Clustering, also known as cluster analysis, is an unsupervised learning technique used to identify patterns and structures in data sets. The main goal of clustering is to group similar objects into the same clusters and dissimilar objects into distinct clusters based on some measure of similarity or dissimilarity between them. Clustering has various applications such as customer segmentation, anomaly detection, general data exploration, and more. As a result, cluster analysis has applications in different fields.
The dataset used in this project pertains to customers of a Portuguese wholesale distributor. It includes annual spending in monetary units (u.m) on various product categories.
The observations refer to customers, and the variables are divided as follows:
FRESH= Annual spending (in monetary units) on fresh products;
MILK= Annual spending (in monetary units) on dairy products;;
GROCERY= Annual spending (in monetary units) on grocery products;
FROZEN= Annual spending (in monetary units) on frozen products;
DETERGENTS_PAPER= Annual spending (in monetary units) on detergents and paper products;
DELICATESSEN= Annual spending (in monetary units) on delicatessen products;
CHANNEL= Customer channel - Horeca (Hotel/Restaurant/Cafe) or Retail channel;
REGION= Customer region - Lisbon, Porto, or Other city.
“CHANNEL” and “REGION” are categorical variables, while the rest are quantitative variables.
In this project, we will perform a hierarchical clustering process and a “k-means” clustering process. In summary, we will conduct a cluster analysis in which the number of clusters will be determined during the process (hierarchical method), and another analysis in which the number of clusters will be predefined. This way, we can use a common practice among data scientists, which involves using the output of the hierarchical method as input for the “k-means” method.”
Database used:
Wholesale
Customers Data (Please right-click and select “open in a new
tab/window.” )
Installation and loading of the used packages
pacotes <- c("plotly", "fastDummies", "tidyverse", "ggrepel", "knitr", "kableExtra", "reshape2",
"misc3d", "plot3D", "cluster", "factoextra", "ade4")
if(sum(as.numeric(!pacotes %in% installed.packages())) != 0){
instalador <- pacotes[!pacotes %in% installed.packages()]
for(i in 1:length(instalador)) {
install.packages(instalador, dependencies = T)
break()}
sapply(pacotes, require, character = T)
} else {
sapply(pacotes, require, character = T)
}
clientesdata <- read.csv("Wholesale customers data.csv")
save(clientesdata, file = "clientesdata.RData")
View(clientesdata)
| Channel | Region | Fresh | Milk | Grocery | Frozen | Detergents_Paper | Delicassen |
|---|---|---|---|---|---|---|---|
| 2 | 3 | 12669 | 9656 | 7561 | 214 | 2674 | 1338 |
| 2 | 3 | 7057 | 9810 | 9568 | 1762 | 3293 | 1776 |
| 2 | 3 | 6353 | 8808 | 7684 | 2405 | 3516 | 7844 |
| 1 | 3 | 13265 | 1196 | 4221 | 6404 | 507 | 1788 |
| 2 | 3 | 22615 | 5410 | 7198 | 3915 | 1777 | 5185 |
| 2 | 3 | 9413 | 8259 | 5126 | 666 | 1795 | 1451 |
| 2 | 3 | 12126 | 3199 | 6975 | 480 | 3140 | 545 |
| 2 | 3 | 7579 | 4956 | 9426 | 1669 | 3321 | 2566 |
| 1 | 3 | 5963 | 3648 | 6192 | 425 | 1716 | 750 |
| 2 | 3 | 6006 | 11093 | 18881 | 1159 | 7425 | 2098 |
showing the first rows only
map(clientesdata[, c("Channel", "Region")], ~ summary(as.factor(.)))
## $Channel
## 1 2
## 298 142
##
## $Region
## 1 2 3
## 77 47 316
Where: Channel(1) = Hotel/Restaurant/Café; Channel(2) = Retail. Region(1) = Lisbon; Region(2) = Porto; Region(3) = Other Region.
glimpse(clientesdata)
## Rows: 440
## Columns: 8
## $ Channel <int> 2, 2, 2, 1, 2, 2, 2, 2, 1, 2, 2, 2, 2, 2, 2, 1, 2, 1,…
## $ Region <int> 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3,…
## $ Fresh <int> 12669, 7057, 6353, 13265, 22615, 9413, 12126, 7579, 5…
## $ Milk <int> 9656, 9810, 8808, 1196, 5410, 8259, 3199, 4956, 3648,…
## $ Grocery <int> 7561, 9568, 7684, 4221, 7198, 5126, 6975, 9426, 6192,…
## $ Frozen <int> 214, 1762, 2405, 6404, 3915, 666, 480, 1669, 425, 115…
## $ Detergents_Paper <int> 2674, 3293, 3516, 507, 1777, 1795, 3140, 3321, 1716, …
## $ Delicassen <int> 1338, 1776, 7844, 1788, 5185, 1451, 545, 2566, 750, 2…
Since the categorical variables are encoded as numerical values, we will change them to factors:
clientesdata2 <- clientesdata
clientesdata2$Channel <- as.factor(clientesdata$Channel)
clientesdata2$Region <- as.factor(clientesdata$Region)
Since we have both categorical and numerical variables in the database, we will separate the variables into two databases so that we can create two distance matrices. This procedure is necessary because we will use different distance calculation methods for numerical and categorical variables. Afterward, we will combine the matrices and perform clustering on the combined matrix.
dados_numericos <- clientesdata2[, c("Fresh", "Milk", "Grocery", "Frozen", "Detergents_Paper", "Delicassen")]
dados_categoricos <- clientesdata2[, c("Channel", "Region")]
dados_padronizados <- as.data.frame(scale(dados_numericos))
Now, all numerical variables have a mean of 0 and a standard deviation of 1. Standardization is necessary in cluster analyses when the data does not have a balanced scale of values among variables.
dados_dummies <- dummy_columns(.data = dados_categoricos,
select_columns = "Channel",
remove_selected_columns = T,
remove_most_frequent_dummy = T)
dados_dummies <- dummy_columns(.data = dados_dummies,
select_columns = "Region",
remove_selected_columns = T,
remove_most_frequent_dummy = T)
matriz_D_numerica <- dados_padronizados %>% dist(method = "euclidean")
matriz_D_categorica <- dados_dummies %>% dist(method = "binary")
