outliers
DecompositionAndOutlierGenerator
Bases: OutliersGenerator
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
__init__(decomposition_transformer=PCA(n_components=2), outlier_generator=ZScoreSamplingGenerator(default_rng(0), n_outliers=10))
:param decomposition_transformer: The dimensionality reduction transformer to be used before the outlier transformer :param outlier_generator: The outlier transformer to be used after the dimensionality has been reduced
Source code in badgers/generators/tabular_data/outliers.py
366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 |
|
generate(X, y=None, **params)
Randomly generate outliers by first applying a dimensionality reduction technique (sklearn.decomposition) and an outlier transformer.
- Standardize the input data (mean = 0, variance = 1)
- Apply the dimensionality reduction transformer
- Generates outliers by applying the outlier transformer
- Inverse the dimensionality reduction and the standardization transformations
:param X: the input features :param y: the regression target, class labels, or None :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
HistogramSamplingGenerator
Bases: OutliersGenerator
Randomly generates outliers from low density regions. Low density regions are estimated through a histogram.
WARNING: This computes a full histogram in d-dimensions (d = nb features / columns), which is O(d²). Should only be used with low dimensionality data! It will raise an error if the number of dimensions is greater than 5.
TODO: this works but is very inefficient, better strategies are welcome!
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
__init__(random_generator=default_rng(seed=0), n_outliers=10, threshold_low_density=0.1, bins=10)
:param random_generator: A random generator :param n_outliers: The number of outliers to generate :param threshold_low_density: the threshold that defines a low density region (must be between 0 and 1) :param bins: number of bins for the histogram
Source code in badgers/generators/tabular_data/outliers.py
219 220 221 222 223 224 225 226 227 228 229 230 231 |
|
generate(X, y=None, **params)
Randomly generates outliers from low density regions. Low density regions are estimated through histograms
- Standardize the input data (mean = 0, variance = 1)
- Compute and normalize histogram for the data
- Sample datapoint uniformly at random within bins of low density
- Inverse the standardization transformation
:param X: the input features :param y: not used :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
HypersphereSamplingGenerator
Bases: OutliersGenerator
Generates outliers by sampling points from a hypersphere with radius at least 3 sigma
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
__init__(random_generator=default_rng(seed=0), n_outliers=10)
:param random_generator: A random generator :param n_outliers: The number of outliers to generate
Source code in badgers/generators/tabular_data/outliers.py
94 95 96 97 98 99 100 |
|
generate(X, y=None, **params)
Randomly generates outliers as data points with a z-score > 3.
- Standardize the input data (mean = 0, variance = 1)
- Generate outliers on a hypersphere (see https://en.wikipedia.org/wiki/N-sphere#Spherical_coordinates):
- angles are chosen uniformly at random
- radius is = 3 + a random number following an exponential distribution function with default parameters (see https://numpy.org/doc/stable/reference/random/generated/numpy.random.Generator.exponential.html)
- Inverse the standardization transformation
:param X: the input features :param y: not used :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
IndependentHistogramsGenerator
Bases: OutliersGenerator
Randomly generates outliers from low density regions. Low density regions are estimated through several independent histograms (one for each feature).
For each feature (column), a histogram is computed (it approximates the marginal distribution). Values are generated from bins with a low number of data points.
All values generated for each feature are simply concatenated (independence hypothesis!).
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
generate(X, y=None, **params)
Randomly generates outliers from low density regions. Low density regions are estimated through several independent histograms (one for each feature).
For each feature (column), a histogram is computed (it approximates the marginal distribution). Values are generated from bins with a low number of data points.
All values generated for each feature are simply concatenated (independence hypothesis!).
:param X: the input features :param y: not used :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
LowDensitySamplingGenerator
Bases: OutliersGenerator
Randomly generates outliers from low density regions. Low density regions are estimated using a KernelDensity estimator. Points are sampled uniformly at random and filtered out if they do not belong to a low density region
TODO: this works but might not be efficient, a better sampling strategy is welcome
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
__init__(random_generator=default_rng(seed=0), n_outliers=10, threshold_low_density=0.1)
:param random_generator: A random generator :param n_outliers: The number of outliers to generate :param threshold_low_density: the threshold that defines a low density region (must be between 0 and 1)
Source code in badgers/generators/tabular_data/outliers.py
293 294 295 296 297 298 299 300 301 302 |
|
generate(X, y=None, **params)
Generate data points belonging to low density regions.
Pseudo code: - Standardize the data X - Estimate the density based upon the original data X - Computes a threshold for determining low density (so far 10th percentile) - Sample uniformly at random within the hypercube [min, max] - Estimate the density of the new points and filter out the ones with a density that is above the threshold
:param X: the input features :param y: not used :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
OutliersGenerator
Bases: GeneratorMixin
Base class for transformers that add outliers to tabular data
Source code in badgers/generators/tabular_data/outliers.py
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 |
|
__init__(random_generator=default_rng(seed=0), n_outliers=10)
:param random_generator: A random generator :param n_outliers: The number of outliers to generate
Source code in badgers/generators/tabular_data/outliers.py
22 23 24 25 26 27 28 |
|
ZScoreSamplingGenerator
Bases: OutliersGenerator
Randomly generates outliers as data points with a z-score > 3.
Source code in badgers/generators/tabular_data/outliers.py
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 |
|
__init__(random_generator=default_rng(seed=0), n_outliers=10)
:param random_generator: A random generator :param n_outliers: The number of outliers to generate
Source code in badgers/generators/tabular_data/outliers.py
40 41 42 43 44 45 46 |
|
generate(X, y=None, **params)
Randomly generates outliers as data points with a z-score > 3.
- Standardize the input data (mean = 0, variance = 1)
- Generate outliers as follows:
- the sign is randomly chosen
- for each dimension: the value is equal to 3 + a random number following an exponential distribution function with default parameters (see https://numpy.org/doc/stable/reference/random/generated/numpy.random.Generator.exponential.html)
- Inverse the standardization transformation
:param X: the input features :param y: not used :param params: :return:
Source code in badgers/generators/tabular_data/outliers.py
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 |
|