aws-sdk-sagemaker 1.175.0 → 1.176.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: ded8df879b09c28afc104def41071b5cc2e6a71203a6e22f6690b57a64b0090d
4
- data.tar.gz: beee8a7492809d230e990f7a9ffaa19e0660c1581a550c9e414cb47ccdb7d20b
3
+ metadata.gz: b11b07add56f85c6f243d0bb5625b83fb317f8c16b23c9835046f445246624d9
4
+ data.tar.gz: 1b437a3a9d8f8e23e313a2deb12267c5ea702d648cba09d8bea942787449734f
5
5
  SHA512:
6
- metadata.gz: 3f5ce239b868319df2f30e6b1d728b49c5e318926f8b7fddde04ab1daff32a1a8dde545fe31cd87547665717906c16b60dcae549ab8f820e81a37b5d137f05e3
7
- data.tar.gz: e703f72d9a827260db39b0ef1a92feac98f487778d5f6b791d787f3628feb077d86af02724d3b4a3ee97f00e48753fd16e81783bf1e2a924d75abfa644db54ce
6
+ metadata.gz: c4a2d3d9cee2273b34f4c4adce3aad3e2de44b6683ee7bf89273df96865a35ad29d62ebdb808635fcc1705c13d91bc95b56875e5325e3e8d135dd3de345fef38
7
+ data.tar.gz: c91d6b9501a9718817346a1e676582a914d70e4bfb2e6a0ed55b19eadaf141e97806132e871b3405f53987bfd3021da3cfd6cdef7192cfadfb7dabebcd854134
data/CHANGELOG.md CHANGED
@@ -1,6 +1,11 @@
1
1
  Unreleased Changes
2
2
  ------------------
3
3
 
4
+ 1.176.0 (2023-05-02)
5
+ ------------------
6
+
7
+ * Feature - Amazon Sagemaker Autopilot supports training models with sample weights and additional objective metrics.
8
+
4
9
  1.175.0 (2023-04-27)
5
10
  ------------------
6
11
 
data/VERSION CHANGED
@@ -1 +1 @@
1
- 1.175.0
1
+ 1.176.0
@@ -1280,6 +1280,7 @@ module Aws::SageMaker
1280
1280
  # target_attribute_name: "TargetAttributeName", # required
1281
1281
  # content_type: "ContentType",
1282
1282
  # channel_type: "training", # accepts training, validation
1283
+ # sample_weight_attribute_name: "SampleWeightAttributeName",
1283
1284
  # },
1284
1285
  # ],
1285
1286
  # output_data_config: { # required
@@ -9929,6 +9930,7 @@ module Aws::SageMaker
9929
9930
  # resp.input_data_config[0].target_attribute_name #=> String
9930
9931
  # resp.input_data_config[0].content_type #=> String
9931
9932
  # resp.input_data_config[0].channel_type #=> String, one of "training", "validation"
9933
+ # resp.input_data_config[0].sample_weight_attribute_name #=> String
9932
9934
  # resp.output_data_config.kms_key_id #=> String
9933
9935
  # resp.output_data_config.s3_output_path #=> String
9934
9936
  # resp.role_arn #=> String
@@ -23470,7 +23472,7 @@ module Aws::SageMaker
23470
23472
  params: params,
23471
23473
  config: config)
23472
23474
  context[:gem_name] = 'aws-sdk-sagemaker'
23473
- context[:gem_version] = '1.175.0'
23475
+ context[:gem_version] = '1.176.0'
23474
23476
  Seahorse::Client::Request.new(handlers, context)
23475
23477
  end
23476
23478
 
@@ -1648,6 +1648,7 @@ module Aws::SageMaker
1648
1648
  SageMakerImageVersionAlias = Shapes::StringShape.new(name: 'SageMakerImageVersionAlias')
1649
1649
  SageMakerImageVersionAliases = Shapes::ListShape.new(name: 'SageMakerImageVersionAliases')
1650
1650
  SagemakerServicecatalogStatus = Shapes::StringShape.new(name: 'SagemakerServicecatalogStatus')
1651
+ SampleWeightAttributeName = Shapes::StringShape.new(name: 'SampleWeightAttributeName')
1651
1652
  SamplingPercentage = Shapes::IntegerShape.new(name: 'SamplingPercentage')
1652
1653
  ScheduleConfig = Shapes::StructureShape.new(name: 'ScheduleConfig')
1653
1654
  ScheduleExpression = Shapes::StringShape.new(name: 'ScheduleExpression')
@@ -2235,6 +2236,7 @@ module Aws::SageMaker
2235
2236
  AutoMLChannel.add_member(:target_attribute_name, Shapes::ShapeRef.new(shape: TargetAttributeName, required: true, location_name: "TargetAttributeName"))
2236
2237
  AutoMLChannel.add_member(:content_type, Shapes::ShapeRef.new(shape: ContentType, location_name: "ContentType"))
2237
2238
  AutoMLChannel.add_member(:channel_type, Shapes::ShapeRef.new(shape: AutoMLChannelType, location_name: "ChannelType"))