dist_total <- matriz_D_categorica + matriz_D_numerica
data.matrix(dist_total)[1:5, ] %>%
kable() %>%
kable_styling(bootstrap_options = c("striped", "hover"), full_width = FALSE) %>%
scroll_box(width = "100%", height = "250px")
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 | 27 | 28 | 29 | 30 | 31 | 32 | 33 | 34 | 35 | 36 | 37 | 38 | 39 | 40 | 41 | 42 | 43 | 44 | 45 | 46 | 47 | 48 | 49 | 50 | 51 | 52 | 53 | 54 | 55 | 56 | 57 | 58 | 59 | 60 | 61 | 62 | 63 | 64 | 65 | 66 | 67 | 68 | 69 | 70 | 71 | 72 | 73 | 74 | 75 | 76 | 77 | 78 | 79 | 80 | 81 | 82 | 83 | 84 | 85 | 86 | 87 | 88 | 89 | 90 | 91 | 92 | 93 | 94 | 95 | 96 | 97 | 98 | 99 | 100 | 101 | 102 | 103 | 104 | 105 | 106 | 107 | 108 | 109 | 110 | 111 | 112 | 113 | 114 | 115 | 116 | 117 | 118 | 119 | 120 | 121 | 122 | 123 | 124 | 125 | 126 | 127 | 128 | 129 | 130 | 131 | 132 | 133 | 134 | 135 | 136 | 137 | 138 | 139 | 140 | 141 | 142 | 143 | 144 | 145 | 146 | 147 | 148 | 149 | 150 | 151 | 152 | 153 | 154 | 155 | 156 | 157 | 158 | 159 | 160 | 161 | 162 | 163 | 164 | 165 | 166 | 167 | 168 | 169 | 170 | 171 | 172 | 173 | 174 | 175 | 176 | 177 | 178 | 179 | 180 | 181 | 182 | 183 | 184 | 185 | 186 | 187 | 188 | 189 | 190 | 191 | 192 | 193 | 194 | 195 | 196 | 197 | 198 | 199 | 200 | 201 | 202 | 203 | 204 | 205 | 206 | 207 | 208 | 209 | 210 | 211 | 212 | 213 | 214 | 215 | 216 | 217 | 218 | 219 | 220 | 221 | 222 | 223 | 224 | 225 | 226 | 227 | 228 | 229 | 230 | 231 | 232 | 233 | 234 | 235 | 236 | 237 | 238 | 239 | 240 | 241 | 242 | 243 | 244 | 245 | 246 | 247 | 248 | 249 | 250 | 251 | 252 | 253 | 254 | 255 | 256 | 257 | 258 | 259 | 260 | 261 | 262 | 263 | 264 | 265 | 266 | 267 | 268 | 269 | 270 | 271 | 272 | 273 | 274 | 275 | 276 | 277 | 278 | 279 | 280 | 281 | 282 | 283 | 284 | 285 | 286 | 287 | 288 | 289 | 290 | 291 | 292 | 293 | 294 | 295 | 296 | 297 | 298 | 299 | 300 | 301 | 302 | 303 | 304 | 305 | 306 | 307 | 308 | 309 | 310 | 311 | 312 | 313 | 314 | 315 | 316 | 317 | 318 | 319 | 320 | 321 | 322 | 323 | 324 | 325 | 326 | 327 | 328 | 329 | 330 | 331 | 332 | 333 | 334 | 335 | 336 | 337 | 338 | 339 | 340 | 341 | 342 | 343 | 344 | 345 | 346 | 347 | 348 | 349 | 350 | 351 | 352 | 353 | 354 | 355 | 356 | 357 | 358 | 359 | 360 | 361 | 362 | 363 | 364 | 365 | 366 | 367 | 368 | 369 | 370 | 371 | 372 | 373 | 374 | 375 | 376 | 377 | 378 | 379 | 380 | 381 | 382 | 383 | 384 | 385 | 386 | 387 | 388 | 389 | 390 | 391 | 392 | 393 | 394 | 395 | 396 | 397 | 398 | 399 | 400 | 401 | 402 | 403 | 404 | 405 | 406 | 407 | 408 | 409 | 410 | 411 | 412 | 413 | 414 | 415 | 416 | 417 | 418 | 419 | 420 | 421 | 422 | 423 | 424 | 425 | 426 | 427 | 428 | 429 | 430 | 431 | 432 | 433 | 434 | 435 | 436 | 437 | 438 | 439 | 440 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.000000 | 0.620152 | 2.412450 | 2.815908 | 1.851574 | 0.4604609 | 0.9288162 | 0.9507279 | 2.024846 | 1.684588 | 1.562602 | 1.337659 | 1.725146 | 1.555848 | 1.209405 | 2.329964 | 1.1172125 | 2.496744 | 1.0442424 | 2.108562 | 0.9198763 | 2.675219 | 3.844109 | 6.840521 | 1.996175 | 0.9491927 | 2.5293261 | 2.435258 | 3.134218 | 3.699474 | 2.195132 | 2.297818 | 2.505257 | 2.996025 | 2.580317 | 1.2570434 | 3.008519 | 1.067672 | 1.709576 | 5.298941 | 3.785478 | 2.225185 | 1.305086 | 2.493664 | 0.8071114 | 2.839579 | 2.015219 | 9.717564 | 0.7914421 | 3.647675 | 2.6683808 | 2.146988 | 2.358176 | 1.288968 | 2.845809 | 2.060335 | 4.725849 | 1.2173930 | 2.360619 | 1.790340 | 1.073556 | 8.622706 | 1.0650610 | 1.513283 | 2.5868882 | 6.324153 | 2.567549 | 1.101571 | 2.673839 | 2.334694 | 3.5061541 | 6.140098 | 3.120222 | 1.888192 | 0.6940152 | 2.8478995 | 3.019711 | 2.988360 | 2.499260 | 2.344101 | 2.460176 | 1.681525 | 0.5825222 | 2.555608 | 1.052818 | 13.01495 | 9.773481 | 6.401152 | 3.477813 | 2.642477 | 2.6554973 | 3.0961414 | 5.130240 | 8.282782 | 0.9918894 | 2.473469 | 1.461051 | 2.902145 | 2.900292 | 2.7550589 | 1.523130 | 1.582254 | 1.0831281 | 6.118035 | 2.197141 | 2.367689 | 1.374260 | 1.632925 | 0.9005545 | 2.678801 | 2.4252862 | 1.340384 | 3.2418136 | 2.4042957 | 2.5472125 | 2.5077883 | 2.272386 | 2.096298 | 2.7631942 | 2.5232248 | 2.4544483 | 2.495549 | 2.645172 | 0.6492380 | 3.436461 | 7.119951 | 3.1927112 | 0.9856275 | 2.219317 | 3.765614 | 2.8714444 | 2.794698 | 2.283785 | 2.497213 | 2.414327 | 2.499788 | 2.115080 | 2.007361 | 2.157247 | 2.072152 | 2.095575 | 2.601640 | 3.166681 | 3.2151061 | 2.266917 | 3.784065 | 2.403025 | 2.141878 | 2.8221279 | 3.055119 | 2.370859 | 2.473120 | 2.565043 | 2.050250 | 2.935470 | 2.422979 | 1.601948 | 2.5475678 | 0.9682642 | 1.731974 | 1.397184 | 2.599930 | 2.3595352 | 3.285321 | 0.9280006 | 1.909736 | 1.2225387 | 2.288886 | 2.649268 | 2.632238 | 1.534524 | 3.483512 | 2.114305 | 2.232225 | 2.683468 | 1.025677 | 4.053654 | 1.906277 | 1.807178 | 2.541857 | 1.775984 | 10.423250 | 2.408049 | 19.87716 | 2.705040 | 2.177720 | 2.800323 | 2.602471 | 1.2428017 | 1.1097610 | 2.820121 | 2.599679 | 2.500320 | 1.982556 | 2.433316 | 2.8208722 | 5.144784 | 1.510205 | 2.324142 | 2.318465 | 3.042739 | 3.733230 | 3.482022 | 2.749006 | 2.599892 | 2.896711 | 2.608319 | 1.885562 | 2.273105 | 2.879442 | 2.477895 | 6.096286 | 2.490382 | 1.922470 | 1.738457 | 2.403091 | 4.477641 | 2.504291 | 2.772863 | 2.767133 | 2.518225 | 1.878491 | 3.138208 | 2.230447 | 2.563644 | 2.044545 | 1.397428 | 2.540775 | 2.777668 | 2.820725 | 2.317607 | 2.023743 | 2.639467 | 2.388312 | 2.490388 | 2.204974 | 2.525689 | 2.588923 | 2.659120 | 4.160830 | 3.644917 | 2.662727 | 2.647354 | 1.973486 | 1.737257 | 2.298772 | 2.345698 | 2.630838 | 1.988616 | 2.721709 | 2.554445 | 5.585288 | 3.020765 | 2.745277 | 2.661735 | 3.222110 | 1.905511 | 2.608944 | 4.857801 | 4.342550 | 2.026456 | 3.059403 | 2.627721 | 2.692940 | 1.918235 | 4.040050 | 2.463733 | 2.851627 | 2.051437 | 2.657270 | 2.975568 | 2.348011 | 2.126944 | 3.458776 | 2.602447 | 2.763284 | 2.796157 | 4.773910 | 2.6949955 | 0.7172437 | 2.408001 | 0.4969023 | 4.243469 | 3.578370 | 5.893521 | 3.625215 | 2.414388 | 3.0983098 | 2.550848 | 3.902122 | 2.433901 | 2.324110 | 2.493222 | 1.645870 | 2.324090 | 1.718870 | 2.497402 | 1.546328 | 1.472328 | 2.755050 | 1.125172 | 2.577483 | 1.779578 | 2.582662 | 3.354545 | 1.692876 | 2.441534 | 2.285174 | 2.455295 | 2.585994 | 3.906485 | 2.884227 | 3.998161 | 2.707455 | 2.172961 | 1.976527 | 2.527092 | 2.351716 | 2.264947 | 3.623340 | 2.411370 | 2.334406 | 2.570775 | 2.309349 | 2.599438 | 13.74114 | 2.658387 | 2.758382 | 3.017390 | 2.720258 | 2.420382 | 3.106613 | 2.549645 | 10.25892 | 3.336563 | 1.837854 | 2.409631 | 3.130803 | 4.610661 | 3.443050 | 1.583159 | 1.211341 | 2.198007 | 3.501663 | 2.681455 | 2.392510 | 1.548609 | 1.476492 | 2.432929 | 2.615152 | 2.614433 | 2.846128 | 2.768601 | 2.202564 | 2.4852457 | 2.798716 | 2.909795 | 2.367083 | 3.323659 | 2.332363 | 2.429414 | 2.680181 | 2.582355 | 2.448988 | 2.669030 | 1.1198898 | 2.365548 | 2.712538 | 2.643917 | 2.689048 | 2.365569 | 3.1660758 | 2.810793 | 0.8606141 | 2.417330 | 2.549851 | 1.553911 | 3.373196 | 2.400184 | 1.115939 | 3.034794 | 3.317088 | 2.908163 | 2.772951 | 3.446483 | 2.5221831 | 2.454920 | 2.527100 | 2.548406 | 2.5323267 | 2.9333884 | 2.465852 | 2.414575 | 3.079484 | 2.438117 | 2.5400628 | 0.9589695 | 2.949996 | 2.8702765 | 2.571611 | 2.744532 | 4.165058 | 2.465951 | 2.224331 | 2.3186967 | 2.573302 | 3.336707 | 1.816368 | 0.9799543 | 2.675998 | 2.225922 | 2.769132 | 2.459025 | 4.782897 | 2.7490648 | 0.7466402 | 1.1566828 | 1.923747 | 1.745959 | 2.050577 | 2.374162 | 0.8046204 | 2.647176 | 1.008799 | 1.174772 | 4.360966 | 2.079440 | 4.558736 | 2.162112 | 2.375680 | 2.526401 | 3.835168 | 2.431647 | 2.481654 | 1.868244 | 4.188151 | 3.714277 | 3.590575 | 2.346928 | 2.573812 |
| 0.620152 | 0.000000 | 2.170333 | 2.781998 | 1.920911 | 0.6784641 | 1.1396351 | 0.7164900 | 2.071625 | 1.333053 | 1.087422 | 1.562242 | 2.060281 | 1.605518 | 1.502044 | 2.539311 | 0.7459657 | 2.439154 | 1.1522504 | 2.125729 | 1.2480569 | 2.668116 | 3.904920 | 6.703197 | 1.962013 | 1.2768979 | 2.6091309 | 2.732916 | 2.802795 | 4.175155 | 2.359594 | 2.218100 | 2.930239 | 3.231743 | 2.603998 | 0.9609913 | 3.300854 | 1.054614 | 1.479775 | 5.575921 | 3.731049 | 2.433893 | 1.248521 | 2.170327 | 0.8381497 | 2.524759 | 1.637652 | 9.588071 | 0.7784813 | 3.369111 | 2.5994380 | 2.129450 | 2.826819 | 0.950533 | 3.226969 | 2.041268 | 4.473040 | 0.9943292 | 2.674058 | 1.716608 | 1.155835 | 8.491809 | 0.8311212 | 1.203668 | 2.4916373 | 6.047459 | 2.453293 | 1.282473 | 2.344356 | 2.532160 | 3.4396212 | 5.943878 | 2.899524 | 1.889524 | 0.5693493 | 3.0881554 | 2.705831 | 2.788034 | 2.673908 | 2.259615 | 2.595191 | 1.369687 | 0.4299921 | 2.893194 | 1.240487 | 12.79296 | 9.704220 | 6.434210 | 3.328765 | 2.888641 | 2.7857988 | 3.0171589 | 4.855283 | 8.002218 | 0.8278463 | 2.391501 | 1.309248 | 2.927100 | 2.919223 | 2.7550136 | 1.248475 | 1.116343 | 0.6858667 | 6.215100 | 2.465498 | 2.646393 | 1.046566 | 1.340833 | 0.6804688 | 2.442299 | 2.5792011 | 1.190899 | 3.2263431 | 2.5734494 | 2.8140789 | 2.6388750 | 2.485385 | 2.130122 | 2.9443148 | 2.5498869 | 2.7077371 | 2.511351 | 2.844241 | 0.7405835 | 3.774496 | 7.352906 | 3.2651965 | 1.2731194 | 2.186244 | 4.210032 | 2.7930352 | 2.864150 | 2.617328 | 2.713299 | 2.569254 | 2.604969 | 1.714314 | 1.945999 | 2.361786 | 2.145760 | 2.286692 | 3.096209 | 3.588161 | 3.1501042 | 2.720074 | 3.646921 | 2.558977 | 2.315999 | 2.8221743 | 3.512329 | 2.702215 | 2.442283 | 2.843362 | 2.047789 | 2.994883 | 2.096226 | 1.263136 | 2.7882482 | 0.5962734 | 1.417875 | 1.050356 | 2.798456 | 2.5842936 | 2.999673 | 0.6713014 | 1.663001 | 0.8904401 | 2.299839 | 2.749483 | 2.658352 | 1.170557 | 3.174518 | 1.958281 | 1.888116 | 2.524530 | 0.655302 | 4.365742 | 2.077059 | 2.081553 | 2.351302 | 1.900888 | 10.617149 | 2.127569 | 19.65213 | 2.701127 | 2.328058 | 2.741031 | 2.205182 | 0.9870185 | 0.6782545 | 2.899628 | 2.879040 | 2.615397 | 1.650343 | 2.539989 | 2.7912159 | 5.057187 | 1.092908 | 2.375833 | 2.545668 | 2.655165 | 3.393023 | 3.492744 | 2.757669 | 2.529296 | 2.550871 | 2.741326 | 1.648181 | 2.066884 | 2.636277 | 2.772169 | 5.834455 | 2.569931 | 1.651675 | 1.412845 | 1.927654 | 4.197112 | 2.780419 | 2.377339 | 2.798864 | 2.778715 | 1.759570 | 2.989934 | 1.989753 | 2.706763 | 2.330559 | 1.719066 | 2.546787 | 2.750273 | 2.719939 | 2.119902 | 1.927955 | 3.084512 | 2.371399 | 2.574239 | 2.169618 | 2.627839 | 2.803196 | 2.809488 | 4.516578 | 3.572688 | 3.042609 | 2.865038 | 2.124495 | 1.343247 | 1.976872 | 2.491640 | 2.856780 | 2.407317 | 2.645438 | 2.561460 | 5.234077 | 2.824141 | 3.072770 | 2.524771 | 3.606649 | 2.192517 | 2.527881 | 5.201899 | 4.771057 | 2.279366 | 3.003769 | 3.044992 | 2.466362 | 1.539153 | 3.709667 | 1.990872 | 3.004652 | 1.899218 | 2.897001 | 2.911250 | 2.321755 | 1.985889 | 3.812413 | 2.574937 | 2.784669 | 3.040520 | 4.574751 | 2.6842830 | 0.9578527 | 2.429211 | 0.7045248 | 4.600446 | 3.651516 | 6.192215 | 4.097986 | 2.532340 | 3.0777267 | 2.878420 | 4.379783 | 2.427477 | 2.342710 | 2.326843 | 1.312131 | 2.668239 | 1.833833 | 2.755576 | 1.543785 | 1.321134 | 2.772909 | 1.479965 | 2.282931 | 1.564784 | 2.363396 | 2.997126 | 1.416350 | 2.105007 | 2.631933 | 2.592239 | 2.302584 | 3.691251 | 3.241143 | 3.756663 | 2.752239 | 2.265259 | 1.591483 | 2.568693 | 2.214394 | 2.462636 | 3.452709 | 2.423809 | 2.377864 | 2.745276 | 2.436623 | 3.004863 | 13.46820 | 2.619321 | 2.772458 | 3.022578 | 2.591115 | 