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
6.89k stars 2k forks source link

dl with offset: when offset same as response, do not get 0 mse #14528

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

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

y =seq(1:14) x= c(0,1,2,3,1,4,9,18,23,31,20,25,37,45) offset =seq(1:14) df = data.frame(y,x,offset) hdf = as.h2o(df,destination_frame = "hdd")

offset as predictor expect mse~0

h1 = h2o.deeplearning(x = 2:3,y = "y",training_frame = hdf,epochs = 100000,hidden = 14,activation = "Tanh") h1@model$training_metrics@metrics$MSE [1] 1.473114e-05

offset as offset expect mse~0

> h2 = h2o.deeplearning(x = 2,y = "y",training_frame = hdf,epochs = #100000,hidden = 14,activation = "Tanh",offset_column = "offset")

Error in args$x_ignore : object of type 'closure' is not subsettable

buildModel 'deeplearning', {"model_id":"deeplearning-9827f7d8-2236-4ed8-83e8-04fd35af1b93","training_frame":"hdf","ignored_columns":[],"ignore_const_cols":true,"response_column":"y","offset_column":"offset","activation":"Tanh","hidden":[14],"epochs":"10000","variable_importances":false,"checkpoint":"","use_all_factor_levels":true,"train_samples_per_iteration":-2,"adaptive_rate":true,"input_dropout_ratio":0,"l1":0,"l2":0,"loss":"Automatic","score_interval":5,"score_training_samples":10000,"score_duty_cycle":0.1,"autoencoder":false,"overwrite_with_best_model":true,"target_ratio_comm_to_comp":0.02,"seed":-7311643238310494000,"rho":0.99,"epsilon":1e-8,"max_w2":"Infinity","initial_weight_distribution":"UniformAdaptive","regression_stop":0.000001,"diagnostics":true,"fast_mode":true,"force_load_balance":true,"single_node_mode":false,"shuffle_training_data":false,"missing_values_handling":"MeanImputation","quiet_mode":false,"sparse":false,"col_major":false,"average_activation":0,"sparsity_beta":0,"max_categorical_features":2147483647,"reproducible":false,"export_weights_and_biases":false}

get MSE 175.289265 screenshot attached

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

Arno Candel commented: Now get 3.35916e-07 with 10,000 epochs.

DinukaH2O commented 1 year ago

JIRA Issue Migration Info

Jira Issue: PUBDEV-1554 Assignee: Arno Candel Reporter: Nidhi Mehta State: Resolved Fix Version: N/A Attachments: Available (Count: 1) Development PRs: N/A

Attachments From Jira

Attachment Name: dl_offset.png Attached By: Nidhi Mehta File Link:https://h2o-3-jira-github-migration.s3.amazonaws.com/PUBDEV-1554/dl_offset.png