011133
## Louisiana 2.4137967 1.6360410 1.7133330 2.6879097 1.7776353 2.4609320
## South Dakota 1.5765126 3.9457686 3.4644086 1.7515852 4.3067435 1.5082173
## Iowa 1.7426214 3.9154083 4.0958166 1.6038155 4.2724405 1.9508929
## North Dakota 2.5296038 4.8794481 4.4694938 2.6181473 5.2524274 2.5546862
## Texas 2.4496576 0.8218968 2.9692463 2.3259192 0.8377979 2.6949264
## Oklahoma Hawaii Missouri New Mexico Louisiana South Dakota
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii 1.6491638
## Missouri 1.3724911 2.3123720
## New Mexico 2.6268378 3.7154012 1.4937447
## Louisiana 2.2916633 3.5012381 1.8909275 1.7882330
## South Dakota 2.1588538 2.9115203 3.2767510 4.4281177 3.7902169
## Iowa 2.1130016 2.3395756 3.3845451 4.6758935 4.0922753 0.9964108
## North Dakota 3.0891779 3.4578871 4.3173165 5.5131433 4.8442635 1.1604313
## Texas 1.8768374 2.5920693 1.1756214 1.5867966 1.3643137 3.8935265
## Iowa North Dakota
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii
## Missouri
## New Mexico
## Louisiana
## South Dakota
## Iowa
## North Dakota 1.1298867
## Texas 3.9137858 4.8837032fviz_dist(dist.eucl)
#Jarak Chebyshev
dist.cheb = dist(df.scaled, method = "maximum"); dist.cheb## Wyoming Illinois Mississippi Kansas New York Kentucky
## Illinois 1.5939604
## Mississippi 2.0521063 2.7028025
## Kansas 0.5464670 1.5918822 2.2286315
## New York 1.8018683 0.3116811 2.9107104 1.6512808
## Kentucky 0.6399041 2.1483815 1.7819577 0.9702368 2.3562894
## Oklahoma 0.6530462 1.1642124 2.0962376 0.4276699 1.2474473 1.1088421
## Hawaii 1.5939604 2.4115828 2.7028025 1.1781447 2.4709814 2.1483815
## Missouri 1.8700867 0.9009342 1.8018683 1.5138797 1.1088421 1.7661930
## New Mexico 2.4489231 1.2021986 2.2262937 2.0927161 1.1088421 2.3450294
## Louisiana 1.8976467 1.1781447 1.5246578 2.0741719 1.3860526 1.6631605
## South Dakota 1.0395394 2.6334999 2.7140760 1.4553552 2.8414078 1.3018739
## Iowa 1.2473704 2.2927856 3.0671266 0.9944112 2.3521842 1.6549244
## North Dakota 1.3780473 2.7028025 3.3760458 1.5880895 2.9107104 1.9638436
## Texas 1.4693539 0.5702265 2.4948946 1.4783991 0.6296251 1.9404736
## Oklahoma Hawaii Missouri New Mexico Louisiana South Dakota
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii 1.2473704
## Missouri 1.2170406 1.5681228
## New Mexico 1.7958770 2.8392526 1.2711298
## Louisiana 1.9417780 2.4115828 1.4122022 1.4693539
## South Dakota 1.5939604 2.6334999 2.2856616 2.8644979 2.5596164
## Iowa 1.2912504 1.8018683 2.5082909 3.0871273 2.9126670 0.8316315
## North Dakota 1.8849287 2.7028025 3.1019693 3.6808057 3.2215862 0.8163077
## Texas 1.3460052 1.8413563 0.8164294 0.9978963 0.9702368 2.4255920
## Iowa North Dakota
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii
## Missouri
## New Mexico
## Louisiana
## South Dakota
## Iowa
## North Dakota 0.9009342
## Texas 2.3168942 2.7012364fviz_dist(dist.cheb)
#Jarak Manhattan
dist.man <- dist(df.scaled, method = "manhattan"); dist.man## Wyoming Illinois Mississippi Kansas New York Kentucky
## Illinois 4.6804639
## Mississippi 4.5477901 5.1034373
## Kansas 1.4950151 4.6314334 5.5975464
## New York 5.4139111 0.7334472 5.4091682 5.3648806
## Kentucky 1.9159642 5.1088324 3.8673166 2.1102578 5.8422796
## Oklahoma 1.3703957 3.6359252 5.4729270 0.9955082 4.3693724 2.8409781
## Hawaii 3.9738430 4.1009258 8.0763743 2.4788279 4.8343730 4.4465291
## Missouri 3.2505127 2.6766756 5.9782446 3.2014823 2.7867606 3.9878005
## New Mexico 5.6300548 2.7514592 5.3741207 5.5810243 2.4338278 6.0584233
## Louisiana 4.3384469 2.5485829 2.5548545 4.2894164 2.9731109 4.7668154
## South Dakota 3.0080629 7.6885267 5.4767741 3.0570933 8.4219740 2.5796943
## Iowa 3.1085028 7.7889667 7.2404771 3.1575333 8.5224139 3.3731605
## North Dakota 5.0427114 9.7231753 7.3728174 5.0917419 10.4566225 4.6143429
## Texas 4.6324690 1.5082739 5.1808752 4.5834386 1.4875431 5.0608376
## Oklahoma Hawaii Missouri New Mexico Louisiana
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii 2.6034473