2.442527 | 2.888650 | 2.872483 | 10.02888 | 3.217769 | 2.155489 | 2.711147 | 3.021697 | 4.289330 | 3.147484 | 1.299998 | 0.828172 | 1.960374 | 3.215699 | 2.678148 | 2.244943 | 1.110863 | 1.728473 | 2.443512 | 2.343804 | 2.615834 | 2.466282 | 2.779942 | 1.861159 | 2.4401561 | 2.818341 | 3.227230 | 2.101995 | 3.048599 | 2.329918 | 2.763011 | 2.763249 | 2.606314 | 2.515372 | 2.696889 | 0.7738633 | 2.562398 | 2.764248 | 3.006901 | 2.812714 | 2.809126 | 3.2393533 | 2.730857 | 0.9984568 | 2.511805 | 2.687418 | 1.018437 | 3.854496 | 2.458623 | 1.079183 | 3.378569 | 3.312857 | 3.258790 | 2.652220 | 3.253306 | 2.6722709 | 2.371169 | 2.884296 | 2.720104 | 2.6302769 | 2.8069378 | 2.448401 | 2.376902 | 3.176585 | 2.659882 | 2.6516382 | 0.4308295 | 2.939790 | 2.9331844 | 2.740642 | 2.706698 | 4.173217 | 2.974483 | 2.775795 | 2.5293165 | 2.661988 | 3.391937 | 1.501668 | 1.0186183 | 2.531567 | 2.211974 | 2.620719 | 2.260941 | 4.446768 | 2.6378491 | 0.5286671 | 0.8481803 | 1.734135 | 1.398695 | 2.214883 | 2.123691 | 1.0509046 | 3.070028 | 1.385235 | 1.379042 | 4.151780 | 1.840948 | 4.546713 | 2.072267 | 2.499529 | 2.296263 | 3.563641 | 2.835683 | 2.440757 | 2.202477 | 4.114048 | 4.070721 | 3.408252 | 2.501223 | 2.648764 |
| 2.412450 | 2.170333 | 0.000000 | 3.680215 | 1.728199 | 2.3524738 | 2.7665390 | 1.9603064 | 3.674148 | 2.524751 | 2.385904 | 2.949354 | 2.768076 | 3.026187 | 2.580937 | 3.958012 | 2.5319328 | 2.531155 | 1.9687098 | 3.780460 | 2.3421551 | 3.940453 | 3.925102 | 5.348163 | 1.636735 | 2.9711537 | 3.8534730 | 4.037325 | 2.882931 | 5.007855 | 3.177927 | 3.631156 | 4.192736 | 4.238655 | 3.979389 | 2.7494280 | 3.352266 | 2.755301 | 3.076675 | 5.821906 | 3.301583 | 3.215371 | 3.016562 | 3.541825 | 2.8324339 | 2.653713 | 2.908926 | 9.624556 | 2.4003158 | 4.243670 | 3.4440827 | 3.828867 | 3.669821 | 2.835624 | 4.149989 | 3.532221 | 5.146671 | 2.5478252 | 4.020210 | 3.312034 | 2.860303 | 8.905342 | 2.4987514 | 2.662840 | 3.6891806 | 6.585274 | 4.019638 | 2.819092 | 2.990695 | 4.010412 | 4.2249561 | 4.134897 | 4.149593 | 3.101062 | 2.2301536 | 4.1004070 | 3.879960 | 3.804500 | 4.067582 | 3.457451 | 3.945953 | 3.030178 | 2.0373102 | 4.056692 | 2.898761 | 13.06836 | 10.166170 | 4.862571 | 4.440246 | 3.421233 | 4.0668181 | 3.9926901 | 4.938167 | 8.075351 | 2.8166317 | 3.820152 | 2.914309 | 4.220496 | 4.198754 | 4.0461746 | 2.062851 | 2.463678 | 2.3385204 | 6.488696 | 3.883498 | 3.766689 | 2.368863 | 2.672426 | 2.4263069 | 3.808322 | 3.6167554 | 2.634636 | 3.7319015 | 3.6361583 | 3.8737187 | 3.9346769 | 3.978052 | 3.568782 | 4.0266475 | 3.8675477 | 3.8537862 | 3.955509 | 3.977221 | 2.7915952 | 4.494778 | 7.774588 | 4.3173897 | 2.2341268 | 3.935849 | 4.828221 | 3.9004866 | 4.129534 | 4.022153 | 4.038778 | 3.930333 | 3.996776 | 3.295251 | 3.866832 | 3.023478 | 3.639446 | 3.212873 | 4.145371 | 4.771290 | 4.2596242 | 4.127026 | 3.873464 | 3.928200 | 3.813328 | 4.0473940 | 4.350990 | 3.874373 | 3.513217 | 4.170583 | 3.407333 | 4.276477 | 2.974998 | 2.048999 | 4.0706085 | 2.2553767 | 3.004318 | 2.263499 | 3.582133 | 3.9249222 | 4.052941 | 2.7054776 | 2.635310 | 1.5882315 | 3.759647 | 4.033008 | 3.897640 | 2.884362 | 2.978529 | 3.647113 | 3.178135 | 3.647284 | 2.306151 | 5.270194 | 3.092686 | 3.797180 | 3.311612 | 2.989873 | 10.382171 | 3.896814 | 17.80138 | 4.139839 | 3.895222 | 3.924740 | 3.919325 | 2.7697375 | 2.2818397 | 4.213903 | 4.209450 | 3.984460 | 3.046341 | 3.749693 | 3.7220556 | 5.305138 | 2.832438 | 3.801972 | 3.973114 | 3.869988 | 4.200078 | 2.808933 | 4.146329 | 3.695127 | 3.827584 | 4.070368 | 3.342502 | 3.625740 | 3.821500 | 3.623593 | 6.244701 | 3.574365 | 3.395354 | 3.137133 | 3.203459 | 5.113406 | 3.884362 | 1.642944 | 4.126829 | 3.983799 | 3.766263 | 4.152313 | 3.130311 | 3.961935 | 3.771890 | 3.052259 | 3.855190 | 4.070657 | 3.848888 | 2.990252 | 3.708561 | 4.200823 | 3.983299 | 3.341802 | 3.820501 | 4.027292 | 3.869409 | 4.122212 | 5.006884 | 4.074887 | 4.097336 | 4.073330 | 3.178164 | 3.129877 | 2.810411 | 3.861713 | 4.039632 | 3.873617 | 3.992520 | 3.862022 | 5.288472 | 3.721074 | 4.187898 | 3.183171 | 4.576966 | 3.842597 | 3.732751 | 5.442140 | 5.528259 | 3.806601 | 4.181001 | 4.194467 | 3.885852 | 3.107252 | 4.077598 | 2.978582 | 3.312225 | 3.449142 | 4.177010 | 4.184746 | 3.518833 | 3.590479 | 4.481024 | 3.751691 | 4.064519 | 3.742454 | 5.304542 | 3.7231694 | 2.3862236 | 3.701392 | 2.1463697 | 5.241309 | 4.398962 | 6.554537 | 5.097766 | 3.999455 | 4.0618881 | 4.171401 | 5.298693 | 3.577082 | 3.772460 | 3.644898 | 3.161897 | 3.679914 | 3.405060 | 3.942550 | 3.214098 | 3.020680 | 4.099847 | 2.780872 | 3.785185 | 3.244511 | 3.763791 | 3.718954 | 3.329117 | 3.356930 | 3.880887 | 3.913745 | 3.646132 | 4.438422 | 4.481652 | 4.707011 | 4.057834 | 3.618998 | 2.383210 | 3.767872 | 3.458730 | 4.007157 | 4.299640 | 3.779717 | 3.678147 | 4.086899 | 3.915756 | 4.073030 | 13.32005 | 3.840394 | 4.031003 | 3.906591 | 3.857343 | 3.723040 | 4.006177 | 3.558475 | 10.40032 | 3.412929 | 3.149494 | 3.948815 | 4.050905 | 5.024055 | 3.856963 | 2.597679 | 2.170255 | 3.450881 | 4.246098 | 4.002351 | 3.864120 | 2.634995 | 2.993674 | 3.816714 | 3.559735 | 3.811566 | 3.003658 | 4.113285 | 3.075857 | 3.3212077 | 4.157043 | 4.305234 | 3.441821 | 3.262242 | 3.762545 | 4.051505 | 4.046494 | 3.898029 | 3.966838 | 3.995446 | 1.7448115 | 3.946988 | 4.057938 | 4.187867 | 4.102992 | 3.443261 | 4.2304696 | 2.604013 | 1.8861251 | 3.849312 | 4.100138 | 2.476441 | 4.677167 | 3.720080 | 2.878641 | 4.486260 | 4.338076 | 3.864348 | 3.986922 | 3.315227 | 3.8257203 | 3.704331 | 4.008631 | 4.028033 | 3.8986296 | 3.8844527 | 3.497784 | 3.815500 | 4.222536 | 4.003279 | 4.0090022 | 2.0561802 | 4.039078 | 4.1601682 | 4.043822 | 3.837331 | 4.692361 | 4.278324 | 4.205623 | 3.8679843 | 3.994594 | 4.327697 | 3.029287 | 2.7588026 | 2.249325 | 3.356761 | 2.706781 | 3.973476 | 4.798692 | 3.9891743 | 2.1599432 | 2.5847750 | 3.509129 | 2.913237 | 3.481593 | 3.558453 | 2.3924940 | 4.094929 | 2.837229 | 2.989307 | 4.642920 | 3.315544 | 5.224369 | 3.782032 | 3.340175 | 3.469769 | 4.282591 | 4.148364 | 3.594356 | 3.732618 | 4.707595 | 4.576351 | 4.143755 | 3.442665 | 4.094651 |
