h2oai / h2o-3

H2O is an Open Source, Distributed, Fast & Scalable Machine Learning Platform: Deep Learning, Gradient Boosting (GBM) & XGBoost, Random Forest, Generalized Linear Modeling (GLM with Elastic Net), K-Means, PCA, Generalized Additive Models (GAM), RuleFit, Support Vector Machine (SVM), Stacked Ensembles, Automatic Machine Learning (AutoML), etc.
http://h2o.ai
Apache License 2.0
6.87k stars 2k forks source link

GLM : Family = Gaussian, and priors = 0.5 => doesn't error out that Priors works only for Binomial #15097

Closed exalate-issue-sync[bot] closed 1 year ago

exalate-issue-sync[bot] commented 1 year ago

TestNG testcase : glm_neg_testcase_144

Priors = 0.5 and family = gaussian

Doesn't error out that Priors works only for Binomial

Validate Parameters object with testcase: glm_neg_testcase_144 Create modelParameter object with testcase: glm_neg_testcase_144 Set _family: gaussian Set _solver: L_BFGS set auto_set params Set _prior: 0.5 Create train frame: airquality_train1 Create frame with airquality_train1.csv dataSetId: airquality_train1 dataSetDirectory: smalldata fileName: airquality_train1.csv responseColumn: Ozone columnNames: Ozone;Solar.R;Wind;Temp;Month;Day; columnTypes: numeric;numeric;numeric;numeric;numeric;numeric; 10-05 15:02:44.426 172.16.2.201:54321 5742 #t worker INFO: Locking cloud to new members, because water.fvec.NFSFileVec 10-05 15:02:44.542 172.16.2.201:54321 5742 #t worker INFO: Total file size: 1.4 KB 10-05 15:02:44.553 172.16.2.201:54321 5742 #t worker INFO: Parse chunk size 4194304 10-05 15:02:44.776 172.16.2.201:54321 5742 FJ-0-15 INFO: Parse result for airquality_train1.hex (77 rows): 10-05 15:02:44.812 172.16.2.201:54321 5742 FJ-0-15 INFO: ColV2 type min max NAs constant cardinality 10-05 15:02:44.813 172.16.2.201:54321 5742 FJ-0-15 INFO: Ozone: numeric 1.00000 135.000
10-05 15:02:44.813 172.16.2.201:54321 5742 FJ-0-15 INFO: Solar.R: numeric 7.00000 334.000 5
10-05 15:02:44.814 172.16.2.201:54321 5742 FJ-0-15 INFO: Wind: numeric 4.00000 20.7000
10-05 15:02:44.815 172.16.2.201:54321 5742 FJ-0-15 INFO: Temp: numeric 57.0000 92.0000
10-05 15:02:44.816 172.16.2.201:54321 5742 FJ-0-15 INFO: Month: numeric 5.00000 8.00000
10-05 15:02:44.816 172.16.2.201:54321 5742 FJ-0-15 INFO: Day: numeric 1.00000 31.0000
10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk compression summary: 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk Type Chunk Name Count Count Percentage Size Size Percentage 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: C1N 1-Byte Integers (w/o NAs) 4 66.667 % 580 B 60.228 % 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: C1S 1-Byte Fractions 1 16.667 % 161 B 16.719 % 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: C2 2-Byte Integers 1 16.667 % 222 B 23.053 % 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: Frame distribution summary: 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: Size Number of Rows Number of Chunks per Column Number of Chunks 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: 172.16.2.201:54321 963 B 77.000000 1.000000 6.000000 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: mean 963 B 77.000000 1.000000 6.000000 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: min 963 B 77.000000 1.000000 6.000000 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: max 963 B 77.000000 1.000000 6.000000 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: stddev 0 B 0.000000 0.000000 0.000000 10-05 15:02:44.831 172.16.2.201:54321 5742 FJ-0-15 INFO: total 963 B 77.000000 1.000000 6.000000 Create validate frame: airquality_train1 Create frame with airquality_validation1.csv dataSetId: airquality_validation1 dataSetDirectory: smalldata fileName: airquality_validation1.csv responseColumn: Ozone columnNames: Ozone;Solar.R;Wind;Temp;Month;Day; columnTypes: numeric;numeric;numeric;numeric;numeric;numeric; 10-05 15:02:44.835 172.16.2.201:54321 5742 #t worker INFO: Total file size: 783 B 10-05 15:02:44.836 172.16.2.201:54321 5742 #t worker INFO: Parse chunk size 4194304 10-05 15:02:44.846 172.16.2.201:54321 5742 FJ-0-15 INFO: Parse result for airquality_validation1.hex (39 rows): 10-05 15:02:44.850 172.16.2.201:54321 5742 FJ-0-15 INFO: ColV2 type min max NAs constant cardinality 10-05 15:02:44.851 172.16.2.201:54321 5742 FJ-0-15 INFO: Ozone: numeric 7.00000 168.000
10-05 15:02:44.851 172.16.2.201:54321 5742 FJ-0-15 INFO: Solar.R: numeric 14.0000 259.000
10-05 15:02:44.852 172.16.2.201:54321 5742 FJ-0-15 INFO: Wind: numeric 2.30000 16.6000
10-05 15:02:44.852 172.16.2.201:54321 5742 FJ-0-15 INFO: Temp: numeric 63.0000 97.0000