## Missouri 2.2059740 4.4728430
## New Mexico 4.5855161 6.8523850 2.3795420
## Louisiana 3.5711187 6.1151982 3.4233902 3.0568606
## South Dakota 4.0526016 4.5379784 6.2585756 8.6381176 7.3465098
## Iowa 4.1530415 3.9256352 6.3590155 8.7385576 7.4469497
## North Dakota 6.0872501 5.6222495 8.2932241 10.6727662 9.3811583
## Texas 3.5879303 4.4687467 2.1834220 2.9573454 2.6260207
## South Dakota Iowa North Dakota
## Illinois
## Mississippi
## Kansas
## New York
## Kentucky
## Oklahoma
## Hawaii
## Missouri
## New Mexico
## Louisiana
## South Dakota
## Iowa 1.7637030
## North Dakota 2.0346485 1.9342086
## Texas 7.6405319 7.7409718 9.6751804fviz_dist(dist.man)
#Jarak Mahalanobis
dist.mah <- mahalanobis.dist(df.scaled); dist.mah## Wyoming Illinois Mississippi Kansas New York Kentucky
## Wyoming 0.000000 1.7186109 2.820779 1.4195095 1.8695558 2.867847
## Illinois 1.718611 0.0000000 3.658323 2.2905255 0.4722069 3.878642
## Mississippi 2.820779 3.6583235 0.000000 3.2139075 3.6566922 2.544477
## Kansas 1.419510 2.2905255 3.213907 0.0000000 2.1522535 2.048031
## New York 1.869556 0.4722069 3.656692 2.1522535 0.0000000 3.698342
## Kentucky 2.867847 3.8786421 2.544477 2.0480310 3.6983422 0.000000
## Oklahoma 1.146496 1.8980286 3.237573 0.6499978 1.7772007 2.505941
## Hawaii 3.466671 3.6449604 4.722203 2.2108491 3.3748818 2.753554
## Missouri 3.198071 3.6796400 3.956918 2.2592572 3.3618939 2.642756
## New Mexico 3.281318 3.5101406 4.057258 3.1016653 3.2869855 3.870023
## Louisiana 2.284940 2.5550539 1.688058 2.2700723 2.4136664 2.119635
## South Dakota 1.826205 3.3564158 3.087365 1.6274307 3.3404110 2.261154
## Iowa 1.327907 2.6329606 3.559587 1.1128197 2.6839965 2.621704
## North Dakota 1.582582 3.1919907 3.553572 1.9466491 3.3317039 3.040465
## Texas 2.540604 2.4769381 3.093919 1.7462066 2.1399545 2.108949
## Oklahoma Hawaii Missouri New Mexico Louisiana South Dakota
## Wyoming 1.1464956 3.466671 3.198071 3.281318 2.284940 1.826205
## Illinois 1.8980286 3.644960 3.679640 3.510141 2.555054 3.356416
## Mississippi 3.2375727 4.722203 3.956918 4.057258 1.688058 3.087365
## Kansas 0.6499978 2.210849 2.259257 3.101665 2.270072 1.627431
## New York 1.7772007 3.374882 3.361894 3.286985 2.413666 3.340411
## Kentucky 2.5059414 2.753554 2.642756 3.870023 2.119635 2.261154
## Oklahoma 0.0000000 2.705865 2.203038 2.660216 2.350208 1.672866
## Hawaii 2.7058650 0.000000 3.193764 4.645567 3.383255 3.551072
## Missouri 2.2030382 3.193764 0.000000 1.836797 3.256319 2.505784
## New Mexico 2.6602159 4.645567 1.836797 0.000000 3.676879 3.026024
## Louisiana 2.3502077 3.383255 3.256319 3.676879 0.000000 3.021642
## South Dakota 1.6728664 3.551072 2.505784 3.026024 3.021642 0.000000
## Iowa 1.3299426 2.790197 3.145245 3.792086 2.954252 1.518854
## North Dakota 1.9813596 3.780966 3.590548 3.950259 3.434074 1.304743
## Texas 1.9635201 2.082005 2.576037 3.501666 1.527269 3.090805
## Iowa North Dakota Texas
## Wyoming 1.327907 1.582582 2.540604
## Illinois 2.632961 3.191991 2.476938
## Mississippi 3.559587 3.553572 3.093919
## Kansas 1.112820 1.946649 1.746207
## New York 2.683996 3.331704 2.139954
## Kentucky 2.621704 3.040465 2.108949
## Oklahoma 1.329943 1.981360 1.963520
## Hawaii 2.790197 3.780966 2.082005
## Missouri 3.145245 3.590548 2.576037
## New Mexico 3.792086 3.950259 3.501666
## Louisiana 2.954252 3.434074 1.527269
## South Dakota 1.518854 1.304743 3.090805
## Iowa 0.000000 1.045923 2.734770
## North Dakota 1.045923 0.000000 3.563193
## Texas 2.734770 3.563193 0.000000dist.mah_matrix <- as.matrix(dist.mah)Interpretasi
Berbagai jenis jarak tersebut memungkinkan analisis yang sesuai dengan bentuk data dan tujuan, antara lain:
Jarak Euclidean - Menghitung jarak garis lurus antar dua koordinat (GPS) - Clustering (K-Means, Hierarchical) → objek yang jaraknya dekat digabungkan
Jarak Chebyshev - Jarak langkah raja antara dua posisi = jarak Chebyshev. - Berguna di quality control multivariat, misalnya mengecek dimensi produk (lebar, panjang, tinggi) → fokus pada dimensi terburuk.