| 2.815908 | 2.781998 | 3.680215 | 0.000000 | 2.659964 | 2.5817546 | 2.4676659 | 2.4619893 | 1.484365 | 3.791430 | 2.809077 | 2.124645 | 3.698296 | 3.114515 | 3.294599 | 1.355338 | 3.2648919 | 1.742488 | 2.5035669 | 1.518280 | 2.2950519 | 1.000697 | 1.841387 | 8.441497 | 3.475779 | 2.7013656 | 0.8118723 | 1.311934 | 5.255089 | 2.624154 | 1.512222 | 1.334066 | 1.517953 | 1.450726 | 1.636961 | 3.0023952 | 1.971609 | 3.228691 | 4.139641 | 3.513575 | 1.850347 | 1.201656 | 3.423722 | 4.523861 | 2.9268672 | 4.996961 | 4.142268 | 11.697694 | 2.8289858 | 5.812456 | 0.8089728 | 1.669532 | 3.562222 | 3.450475 | 1.411790 | 1.374538 | 6.917709 | 3.4836847 | 1.323084 | 1.662718 | 2.705109 | 10.651445 | 2.2541113 | 3.524589 | 0.9184272 | 8.343130 | 1.906755 | 2.993391 | 1.279767 | 1.406272 | 0.9434664 | 4.914947 | 1.189026 | 2.051493 | 2.4794017 | 0.8824137 | 1.141385 | 5.025585 | 1.100070 | 1.500158 | 1.406342 | 3.786775 | 2.8850909 | 1.125978 | 2.447738 | 15.03498 | 12.106267 | 5.095091 | 1.023756 | 1.266314 | 0.8737758 | 0.5982309 | 7.169875 | 5.913754 | 3.4468920 | 1.690850 | 2.942340 | 1.709681 | 1.677065 | 0.6758855 | 3.148556 | 3.624213 | 2.6378354 | 4.200038 | 1.163123 | 1.158808 | 3.291353 | 3.616374 | 2.9286074 | 4.728111 | 0.9044944 | 3.584240 | 0.8058153 | 0.7322717 | 0.9312205 | 0.8958048 | 1.179543 | 1.185160 | 0.6891632 | 0.8155746 | 0.9137331 | 1.293343 | 1.070037 | 2.7693769 | 1.838838 | 5.425332 | 0.8082797 | 2.7665028 | 2.057036 | 2.507939 | 0.7047406 | 1.671894 | 1.334632 | 1.395917 | 1.359376 | 1.329725 | 1.367437 | 1.867532 | 1.454710 | 1.439952 | 1.102668 | 2.470508 | 2.295225 | 0.7795744 | 1.534569 | 5.560381 | 1.314872 | 1.241680 | 0.9803606 | 1.891904 | 1.227450 | 1.581136 | 1.011857 | 1.782470 | 1.821237 | 4.470510 | 3.369998 | 0.8887983 | 2.8276286 | 3.712925 | 2.943076 | 1.190500 | 0.8990607 | 5.305217 | 2.6569369 | 3.697914 | 2.9562355 | 1.577969 | 1.336613 | 1.123088 | 3.518496 | 5.673672 | 1.814861 | 4.117356 | 1.314139 | 3.125465 | 2.759067 | 1.308235 | 1.446828 | 1.048634 | 1.538103 | 9.448508 | 2.461656 | 18.60734 | 1.774564 | 1.355737 | 1.143193 | 1.647729 | 3.0294447 | 3.3820755 | 1.012749 | 1.330285 | 1.455856 | 3.567467 | 1.279874 | 0.6927844 | 4.021769 | 2.904509 | 1.780159 | 2.371938 | 4.405746 | 5.119150 | 3.539277 | 2.718168 | 2.385583 | 4.448295 | 2.414374 | 2.691850 | 2.539324 | 4.496528 | 2.089752 | 7.590775 | 2.178035 | 2.700387 | 3.132026 | 3.479389 | 6.060813 | 1.942809 | 3.779772 | 2.298546 | 2.078218 | 3.038748 | 1.864869 | 1.885264 | 2.349276 | 2.268121 | 2.465689 | 2.225849 | 2.827800 | 1.549137 | 1.808337 | 2.924863 | 2.594698 | 2.698797 | 1.591233 | 2.665791 | 2.081620 | 1.801455 | 2.371508 | 3.729446 | 2.350863 | 2.305628 | 1.812935 | 2.302096 | 2.747769 | 3.647298 | 2.138900 | 2.064045 | 2.419290 | 1.797060 | 2.336948 | 6.873127 | 1.601197 | 3.413696 | 3.380975 | 2.673463 | 2.436313 | 1.892557 | 4.466465 | 4.309992 | 2.471139 | 1.768747 | 2.451022 | 2.229582 | 3.589317 | 4.393968 | 3.754397 | 2.035430 | 3.539624 | 2.050461 | 1.994860 | 2.524495 | 2.889188 | 1.930269 | 1.561092 | 1.652443 | 1.328193 | 2.299001 | 0.3215676 | 2.5593155 | 1.419153 | 2.4305693 | 2.862073 | 1.343720 | 4.541792 | 2.485526 | 1.301201 | 0.5483017 | 1.327319 | 2.662499 | 1.495850 | 1.106607 | 1.046843 | 3.412220 | 2.326287 | 2.471708 | 1.892173 | 2.419688 | 2.700903 | 2.725357 | 2.453613 | 4.194980 | 3.123367 | 3.780003 | 4.678831 | 3.446235 | 4.033274 | 2.190011 | 2.395910 | 4.434501 | 2.397155 | 2.646797 | 5.118988 | 1.748511 | 2.118939 | 3.699871 | 1.972003 | 2.508169 | 2.104538 | 5.466758 | 2.603364 | 1.873493 | 1.843863 | 2.458574 | 2.380959 | 12.64196 | 2.025911 | 2.678597 | 1.459571 | 1.791216 | 1.711221 | 4.736874 | 2.227990 | 11.40878 | 2.263361 | 2.978798 | 2.279176 | 1.603188 | 3.234368 | 2.074589 | 3.254689 | 2.998937 | 1.946721 | 5.257493 | 1.308167 | 1.851864 | 3.233225 | 2.998805 | 1.396550 | 4.697748 | 1.278067 | 4.686137 | 1.651276 | 4.059662 | 0.8990239 | 1.757234 | 1.185162 | 3.904381 | 2.707018 | 1.765276 | 1.090801 | 1.406562 | 1.553524 | 1.552081 | 1.484239 | 2.7212014 | 1.252214 | 1.433638 | 1.246639 | 1.396897 | 3.368416 | 0.7858194 | 1.791206 | 2.3890090 | 1.253365 | 1.500449 | 2.851814 | 2.332468 | 1.557689 | 2.800429 | 1.419227 | 1.068434 | 2.138251 | 1.151709 | 3.314617 | 0.8844676 | 1.402510 | 1.175003 | 1.171864 | 0.9087332 | 0.9062629 | 1.442185 | 1.658977 | 1.250127 | 1.145328 | 0.9913819 | 2.9733043 | 0.512913 | 0.7199765 | 1.199983 | 1.003238 | 1.812797 | 1.799818 | 2.162393 | 0.8373688 | 1.185714 | 1.318033 | 3.853838 | 2.3349533 | 1.581414 | 1.200071 | 1.646237 | 2.063986 | 2.524649 | 0.8205687 | 2.4735965 | 3.4129799 | 1.811518 | 3.788177 | 1.327051 | 2.277192 | 2.4799903 | 1.564270 | 2.294350 | 2.663343 | 1.866637 | 1.928986 | 3.104099 | 1.450612 | 1.390074 | 1.983874 | 1.631795 | 1.503086 | 1.404364 | 1.421901 | 2.710103 | 2.132232 | 5.669890 | 1.162555 | 1.675524 |