2239
+ AutoMLChannel.add_member(:sample_weight_attribute_name, Shapes::ShapeRef.new(shape: SampleWeightAttributeName, location_name: "SampleWeightAttributeName"))
2238
2240
  AutoMLChannel.struct_class = Types::AutoMLChannel
2239
2241
 
2240
2242
  AutoMLContainerDefinition.add_member(:image, Shapes::ShapeRef.new(shape: ContainerImage, required: true, location_name: "Image"))
@@ -1939,6 +1939,27 @@ module Aws::SageMaker
1939
1939
  # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-datasets-problem-types.html#autopilot-data-sources-training-or-validation
1940
1940
  # @return [String]
1941
1941
  #
1942
+ # @!attribute [rw] sample_weight_attribute_name
1943
+ # If specified, this column name indicates which column of the dataset
1944
+ # should be treated as sample weights for use by the objective metric
1945
+ # during the training, evaluation, and the selection of the best
1946
+ # model. This column is not considered as a predictive feature. For
1947
+ # more information on Autopilot metrics, see [Metrics and
1948
+ # validation][1].
1949
+ #
1950
+ # Sample weights should be numeric, non-negative, with larger values
1951
+ # indicating which rows are more important than others. Data points
1952
+ # that have invalid or no weight value are excluded.
1953
+ #
1954
+ # Support for sample weights is available in [Ensembling][2] mode
1955
+ # only.
1956
+ #
1957
+ #
1958
+ #
1959
+ # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-metrics-validation.html
1960
+ # [2]: https://docs.aws.amazon.com/sagemaker/latest/APIReference/API_AutoMLAlgorithmConfig.html
1961
+ # @return [String]
1962
+ #
1942
1963
  # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/AutoMLChannel AWS API Documentation
1943
1964
  #
1944
1965
  class AutoMLChannel < Struct.new(
@@ -1946,7 +1967,8 @@ module Aws::SageMaker
1946
1967
  :compression_type,
1947
1968
  :target_attribute_name,
1948
1969
  :content_type,
1949
- :channel_type)
1970
+ :channel_type,
1971
+ :sample_weight_attribute_name)
1950
1972
  SENSITIVE = []
1951
1973
  include Aws::Structure
1952
1974
  end
@@ -2217,177 +2239,13 @@ module Aws::SageMaker
2217
2239
  #
2218
2240
  # @!attribute [rw] metric_name
2219
2241
  # The name of the objective metric used to measure the predictive
2220
- # quality of a machine learning system. This metric is optimized
2221
- # during training to provide the best estimate for model parameter
2222
- # values from data.
2223
- #
2224
- # Here are the options:
2225
- #
2226
- # Accuracy
2227
- #
2228
- # : The ratio of the number of correctly classified items to the total
2229
- # number of (correctly and incorrectly) classified items. It is used
2230
- # for both binary and multiclass classification. Accuracy measures
2231
- # how close the predicted class values are to the actual values.
2232
- # Values for accuracy metrics vary between zero (0) and one (1). A
2233
- # value of 1 indicates perfect accuracy, and 0 indicates perfect
2234
- # inaccuracy.
2235
- #
2236
- # AUC
2237
- #
2238
- # : The area under the curve (AUC) metric is used to compare and
2239
- # evaluate binary classification by algorithms that return
2240
- # probabilities, such as logistic regression. To map the
2241
- # probabilities into classifications, these are compared against a
2242
- # threshold value.
2243
- #
2244
- # The relevant curve is the receiver operating characteristic curve
2245
- # (ROC curve). The ROC curve plots the true positive rate (TPR) of
2246
- # predictions (or recall) against the false positive rate (FPR) as a
2247
- # function of the threshold value, above which a prediction is
2248
- # considered positive. Increasing the threshold results in fewer
2249
- # false positives, but more false negatives.