10-05 15:02:44.853 172.16.2.201:54321 5742 FJ-0-15 INFO: Month: numeric 8.00000 9.00000
10-05 15:02:44.854 172.16.2.201:54321 5742 FJ-0-15 INFO: Day: numeric 1.00000 31.0000
10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk compression summary: 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk Type Chunk Name Count Count Percentage Size Size Percentage 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: C1N 1-Byte Integers (w/o NAs) 4 66.667 % 428 B 63.501 % 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: C1S 1-Byte Fractions 2 33.333 % 246 B 36.499 % 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: Frame distribution summary: 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: Size Number of Rows Number of Chunks per Column Number of Chunks 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: 172.16.2.201:54321 674 B 39.000000 1.000000 6.000000 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: mean 674 B 39.000000 1.000000 6.000000 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: min 674 B 39.000000 1.000000 6.000000 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: max 674 B 39.000000 1.000000 6.000000 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: stddev 0 B 0.000000 0.000000 0.000000 10-05 15:02:44.859 172.16.2.201:54321 5742 FJ-0-15 INFO: total 674 B 39.000000 1.000000 6.000000 10-05 15:02:44.883 172.16.2.201:54321 5742 #t worker INFO: Total file size: 92 B 10-05 15:02:44.906 172.16.2.201:54321 5742 FJ-0-15 INFO: Parse result for beta_constraints.hex (5 rows): 10-05 15:02:44.908 172.16.2.201:54321 5742 FJ-0-15 INFO: ColV2 type min max NAs constant cardinality 10-05 15:02:44.908 172.16.2.201:54321 5742 FJ-0-15 INFO: names: string
10-05 15:02:44.909 172.16.2.201:54321 5742 FJ-0-15 INFO: lower_bounds: numeric 100.000 100.000 constant
10-05 15:02:44.910 172.16.2.201:54321 5742 FJ-0-15 INFO: upper_bounds: numeric 0.00000 0.00000 constant
10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk compression summary: 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: Chunk Type Chunk Name Count Count Percentage Size Size Percentage 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: C0L Constant Integers 2 66.667 % 160 B 56.939 % 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: CStr String 1 33.333 % 121 B 43.061 % 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: Frame distribution summary: 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: Size Number of Rows Number of Chunks per Column Number of Chunks 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: 172.16.2.201:54321 281 B 5.000000 1.000000 3.000000 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: mean 281 B 5.000000 1.000000 3.000000 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: min 281 B 5.000000 1.000000 3.000000 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: max 281 B 5.000000 1.000000 3.000000 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: stddev 0 B 0.000000 0.000000 0.000000 10-05 15:02:44.915 172.16.2.201:54321 5742 FJ-0-15 INFO: total 281 B 5.000000 1.000000 3.000000 Set train frame Set validate frame Create success modelParameter object. Build model Train model 10-05 15:02:45.044 172.16.2.201:54321 5742 FJ-0-15 INFO: Building H2O GLM model with these parameters: 10-05 15:02:45.048 172.16.2.201:54321 5742 FJ-0-15 INFO: {"_model_id":null,"_train":{"name":"airquality_train1.hex","type":"Key"},"_valid":{"name":"airquality_validation1.hex","type":"Key"},"_nfolds":0,"_keep_cross_validation_predictions":false,"_fold_assignment":"AUTO","_distribution":"AUTO","_tweedie_power":1.5,"_ignored_columns":null,"_ignore_const_cols":true,"_weights_column":null,"_offset_column":null,"_fold_column":null,"_score_each_iteration":false,"_response_column":"Ozone","_balance_classes":false,"_max_after_balance_size":5.0,"_class_sampling_factors":null,"_max_hit_ratio_k":10,"_max_confusion_matrix_size":20,"_checkpoint":null,"_standardize":true,"_family":"gaussian","_link":"family_default","_solver":"L_BFGS","_tweedie_variance_power":0.0,"_tweedie_link_power":1.0,"_alpha":null,"_lambda":null,"_prior":0.5,"_lambda_search":false,"_nlambdas":100,"_non_negative":false,"_exactLambdas":false,"_lambda_min_ratio":-1.0,"_use_all_factor_levels":false,"_max_iterations":-1,"_intercept":true,"_beta_epsilon":1.0E-5,"_objective_epsilon":1.0E-5,"_gradient_epsilon":1.0E-4,"_beta_constraints":{"name":"beta_constraints.hex","type":"Key"},"_max_active_predictors":-1} 10-05 15:02:45.943 172.16.2.201:54321 5742 FJ-0-15 INFO: GLM[dest=model, iteration=0, lambda = 2199.5]: All 5 coefficients are active 10-05 15:02:46.070 172.16.2.201:54321 5742 FJ-0-15 WARN: ADMM solver reached maximum number of iterations (500) 10-05 15:02:46.070 172.16.2.201:54321 5742 FJ-0-15 WARN: ADMM solver finished with gerr = 17.80996283234473 > eps = 0.01 10-05 15:02:46.075 172.16.2.201:54321 5742 FJ-0-9 INFO: GLM[dest=model, iteration=516, lambda = 2199.5]: hold-out set validation = mse = NaN, explained_dev = 0.0 10-05 15:02:46.289 172.16.2.201:54321 5742 FJ-0-9 INFO: Solution at lambda = 2199.4816673930377 has 0 nonzeros, gradient err = 0.