| 1.851574 | 1.920911 | 1.728199 | 2.659964 | 0.000000 | 1.8674171 | 2.0168263 | 1.6289660 | 3.189042 | 2.598596 | 2.227244 | 2.017647 | 1.736699 | 2.105596 | 1.663532 | 3.199823 | 2.5397479 | 2.585804 | 0.9388877 | 3.195140 | 1.3351380 | 3.295843 | 2.470645 | 6.057667 | 1.113528 | 2.0969324 | 3.0230195 | 3.092124 | 3.546989 | 3.404093 | 2.127045 | 3.259521 | 2.997422 | 2.654035 | 3.597712 | 2.5844163 | 1.818872 | 2.192622 | 3.105833 | 4.198962 | 2.226187 | 2.041299 | 2.584810 | 3.617476 | 2.3126815 | 3.290611 | 3.039870 | 9.777747 | 1.9030039 | 4.497656 | 2.8769500 | 3.435271 | 2.131585 | 2.811861 | 2.673212 | 3.099270 | 5.529063 | 2.5248627 | 3.110760 | 3.000515 | 2.299077 | 8.947871 | 1.9647477 | 2.467535 | 3.1337884 | 6.980894 | 3.745133 | 1.921386 | 2.863909 | 3.334281 | 3.1168152 | 4.707001 | 3.549357 | 1.859996 | 1.7729800 | 2.8211188 | 3.449894 | 3.729654 | 3.213797 | 3.225192 | 3.324651 | 3.060857 | 1.7171825 | 2.799504 | 2.143607 | 13.41539 | 10.458262 | 4.672532 | 3.603439 | 1.842502 | 3.1500418 | 3.0212796 | 5.460610 | 7.566477 | 2.6392163 | 3.541261 | 2.691365 | 3.810879 | 3.784991 | 3.1809175 | 1.859879 | 2.643253 | 2.1636083 | 5.057602 | 2.948317 | 2.761389 | 2.468992 | 2.545723 | 2.3448837 | 3.779182 | 2.7679715 | 2.341310 | 2.5280613 | 2.6313287 | 2.7043275 | 3.0502858 | 3.137994 | 2.988642 | 2.7231862 | 3.1088550 | 2.7145865 | 3.421402 | 3.086661 | 2.1677133 | 2.813566 | 6.209507 | 3.0719138 | 1.1652848 | 3.739327 | 3.205562 | 3.1798728 | 3.690584 | 3.072899 | 3.308518 | 3.288838 | 3.368294 | 2.978628 | 3.521884 | 2.324641 | 3.182519 | 2.073377 | 2.995935 | 3.217290 | 3.3529613 | 3.061667 | 3.801074 | 3.263139 | 3.113746 | 3.3657920 | 2.859271 | 2.860581 | 3.295709 | 2.959816 | 3.219670 | 3.869826 | 3.285173 | 2.229991 | 2.8985587 | 2.1606237 | 3.013678 | 2.236457 | 2.784716 | 2.8672305 | 4.155701 | 2.3056653 | 2.301177 | 1.7795453 | 3.396864 | 3.420889 | 3.287949 | 2.874326 | 3.915474 | 3.460383 | 3.263822 | 3.333320 | 2.310952 | 3.637975 | 2.381347 | 3.068790 | 2.968780 | 2.395028 | 9.429913 | 3.855605 | 18.47942 | 3.761925 | 3.248183 | 3.413519 | 3.545707 | 2.3875797 | 2.4595761 | 3.104732 | 3.269949 | 3.418117 | 2.949699 | 3.147229 | 2.5750863 | 4.244018 | 2.750165 | 2.882476 | 3.242590 | 3.929788 | 4.412849 | 2.132454 | 3.758620 | 3.316654 | 3.957551 | 3.446000 | 3.021448 | 3.356352 | 3.908351 | 2.480572 | 6.533455 | 2.993348 | 3.197933 | 3.021729 | 3.393607 | 5.328885 | 2.717079 | 2.703309 | 3.533670 | 2.983707 | 3.452213 | 3.456265 | 2.657767 | 3.316292 | 2.932316 | 1.897906 | 3.334952 | 3.737071 | 3.057166 | 2.199549 | 3.497583 | 2.836041 | 3.632622 | 2.236743 | 3.501411 | 3.254647 | 2.662200 | 3.439574 | 3.311387 | 2.846401 | 2.696193 | 2.862875 | 2.541809 | 2.905333 | 3.040426 | 3.138399 | 3.081334 | 2.942822 | 3.227264 | 3.391896 | 5.787395 | 3.035071 | 2.971601 | 3.095383 | 3.019494 | 3.092747 | 3.048485 | 3.837723 | 3.883644 | 3.169394 | 3.370425 | 2.914330 | 3.426286 | 3.207091 | 4.198733 | 3.198514 | 1.964180 | 3.034048 | 3.128474 | 3.511349 | 3.258286 | 3.445858 | 2.812684 | 3.456831 | 3.689957 | 2.097508 | 4.526944 | 2.7920865 | 1.4842460 | 3.293382 | 1.5168042 | 3.551096 | 3.015270 | 4.968732 | 3.508484 | 3.350417 | 2.9599954 | 3.143251 | 3.698779 | 3.260891 | 3.169667 | 3.194354 | 3.089427 | 2.430859 | 2.794702 | 2.708900 | 2.634048 | 2.632102 | 3.730616 | 1.903442 | 3.731616 | 3.071544 | 3.641928 | 4.033194 | 3.314749 | 3.373793 | 2.802797 | 3.296833 | 3.884542 | 3.721861 | 2.960409 | 4.755579 | 3.186532 | 2.842315 | 2.891959 | 3.099208 | 3.230008 | 3.124316 | 4.529050 | 3.451041 | 2.912207 | 2.987440 | 3.042640 | 2.644866 | 12.88020 | 3.275343 | 3.671748 | 2.729556 | 3.212211 | 2.895227 | 3.939065 | 2.236998 | 10.59294 | 2.431931 | 1.909299 | 3.034453 | 3.196788 | 4.507477 | 3.369767 | 2.622620 | 2.198790 | 3.430093 | 4.320282 | 3.486536 | 3.595175 | 2.538697 | 1.753863 | 3.373085 | 3.610407 | 3.330488 | 3.420392 | 3.717194 | 3.179745 | 2.4339401 | 3.782850 | 2.978898 | 3.308715 | 3.640450 | 3.521631 | 2.836622 | 3.483165 | 3.499422 | 3.458686 | 3.537415 | 1.7671353 | 3.199927 | 3.542392 | 2.882688 | 3.470150 | 1.831691 | 2.9539267 | 2.753418 | 1.0752942 | 3.250849 | 3.506326 | 2.298820 | 3.095307 | 3.348540 | 2.463526 | 2.986524 | 3.205219 | 2.307815 | 3.483843 | 3.589185 | 2.9457114 | 3.371478 | 2.761604 | 3.237613 | 3.1040801 | 3.3309433 | 3.193625 | 3.524965 | 2.893283 | 3.100779 | 3.2022030 | 2.0460238 | 3.089523 | 3.2142474 | 3.265339 | 3.266703 | 3.386504 | 2.954890 | 3.094551 | 2.6878725 | 3.326201 | 2.957140 | 2.883625 | 2.1316289 | 2.178977 | 2.873471 | 2.818366 | 3.740494 | 4.345866 | 3.2952817 | 1.7655472 | 2.5451139 | 3.211381 | 2.961067 | 2.911047 | 3.456082 | 1.4703631 | 2.703637 | 1.788045 | 2.035403 | 3.871588 | 2.809540 | 4.195951 | 3.366462 | 2.812272 | 3.269620 | 3.624700 | 2.967063 | 3.278835 | 2.759970 | 3.612115 | 2.904650 | 4.190340 | 2.755297 | 3.640150 |
showing only the first 5 rows
Since our distances are relatively small, we will use the complete linkage method during hierarchical clustering.
cluster_hier <- agnes(x = dist_total, method = "complete")
dendo1 <- fviz_dend(x = cluster_hier, show_labels = FALSE)
## Warning: The `<scale>` argument of `guides()` cannot be `FALSE`. Use "none" instead as
## of ggplot2 3.3.4.
## ℹ The deprecated feature was likely used in the factoextra package.
## Please report the issue at <]8;;https://github.com/kassambara/factoextra/issueshttps://github.com/kassambara/factoextra/issues]8;;>.
dendo1
We can observe the presence of some significant “jumps” in the dendrogram. We also notice the presence of well-defined clusters and others that are less clear. This suggests that there may be some outliers in the database.
After analyzing the dendrogram in a hierarchical clustering process, we can choose the number of clusters by examining the dendrogram’s structure and identifying cuts that appear to be the most meaningful or relevant for our objective.
