tesselle: kairos 1.1.0

Nicolas Frerebeau

kairos 1.1.0 is now on CRAN! kairos is a toolkit for the analysis of chronological patterns from archaeological count data.

You can install it from CRAN with:

install.packages("kairos")

The range of tools provided by kairos is now extended by two methods for matrix seriation¹, allowing the construction of relative chronologies:

Reciprocal ranking: seriate_rank()
Correspondence analysis: seriate_average()

The seriate_*() methods return a permutation of the rows and columns (according to the margin argument) of a count data matrix or data.frame. Then, permute() performs the reordering.

This post briefly illustrates how to perform a correspondence analysis based seriation.

library(kairos)

Correspondence Analysis (CA) is an effective method for the seriation of archaeological assemblages. The order of the rows and columns is given by the coordinates along one dimension of the CA space, assumed to account for temporal variation. The direction of temporal change within the correspondence analysis space is arbitrary: additional information is needed to determine the actual order in time.

## Load the zuni dataset
# install.packages("folio")
data("zuni", package = "folio")

## Plot the original matrix
tabula::plot_ford(zuni) +
  ggplot2::theme(axis.text.y = ggplot2::element_blank())

## Get row and column permutations from CA coordinates along the first axis
indices <- seriate_average(zuni, margin = c(1, 2), axes = 1)
get_order(indices)

$rows
  [1] 372 387 350 367 110 417 364 407 357 160 373 406 356 344 348 384
 [17] 362 378 339 383 376 386 338 419 416 408 377 392 347 371 374 351
 [33] 353 402 366 375 397 363 109 395 400 389 420 155 381 412 178 382
 [49] 415 368 404 172 385 361 126 352 354 360 358 343 182 414 342  86
 [65]  94 403  93 410 413 396 379 388 401 108 345 176 341 315 399 113
 [81] 314 369  98 321 112 179 323  87 411 365 316 359 197 115 370 253
 [97] 104 289 320 174 390 327 409 332 317 391 103 355  37 418 322 318
[113] 398 194 329 335 334 325 331 326 319 394 328 156  47 100 147 173
[129] 349 290 333 340  99 153  32  88 330 157 142  38 223  17  27 405
[145] 336  26 242 116 256 159 184 158 163 255 121  83 162  96 324 148
[161] 198 228 232  97 243  40 250 150 165  31 154  35 169   1 171  54
[177]  53 229   3 263 224 247 175 164 137 246 240  39 225 212  84 238
[193] 299 230  80   9  46   2  85 257   7 234 278 233 199 111 183  29
[209] 141  57 196  92  42 166  28  22 101 380 279 283  24  45 151  68
[225]  89  41   8 144 209  61 226 239  49  52 145 248 119  82 149   5
[241]  43  48 252 210 227 245  75 152 193  77 102  74 277 237 211  73
[257] 241 222  66 249 107 292 273 130 217 208 118 117  58 393  95 251
[273] 190  65 213 244 143 204  67  64  44 132 301 136 270 177 114   4
[289]  18 287 139  91  51 195 161 236 309 131 189 201 268 180 205 135
[305]  79 187  50 146 168 266 259  62  76  81 106 235 294 214 207 105
[321] 303 254  90 185 203  15 202 281 286  14 124 272 282 284  33 140
[337] 291 271 128 127 129 167 313 170 337 285 133 298 261 125  55 346
[353]  12  63 262 295 191 310  13  70 264 219 134  71 138 258 188 216
[369] 192 312 300 296 260  72 269  36  69 288  23   6 267 280  34 302
[385]  30 181 308 123 122 304 311 186  19 218 306  10 307 200 215 120
[401] 274 305 293 206 221 276  56  21  25  78 220 231 275 297  20  59
[417]  60  16 265  11

$columns
 [1] 18 14 17 16 13 15  9  8 12 11  6  7  5 10  4  2  3  1

seriate_average() also contains the full results of the correspondence analysis. You can use dimensio to explore these results:

## Plot CA row coordinates (get a nice arch effect!)
dimensio::plot_rows(indices) +
  ggplot2::theme_bw() +
  ggplot2::theme(legend.position = "none")

## Plot column contributions to the first two axes
dimensio::plot_contributions(indices, margin = 2, axes = c(1, 2)) +
  ggplot2::theme_bw()

If the results of the correspondence analysis are satisfactory, we can then permute the rows and columns of the initial data matrix:

## Permute rows and columns
permuted <- permute(zuni, indices)

## Plot permuted matrix
tabula::plot_ford(permuted) +
  ggplot2::theme(axis.text.y = ggplot2::element_blank())

The move to add these matrix seriation methods, previously implemented in tabula, follows a reorganization for consistency so that we no longer have chronology tools in two separate packages. This implies that the seriation methods are temporarily re-exported by tabula from kairos and will be removed in a future version of tabula.↩︎

kairos 1.1.0

References

Corrections

Reuse

Citation