2250
- #
2251
- # AUC is the area under this ROC curve. Therefore, AUC provides an
2252
- # aggregated measure of the model performance across all possible
2253
- # classification thresholds. AUC scores vary between 0 and 1. A
2254
- # score of 1 indicates perfect accuracy, and a score of one half
2255
- # (0.5) indicates that the prediction is not better than a random
2256
- # classifier.
2257
- #
2258
- # BalancedAccuracy
2259
- #
2260
- # : `BalancedAccuracy` is a metric that measures the ratio of accurate
2261
- # predictions to all predictions. This ratio is calculated after
2262
- # normalizing true positives (TP) and true negatives (TN) by the
2263
- # total number of positive (P) and negative (N) values. It is used
2264
- # in both binary and multiclass classification and is defined as
2265
- # follows: 0.5*((TP/P)+(TN/N)), with values ranging from 0 to 1.
2266
- # `BalancedAccuracy` gives a better measure of accuracy when the
2267
- # number of positives or negatives differ greatly from each other in
2268
- # an imbalanced dataset. For example, when only 1% of email is spam.
2269
- #
2270
- # F1
2271
- #
2272
- # : The `F1` score is the harmonic mean of the precision and recall,
2273
- # defined as follows: F1 = 2 * (precision * recall) / (precision +
2274
- # recall). It is used for binary classification into classes
2275
- # traditionally referred to as positive and negative. Predictions
2276
- # are said to be true when they match their actual (correct) class,
2277
- # and false when they do not.
2278
- #
2279
- # Precision is the ratio of the true positive predictions to all
2280
- # positive predictions, and it includes the false positives in a
2281
- # dataset. Precision measures the quality of the prediction when it
2282
- # predicts the positive class.
2283
- #
2284
- # Recall (or sensitivity) is the ratio of the true positive
2285
- # predictions to all actual positive instances. Recall measures how
2286
- # completely a model predicts the actual class members in a dataset.
2287
- #
2288
- # F1 scores vary between 0 and 1. A score of 1 indicates the best
2289
- # possible performance, and 0 indicates the worst.
2290
- #
2291
- # F1macro
2292
- #
2293
- # : The `F1macro` score applies F1 scoring to multiclass
2294
- # classification problems. It does this by calculating the precision
2295
- # and recall, and then taking their harmonic mean to calculate the
2296
- # F1 score for each class. Lastly, the F1macro averages the
2297
- # individual scores to obtain the `F1macro` score. `F1macro` scores
2298
- # vary between 0 and 1. A score of 1 indicates the best possible
2299
- # performance, and 0 indicates the worst.
2300
- #
2301
- # MAE
2302
- #
2303
- # : The mean absolute error (MAE) is a measure of how different the
2304
- # predicted and actual values are, when they're averaged over all
2305
- # values. MAE is commonly used in regression analysis to understand
2306
- # model prediction error. If there is linear regression, MAE
2307
- # represents the average distance from a predicted line to the
2308
- # actual value. MAE is defined as the sum of absolute errors divided
2309
- # by the number of observations. Values range from 0 to infinity,
2310
- # with smaller numbers indicating a better model fit to the data.
2311
- #
2312
- # MSE
2313
- #
2314
- # : The mean squared error (MSE) is the average of the squared
2315
- # differences between the predicted and actual values. It is used
2316
- # for regression. MSE values are always positive. The better a model
2317
- # is at predicting the actual values, the smaller the MSE value is
2318
- #
2319
- # Precision
2320
- #
2321
- # : Precision measures how well an algorithm predicts the true
2322
- # positives (TP) out of all of the positives that it identifies. It
2323
- # is defined as follows: Precision = TP/(TP+FP), with values ranging
2324
- # from zero (0) to one (1), and is used in binary classification.
2325
- # Precision is an important metric when the cost of a false positive
2326
- # is high. For example, the cost of a false positive is very high if
2327
- # an airplane safety system is falsely deemed safe to fly. A false
2328
- # positive (FP) reflects a positive prediction that is actually
2329
- # negative in the data.
2330
- #
2331
- # PrecisionMacro
2332
- #
2333
- # : The precision macro computes precision for multiclass
2334
- # classification problems. It does this by calculating precision for
2335
- # each class and averaging scores to obtain precision for several
2336
- # classes. `PrecisionMacro` scores range from zero (0) to one (1).
2337
- # Higher scores reflect the model's ability to predict true
2338
- # positives (TP) out of all of the positives that it identifies,
2339
- # averaged across multiple classes.