Setting a height of 7 for the dendrogram cluster definition:
dendo_clusters <- fviz_dend(x = cluster_hier,
h = 7,
color_labels_by_k = F,
rect = T,
rect_fill = T,
lwd = 1,
ggtheme = theme_bw(),
show_labels = FALSE)
dendo_clusters
The height 7 was chosen simply because it seems to provide a good separation of clusters based on the size of the jumps seen in the dendrogram. The cut at height 7 resulted in 12 different clusters, but half of these clusters are clustered on the far right of the dendrogram due to the presence of the outliers mentioned earlier.
dados_completos <- cbind(dados_padronizados, Channel_2=dados_dummies$Channel_2, Region_1=dados_dummies$Region_1, Region_2=dados_dummies$Region_2)
dados_completos$cluster_hier <- factor(cutree(tree = cluster_hier, k = 12))
Note: 12 is the number of clusters created by cutting at height 7. Therefore, the argument ‘k’ indicates the number of clusters. Next, we will check if all variables contribute to the formation of the groups.
summary(anova_channel2 <- aov(formula = Channel_2 ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 63.77 5.798 76.59 <2e-16 ***
## Residuals 428 32.40 0.076
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_region1 <- aov(formula = Region_1 ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 1.77 0.1607 1.114 0.348
## Residuals 428 61.76 0.1443
summary(anova_region2 <- aov(formula = Region_2 ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 2.04 0.18526 1.985 0.0284 *
## Residuals 428 39.94 0.09332
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_fresh <- aov(formula = Fresh ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 234.5 21.316 44.6 <2e-16 ***
## Residuals 428 204.5 0.478
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_milk <- aov(formula = Milk ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 311.8 28.342 95.34 <2e-16 ***
## Residuals 428 127.2 0.297
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_grocery <- aov(formula = Grocery ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 354.6 32.24 163.6 <2e-16 ***
## Residuals 428 84.4 0.20
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_frozen <- aov(formula = Frozen ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 258.7 23.522 55.85 <2e-16 ***
## Residuals 428 180.3 0.421
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_detergents <- aov(formula = Detergents_Paper ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 380.9 34.63 255.2 <2e-16 ***
## Residuals 428 58.1 0.14
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_delicassen <- aov(formula = Delicassen ~ cluster_hier,
data = dados_completos))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_hier 11 356.5 32.41 168.1 <2e-16 ***
## Residuals 428 82.5 0.19
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
For a confidence level of 95%, only the variable “Region_1” cannot be considered significant for the formation of at least one cluster.
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Fresh),
sd = sd(Fresh),
min = min(Fresh),
max = max(Fresh))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 -0.560 0.325 -0.947 0.286
## 2 2 -0.0112 0.760 -0.949 3.49
## 3 3 1.37 1.01 0.497 2.47
## 4 4 2.70 0.989 1.40 5.08
## 5 5 1.77 0.858 0.864 2.57
## 6 6 -0.494 0.436 -0.942 0.794
## 7 7 0.325 NA 0.325 0.325
## 8 8 -0.0543 NA -0.0543 -0.0543
## 9 9 7.92 NA 7.92 7.92
## 10 10 1.96 NA 1.96 1.96
## 11 11 1.64 NA 1.64 1.64
## 12 12 -0.272 NA -0.272 -0.272
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Milk),
sd = sd(Milk),
min = min(Milk),
max = max(Milk))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.533 0.701 -0.613 2.72
## 2 2 -0.359 0.355 -0.778 1.48
## 3 3 1.14 2.62 -0.614 4.15
## 4 4 -0.379 0.288 -0.747 0.184
## 5 5 6.72 2.38 4.41 9.17
## 6 6 1.32 0.984 -0.279 3.26
## 7 7 5.47 NA 5.47 5.47
## 8 8 -0.367 NA -0.367 -0.367
## 9 9 3.23 NA 3.23 3.23
## 10 10 5.17 NA 5.17 5.17
## 11 11 1.49 NA 1.49 1.49
## 12 12 -0.111 NA -0.111 -0.111
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Grocery),
sd = sd(Grocery),
min = min(Grocery),
max = max(Grocery))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.570 0.570 -0.662 2.21
## 2 2 -0.408 0.343 -0.836 0.949
## 3 3 0.958 0.817 0.0174 1.48
## 4 4 -0.364 0.374 -0.787 0.490
## 5 5 4.33 1.56 2.54 5.43
## 6 6 2.09 0.766 0.928 3.99
## 7 7 8.93 NA 8.93 8.93
## 8 8 -0.620 NA -0.620 -0.620
## 9 9 1.07 NA 1.07 1.07
## 10 10 1.29 NA 1.29 1.29
## 11 11 0.597 NA 0.597 0.597
## 12 12 6.24 NA 6.24 6.24
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Frozen),
sd = sd(Frozen),
min = min(Frozen),
max = max(Frozen))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 -0.327 0.326 -0.626 1.46
## 2 2 -0.0149 0.707 -0.628 3.22
## 3 3 0.523 0.127 0.429 0.667
## 4 4 0.606 1.07 -0.420 3.08
## 5 5 0.193 0.713 -0.429 0.970
## 6 6 -0.243 0.360 -0.625 0.757
## 7 7 -0.421 NA -0.421 -0.421
## 8 8 6.58 NA 6.58 6.58
## 9 9 2.82 NA 2.82 2.82
## 10 10 6.89 NA 6.89 6.89
## 11 11 11.9 NA 11.9 11.9
## 12 12 -0.606 NA -0.606 -0.606
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Detergents_Paper),
sd = sd(Detergents_Paper),
min = min(Detergents_Paper),
max = max(Detergents_Paper))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.549 0.512 -0.545 2.00
## 2 2 -0.401 0.269 -0.604 0.844
## 3 3 0.101 0.324 -0.273 0.305
## 4 4 -0.486 0.105 -0.600 -0.283
## 5 5 4.36 0.702 3.61 5.00
## 6 6 2.45 0.760 1.34 4.48
## 7 7 7.96 NA 7.96 7.96
## 8 8 -0.589 NA -0.589 -0.589
## 9 9 0.433 NA 0.433 0.433
## 10 10 -0.554 NA -0.554 -0.554
## 11 11 -0.338 NA -0.338 -0.338
## 12 12 7.39 NA 7.39 7.39
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Delicassen),
sd = sd(Delicassen),
min = min(Delicassen),
max = max(Delicassen))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.0384 0.545 -0.540 2.24
## 2 2 -0.139 0.391 -0.540 1.89
## 3 3 4.82 0.433 4.55 5.32
## 4 4 -0.0715 0.337 -0.540 0.493
## 5 5 0.569 1.04 -0.221 1.75
## 6 6 0.0961 0.549 -0.528 1.28
## 7 7 0.503 NA 0.503 0.503
## 8 8 0.416 NA 0.416 0.416
## 9 9 2.49 NA 2.49 2.49
## 10 10 16.5 NA 16.5 16.5
## 11 11 1.45 NA 1.45 1.45
## 12 12 -0.110 NA -0.110 -0.110
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Channel_2),
sd = sd(Channel_2),
min = min(Channel_2),
max = max(Channel_2))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.944 0.230 0 1
## 2 2 0.0997 0.300 0 1
## 3 3 0.333 0.577 0 1
## 4 4 0 0 0 0
## 5 5 1 0 1 1
## 6 6 1 0 1 1
## 7 7 1 NA 1 1
## 8 8 0 NA 0 0
## 9 9 0 NA 0 0
## 10 10 0 NA 0 0
## 11 11 0 NA 0 0
## 12 12 1 NA 1 1
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Region_1),
sd = sd(Region_1),
min = min(Region_1),
max = max(Region_1))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.122 0.329 0 1
## 2 2 0.196 0.398 0 1
## 3 3 0 0 0 0
## 4 4 0 0 0 0
## 5 5 0 0 0 0
## 6 6 0.333 0.483 0 1
## 7 7 0 NA 0 0
## 8 8 0 NA 0 0
## 9 9 0 NA 0 0
## 10 10 0 NA 0 0
## 11 11 0 NA 0 0
## 12 12 0 NA 0 0
group_by(dados_completos, cluster_hier) %>%
summarise(
mean = mean(Region_2),
sd = sd(Region_2),
min = min(Region_2),
max = max(Region_2))
## # A tibble: 12 × 5
## cluster_hier mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.144 0.354 0 1
## 2 2 0.0997 0.300 0 1
## 3 3 0 0 0 0
## 4 4 0 0 0 0
## 5 5 0 0 0 0
## 6 6 0.0952 0.301 0 1
## 7 7 0 NA 0 0
## 8 8 0 NA 0 0
## 9 9 0 NA 0 0
## 10 10 0 NA 0 0
## 11 11 1 NA 1 1
## 12 12 1 NA 1 1
Through these statistics, we can understand the characteristics of each cluster, and consequently, the retail network would know how to better allocate its resources to meet the demand of its customers. In our example, we can also observe the presence of outliers because clusters 7, 8, 9, 10, 11, and 12 are formed by a single observation. We could remove the outlier observations and rerun the hierarchical clustering algorithm. However, since our goal is to use the output of the hierarchical method as input for the “k-means” method, we will execute the “k-means” clustering algorithm with only 6 clusters. This way, we will obtain more well-defined clusters.
fviz_nbclust(dados_completos[,1:9], kmeans, method = "wss", k.max = 10)
The elbow method appears to indicate that the optimal number of clusters is indeed around 6 or 7, which aligns with our analysis at the end of the hierarchical procedure. The optimal number is indicated by the point on the X-axis where the distances on the Y-axis between the points start to decrease more significantly.
