Réaliser une ACP en R
Algèbre linéaire pour la data science en R
Eric Eager
Data Scientist at Pro Football Focus
Données du NFL Combine
head(select(combine, height:shuttle))
height weight forty vertical bench broad_jump three_cone shuttle
1 71 192 4.38 35.0 14 127 6.71 3.98
2 73 298 5.34 26.5 27 99 7.81 4.71
3 77 256 4.67 31.0 17 113 7.34 4.38
4 74 198 4.34 41.0 16 131 6.56 4.03
5 76 257 4.87 30.0 20 118 7.12 4.23
6 78 262 4.60 38.5 18 128 7.53 4.48
Données du NFL Combine
prcomp(A)
Écarts types (1, .., p=8):
[1] 46.7720885 6.6356959 4.7108443 2.2950226 1.6430770 0.2513368 0.1216908 0.1143365
Rotation (n x k) = (8 x 8):
PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8
height 0.042047079 -0.061885367 0.1454490039 -0.1040556410 -0.980792060 0.020679696 -6.155636e-03 0.0008055445
weight 0.980711529 -0.130912788 0.1270100265 0.0193388930 0.066908382 -0.008423587 6.988341e-04 0.0036087841
forty 0.006112061 0.012525260 0.0025260713 -0.0021291637 0.004096693 0.152469227 -2.539868e-01 -0.9549983725
vertical -0.062926466 -0.333556369 0.0398922845 0.9366594549 -0.074901137 0.012214516 7.045063e-03 -0.0070051256
bench 0.088291423 -0.313533433 -0.9363461471 -0.0745692157 -0.107188391 0.009167322 -8.604309e-05 -0.0048308793
broad_jump -0.156742686 -0.876925849 0.2904565302 -0.3252903706 0.126494599 0.013753112 -2.187651e-03 -0.0076907609
three_cone 0.007468520 0.014691994 0.0009057581 0.0003320888 0.020902644 0.894560357 -3.743559e-01 0.2427137770
shuttle 0.004518826 0.009863931 0.0023111814 -0.0094052914 0.004010629 0.419039274 8.917710e-01 -0.1700673446
Données du NFL Combine
summary(prcomp(A))
Importance des composantes :
PC1 PC2 PC3 PC4 PC5 PC6 PC7 PC8
Écart type 46.7721 6.63570 4.71084 2.29502 1.64308 0.25134 0.12169 0.11434
Part de variance 0.9672 0.01947 0.00981 0.00233 0.00119 0.00003 0.00001 0.00001
Proportion cumulée 0.9672 0.98663 0.99644 0.99877 0.99996 0.99999 0.99999 1.00000
head(prcomp(A)$x[, 1:2])
PC1 PC2
[1,] -62.005067 -2.654645
[2,] 48.123290 6.693433
[3,] 3.732016 1.283046
[4,] -56.823742 -9.764098
[5,] 4.213670 -3.779862
[6,] 6.924978 -15.530509
head(cbind(combine[, 1:4], prcomp(A)$x[, 1:2]))
player position school year PC1 PC2
1 Jaire Alexander CB Louisville 2018 -62.005067 -2.654645
2 Brian Allen C Michigan St. 2018 48.123290 6.693433
3 Mark Andrews TE Oklahoma 2018 3.732016 1.283046
4 Troy Apke S Penn St. 2018 -56.823742 -9.764098
5 Dorance Armstrong EDGE Kansas 2018 4.213670 -3.779862
6 Ade Aruna DE Tulane 2018 6.924978 -15.530509
Que faire après une ACP
- Préparation des données / contrôle qualité
- Visualisation des données
- Apprentissage non supervisé (clustering)
- Apprentissage supervisé (prédiction ou explication)
- Et bien plus encore !
Exemple : visualisation des données
Passons à la pratique !
Algèbre linéaire pour la data science en R