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Model Metrics Type: RegressionGLM 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Description: N/A 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: model id: model 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: frame id: airquality_train1.hex 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: MSE: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: R^2: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: mean residual deviance: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: null DOF: 71.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: residual DOF: 71.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: null deviance: 198494.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: residual deviance: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: AIC: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Model Metrics Type: RegressionGLM 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Description: N/A 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: model id: model 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: frame id: airquality_validation1.hex 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: MSE: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: R^2: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: mean residual deviance: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: null DOF: 38.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: residual DOF: 38.0 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: null deviance: 115391.75 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: residual deviance: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: AIC: NaN 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: GLM Model (summary): 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Family Link Regularization Number of Predictors Total Number of Active Predictors Number of Iterations Training Frame 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: gaussian identity Ridge ( lambda = 2199.5 ) 5 1 516 airquality_train1.hex 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: Scoring History: 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: timestamp duration iteration log_likelihood objective 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.000 sec 0 36663.63889 509.21721 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.837 sec 1 36624.07281 508.96291 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.842 sec 2 36626.99703 508.96148 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.847 sec 3 36626.99703 508.96148 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.854 sec 4 36626.99703 508.96148 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.858 sec 5 36627.17587 508.96150 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.862 sec 6 36627.16264 508.96150 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.866 sec 7 36627.16264 508.96150 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.869 sec 8 36627.16264 508.96150 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:45 0.873 sec 9 36627.16264 508.96150 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: --- 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.944 sec 507 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.944 sec 508 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.944 sec 509 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.944 sec 510 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.945 sec 511 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.945 sec 512 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.945 sec 513 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.945 sec 514 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.945 sec 515 36627.49162 508.96156 10-05 15:02:46.299 172.16.2.201:54321 5742 FJ-0-9 INFO: 2015-10-05 15:02:46 0.950 sec 516 NaN NaN

DinukaH2O commented 1 year ago

JIRA Issue Migration Info

Jira Issue: PUBDEV-2186 Assignee: Tomas Nykodym Reporter: Neeraja Madabhushi State: Resolved Fix Version: N/A Attachments: N/A Development PRs: N/A