2340
- #
2341
- # R2
2342
- #
2343
- # : R2, also known as the coefficient of determination, is used in
2344
- # regression to quantify how much a model can explain the variance
2345
- # of a dependent variable. Values range from one (1) to negative one
2346
- # (-1). Higher numbers indicate a higher fraction of explained
2347
- # variability. `R2` values close to zero (0) indicate that very
2348
- # little of the dependent variable can be explained by the model.
2349
- # Negative values indicate a poor fit and that the model is
2350
- # outperformed by a constant function. For linear regression, this
2351
- # is a horizontal line.
2352
- #
2353
- # Recall
2354
- #
2355
- # : Recall measures how well an algorithm correctly predicts all of
2356
- # the true positives (TP) in a dataset. A true positive is a
2357
- # positive prediction that is also an actual positive value in the
2358
- # data. Recall is defined as follows: Recall = TP/(TP+FN), with
2359
- # values ranging from 0 to 1. Higher scores reflect a better ability
2360
- # of the model to predict true positives (TP) in the data, and is
2361
- # used in binary classification.
2362
- #
2363
- # Recall is important when testing for cancer because it's used to
2364
- # find all of the true positives. A false positive (FP) reflects a
2365
- # positive prediction that is actually negative in the data. It is
2366
- # often insufficient to measure only recall, because predicting
2367
- # every output as a true positive yield a perfect recall score.
2368
- #
2369
- # RecallMacro
2370
- #
2371
- # : The RecallMacro computes recall for multiclass classification
2372
- # problems by calculating recall for each class and averaging scores
2373
- # to obtain recall for several classes. RecallMacro scores range
2374
- # from 0 to 1. Higher scores reflect the model's ability to predict
2375
- # true positives (TP) in a dataset. Whereas, a true positive
2376
- # reflects a positive prediction that is also an actual positive
2377
- # value in the data. It is often insufficient to measure only
2378
- # recall, because predicting every output as a true positive yields
2379
- # a perfect recall score.
2380
- #
2381
- # RMSE
2382
- #
2383
- # : Root mean squared error (RMSE) measures the square root of the
2384
- # squared difference between predicted and actual values, and it's
2385
- # averaged over all values. It is used in regression analysis to
2386
- # understand model prediction error. It's an important metric to
2387
- # indicate the presence of large model errors and outliers. Values
2388
- # range from zero (0) to infinity, with smaller numbers indicating a
2389
- # better model fit to the data. RMSE is dependent on scale, and
2390
- # should not be used to compare datasets of different sizes.
2242
+ # quality of a machine learning system. During training, the model's
2243
+ # parameters are updated iteratively to optimize its performance based
2244
+ # on the feedback provided by the objective metric when evaluating the
2245
+ # model on the validation dataset.
2246
+ #
2247
+ # For the list of all available metrics supported by Autopilot, see
2248
+ # [Autopilot metrics][1].
2391
2249
  #
2392
2250
  # If you do not specify a metric explicitly, the default behavior is
2393
2251
  # to automatically use:
@@ -2397,6 +2255,10 @@ module Aws::SageMaker
2397
2255
  # * `F1`: for binary classification
2398
2256
  #
2399
2257
  # * `Accuracy`: for multiclass classification.
2258
+ #
2259
+ #
2260
+ #
2261
+ # [1]: https://docs.aws.amazon.com/sagemaker/latest/dg/autopilot-metrics-validation.html#autopilot-metrics
2400
2262
  # @return [String]
2401
2263
  #
2402
2264
  # @see http://docs.aws.amazon.com/goto/WebAPI/sagemaker-2017-07-24/AutoMLJobObjective AWS API Documentation
@@ -53,6 +53,6 @@ require_relative 'aws-sdk-sagemaker/customizations'
53
53
  # @!group service
54
54
  module Aws::SageMaker
55
55
 
56
- GEM_VERSION = '1.175.0'
56
+ GEM_VERSION = '1.176.0'
57
57
 
58
58
  end
metadata CHANGED
@@ -1,14 +1,14 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: aws-sdk-sagemaker
3
3
  version: !ruby/object:Gem::Version
4
- version: 1.175.0
4
+ version: 1.176.0
5
5
  platform: ruby
6
6
  authors:
7
7
  - Amazon Web Services
8
8
  autorequire:
9
9
  bindir: bin
10
10
  cert_chain: []
11
- date: 2023-04-27 00:00:00.000000000 Z
11
+ date: 2023-05-02 00:00:00.000000000 Z
12
12
  dependencies:
13
13
  - !ruby/object:Gem::Dependency
14
14
  name: aws-sdk-core