cluster_kmeans <- kmeans(select(dados_completos, -cluster_hier),
centers = 6)
dados_completos2 <- dados_completos
dados_completos2$cluster_K <- factor(cluster_kmeans$cluster)
summary(anova_channel2 <- aov(formula = Channel_2 ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 61.17 12.234 151.7 <2e-16 ***
## Residuals 434 35.00 0.081
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_region1 <- aov(formula = Region_1 ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 0.48 0.09555 0.658 0.656
## Residuals 434 63.05 0.14527
summary(anova_region2 <- aov(formula = Region_2 ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 0.22 0.04388 0.456 0.809
## Residuals 434 41.76 0.09622
summary(anova_fresh <- aov(formula = Fresh ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 256.9 51.37 122.4 <2e-16 ***
## Residuals 434 182.2 0.42
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_milk <- aov(formula = Milk ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 279.8 55.96 152.5 <2e-16 ***
## Residuals 434 159.2 0.37
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_grocery <- aov(formula = Grocery ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 316.3 63.26 223.8 <2e-16 ***
## Residuals 434 122.7 0.28
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_frozen <- aov(formula = Frozen ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 257.1 51.41 122.6 <2e-16 ***
## Residuals 434 181.9 0.42
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_detergents <- aov(formula = Detergents_Paper ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 345.4 69.07 320.1 <2e-16 ***
## Residuals 434 93.6 0.22
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
summary(anova_delicassen <- aov(formula = Delicassen ~ cluster_K,
data = dados_completos2))
## Df Sum Sq Mean Sq F value Pr(>F)
## cluster_K 5 183.6 36.71 62.37 <2e-16 ***
## Residuals 434 255.4 0.59
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
In the case of the “k-means” procedure, the variables “Region_1” referring to the city of Lisbon and “Region_2” referring to the city of Porto do not contribute to the formation of any cluster. All the other variables are significant for the formation of at least one cluster at a 95% confidence level.
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Fresh),
sd = sd(Fresh),
min = min(Fresh),
max = max(Fresh))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.170 0.659 -0.949 1.50
## 2 2 3.16 3.19 1.14 7.92
## 3 3 0.313 1.14 -0.942 2.57
## 4 4 -0.280 0.483 -0.949 0.890
## 5 5 -0.490 0.455 -0.947 1.00
## 6 6 1.87 0.986 0.497 5.08
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Milk),
sd = sd(Milk),
min = min(Milk),
max = max(Milk))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 -0.127 0.704 -0.740 2.40
## 2 2 3.51 1.56 1.49 5.17
## 3 3 3.92 2.62 -0.111 9.17
## 4 4 -0.416 0.287 -0.778 0.661
## 5 5 0.574 0.637 -0.613 2.72
## 6 6 -0.225 0.429 -0.747 1.01
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Grocery),
sd = sd(Grocery),
min = min(Grocery),
max = max(Grocery))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 -0.401 0.317 -0.765 0.850
## 2 2 1.11 0.380 0.597 1.48
## 3 3 4.27 2.16 1.99 8.93
## 4 4 -0.454 0.294 -0.836 0.898
## 5 5 0.836 0.673 -0.102 3.00
## 6 6 -0.230 0.451 -0.787 1.38
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Frozen),
sd = sd(Frozen),
min = min(Frozen),
max = max(Frozen))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 1.55 1.03 0.332 6.58
## 2 2 5.51 5.03 0.429 11.9
## 3 3 -0.00357 0.554 -0.625 0.970
## 4 4 -0.264 0.319 -0.623 0.772
## 5 5 -0.358 0.234 -0.628 0.529
## 6 6 0.154 0.772 -0.607 3.08
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Detergents_Paper),
sd = sd(Detergents_Paper),
min = min(Detergents_Paper),
max = max(Detergents_Paper))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 -0.494 0.0826 -0.601 -0.280
## 2 2 -0.0383 0.482 -0.554 0.433
## 3 3 4.61 1.73 3.12 7.96
## 4 4 -0.406 0.243 -0.604 0.511
## 5 5 0.860 0.696 -0.545 2.99
## 6 6 -0.396 0.243 -0.602 0.365
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Delicassen),
sd = sd(Delicassen),
min = min(Delicassen),
max = max(Delicassen))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <dbl> <dbl>
## 1 1 0.0305 0.431 -0.524 1.54
## 2 2 6.43 6.88 1.45 16.5
## 3 3 0.503 0.697 -0.221 1.75
## 4 4 -0.220 0.305 -0.540 1.28
## 5 5 0.0375 0.537 -0.540 2.24
## 6 6 0.299 1.02 -0.540 4.59
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Channel_2),
sd = sd(Channel_2),
min = min(Channel_2),
max = max(Channel_2))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.0682 0.255 0 1
## 2 2 0.25 0.5 0 1
## 3 3 1 0 1 1
## 4 4 0.1 0.301 0 1
## 5 5 0.951 0.216 0 1
## 6 6 0.143 0.354 0 1
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Region_1),
sd = sd(Region_1),
min = min(Region_1),
max = max(Region_1))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.182 0.390 0 1
## 2 2 0 0 0 0
## 3 3 0.2 0.422 0 1
## 4 4 0.196 0.398 0 1
## 5 5 0.126 0.334 0 1
## 6 6 0.184 0.391 0 1
group_by(dados_completos2, cluster_K) %>%
summarise(
mean = mean(Region_2),
sd = sd(Region_2),
min = min(Region_2),
max = max(Region_2))
## # A tibble: 6 × 5
## cluster_K mean sd min max
## <fct> <dbl> <dbl> <int> <int>
## 1 1 0.136 0.347 0 1
## 2 2 0.25 0.5 0 1
## 3 3 0.1 0.316 0 1
## 4 4 0.0957 0.295 0 1
## 5 5 0.126 0.334 0 1
## 6 6 0.0816 0.277 0 1
As per our analysis of variance above, we can see that none of the clusters is characterized by the predominance of customers from a specific region. However, it is possible to conclude that retail customers are more grouped in cluster 3. Consequently, we can observe that grocery items and detergent_paper items are more purchased by customers in cluster 3. Obviously, many other inferences could be drawn by analyzing the statistics above, but the example described demonstrates how cluster analyses can improve resource allocation for businesses through customer segmentation into groups.
scatter3D(x=dados_completos2$Channel_2,
y=dados_completos2$Grocery,
z=dados_completos2$Detergents_Paper,
phi = 1, bty = "g", pch = 20, cex = 1,
xlab = "Varejo",
ylab = "Mercearia",
zlab = "Papelaria",
main = "Clientes",
colkey = F)
Plotted from the original data, the above graph indeed demonstrates the existence of a group of customers that “stand out” from the others due to the characteristics we noticed when analyzing the descriptive statistics by cluster. The customers represented with the light blue, yellow, and red colors in the graph above are likely some of the customers that were grouped in cluster 3.
tabela_contingencia <- table(dados_completos2$cluster_hier, dados_completos2$cluster_K)
matriz_confusao <- prop.table(tabela_contingencia, margin = 1)
print(matriz_confusao, digits = 2)
##
## 1 2 3 4 5 6
## 1 0.033 0.000 0.000 0.100 0.867 0.000
## 2 0.133 0.000 0.000 0.734 0.030 0.103
## 3 0.000 0.333 0.000 0.000 0.000 0.667
## 4 0.000 0.000 0.000 0.000 0.000 1.000
## 5 0.000 0.000 1.000 0.000 0.000 0.000
## 6 0.000 0.000 0.238 0.000 0.762 0.000
## 7 0.000 0.000 1.000 0.000 0.000 0.000
## 8 1.000 0.000 0.000 0.000 0.000 0.000
## 9 0.000 1.000 0.000 0.000 0.000 0.000
## 10 0.000 1.000 0.000 0.000 0.000 0.000
## 11 0.000 1.000 0.000 0.000 0.000 0.000
## 12 0.000 0.000 1.000 0.000 0.000 0.000
By completing both procedures and analyzing the confusion matrix, we can see that the two procedures tended to group the observations in a similar manner. The confusion matrix shows us the percentage of observations that were grouped in the same cluster during hierarchical clustering and remained grouped in the same cluster after k-means clustering. It’s noticeable that there was no significant dispersion of observations, and the clusters from the hierarchical procedure that experienced more dispersion of observations still had at least 66% of those observations grouped together again.
Thus, it demonstrates the effectiveness of clustering algorithms for grouping observations and how the approach of using the output of a hierarchical method as input for the k-means method can be a valid strategy to further refine the identified groups, allowing for more precise segmentation and additional insights.