Jarak Manhattan - Menghitung jarak dalam gudang/grid jalan yang tidak memungkinkan jalur diagonal. - Menghitung jarak antar dokumen berdasarkan frekuensi kata (NLP).
Jarak Mahalanobis - Mendeteksi transaksi keuangan yang tidak wajar - Memisahkan kelompok dengan varians dan korelasi berbeda (Analisis Diskriminan)
Vektor Rata-Rata dan Statistik Deskriptif Matriks
Statistik deskriptif matriks seperti vektor rata-rata, matriks kovarians, dan korelasi membantu memahami pusat dan hubungan antar variabel.
Rumus Dasar
- Vektor rata-rata: \(\bar{x}_i =
\frac{1}{n} \sum x_{ij}\)
- Matriks kovarians: \(Cov(X) =
\frac{1}{n-1} (X - \bar{X})^T (X - \bar{X})\)
- Matriks korelasi berdasarkan kovarians terstandarisasi
Contoh Implementasi
#Input Data
BB = c(6.2,11.5,8.7,10.1,7.8,6.9,12.0,3.1,14.8,9.4)
PM = c(61,73,68,70,64,60,76,49,84,71)
RTB = c(115,138,127,123,131,120,143,95,160,128)
lizard = as.matrix(cbind(BB,PM,RTB)); lizard## BB PM RTB
## [1,] 6.2 61 115
## [2,] 11.5 73 138
## [3,] 8.7 68 127
## [4,] 10.1 70 123
## [5,] 7.8 64 131
## [6,] 6.9 60 120
## [7,] 12.0 76 143
## [8,] 3.1 49 95
## [9,] 14.8 84 160
## [10,] 9.4 71 128#Matriks Rata-Rata
vecMeans = as.matrix(colMeans(lizard)); vecMeans## [,1]
## BB 9.05
## PM 67.60
## RTB 128.00vecRata = matrix(c(mean(BB), mean(PM), mean(RTB)), nrow=3, ncol=1); vecRata## [,1]
## [1,] 9.05
## [2,] 67.60
## [3,] 128.00#Matriks Kovarians
varkov = cov(lizard); varkov## BB PM RTB
## BB 10.98056 31.80000 54.96667
## PM 31.80000 94.04444 160.22222
## RTB 54.96667 160.22222 300.66667#Matriks Korelasi
korel = cor(lizard); korel## BB PM RTB
## BB 1.0000000 0.9895743 0.9566313
## PM 0.9895743 1.0000000 0.9528259
## RTB 0.9566313 0.9528259 1.0000000#Matriks Standardisasi (akar dari variansi masing-masing variabel)
n = nrow(lizard);n## [1] 10u = matrix(1,n,1); u## [,1]
## [1,] 1
## [2,] 1
## [3,] 1
## [4,] 1
## [5,] 1
## [6,] 1
## [7,] 1
## [8,] 1
## [9,] 1
## [10,] 1xbar = cbind((1/n)*t(u)%*%lizard); xbar## BB PM RTB
## [1,] 9.05 67.6 128D = lizard - u %*% xbar; D## BB PM RTB
## [1,] -2.85 -6.6 -13
## [2,] 2.45 5.4 10
## [3,] -0.35 0.4 -1
## [4,] 1.05 2.4 -5
## [5,] -1.25 -3.6 3
## [6,] -2.15 -7.6 -8
## [7,] 2.95 8.4 15
## [8,] -5.95 -18.6 -33
## [9,] 5.75 16.4 32
## [10,] 0.35 3.4 0S = (1/(n-1))*t(D)%*%D; S## BB PM RTB
## BB 10.98056 31.80000 54.96667
## PM 31.80000 94.04444 160.22222
## RTB 54.96667 160.22222 300.66667Ds = diag(sqrt(diag(S))); Ds## [,1] [,2] [,3]
## [1,] 3.313692 0.000000 0.00000
## [2,] 0.000000 9.697651 0.00000
## [3,] 0.000000 0.000000 17.33974R = solve(Ds) %*% S %*% solve(Ds); R## [,1] [,2] [,3]
## [1,] 1.0000000 0.9895743 0.9566313
## [2,] 0.9895743 1.0000000 0.9528259
## [3,] 0.9566313 0.9528259 1.0000000Interpretasi
Ini menunjukkan lokasi pusat (rata-rata) dan hubungan linear antar variabel (kovarians dan korelasi), dasar penting sebelum analisis lebih lanjut.