Goal is to automate building and tuning a classification model to predict employee attrition, using the h2o::h2o.automl.
Import the cleaned data from Module 7.
library(h2o)## Warning: package 'h2o' was built under R version 4.3.3##
## ----------------------------------------------------------------------
##
## Your next step is to start H2O:
## > h2o.init()
##
## For H2O package documentation, ask for help:
## > ??h2o
##
## After starting H2O, you can use the Web UI at http://localhost:54321
## For more information visit https://docs.h2o.ai
##
## ----------------------------------------------------------------------##
## Attaching package: 'h2o'## The following objects are masked from 'package:stats':
##
## cor, sd, var## The following objects are masked from 'package:base':
##
## %*%, %in%, &&, ||, apply, as.factor, as.numeric, colnames,
## colnames<-, ifelse, is.character, is.factor, is.numeric, log,
## log10, log1p, log2, round, signif, trunclibrary(tidyverse)## Warning: package 'ggplot2' was built under R version 4.3.2## ── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
## ✔ dplyr 1.1.3 ✔ readr 2.1.4
## ✔ forcats 1.0.0 ✔ stringr 1.5.0
## ✔ ggplot2 3.5.0 ✔ tibble 3.2.1
## ✔ lubridate 1.9.2 ✔ tidyr 1.3.0
## ✔ purrr 1.0.2## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ lubridate::day() masks h2o::day()
## ✖ dplyr::filter() masks stats::filter()
## ✖ lubridate::hour() masks h2o::hour()
## ✖ dplyr::lag() masks stats::lag()
## ✖ lubridate::month() masks h2o::month()
## ✖ lubridate::week() masks h2o::week()
## ✖ lubridate::year() masks h2o::year()
## ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errorslibrary(tidymodels)## Warning: package 'tidymodels' was built under R version 4.3.2## ── Attaching packages ────────────────────────────────────── tidymodels 1.1.1 ──
## ✔ broom 1.0.5 ✔ rsample 1.2.0
## ✔ dials 1.2.0 ✔ tune 1.1.2
## ✔ infer 1.0.6 ✔ workflows 1.1.3
## ✔ modeldata 1.3.0 ✔ workflowsets 1.0.1
## ✔ parsnip 1.1.1 ✔ yardstick 1.3.0
## ✔ recipes 1.0.8## Warning: package 'dials' was built under R version 4.3.2## Warning: package 'scales' was built under R version 4.3.2## Warning: package 'infer' was built under R version 4.3.2## Warning: package 'modeldata' was built under R version 4.3.2## Warning: package 'parsnip' was built under R version 4.3.2## Warning: package 'tune' was built under R version 4.3.2## Warning: package 'workflows' was built under R version 4.3.2## Warning: package 'workflowsets' was built under R version 4.3.2## Warning: package 'yardstick' was built under R version 4.3.2## ── Conflicts ───────────────────────────────────────── tidymodels_conflicts() ──
## ✖ scales::discard() masks purrr::discard()
## ✖ dplyr::filter() masks stats::filter()
## ✖ recipes::fixed() masks stringr::fixed()
## ✖ dplyr::lag() masks stats::lag()
## ✖ yardstick::spec() masks readr::spec()
## ✖ recipes::step() masks stats::step()
## • Learn how to get started at https://www.tidymodels.org/start/library(tidyquant)## Loading required package: PerformanceAnalytics
## Loading required package: xts
## Loading required package: zoo
##
## Attaching package: 'zoo'
##
## The following objects are masked from 'package:base':
##
## as.Date, as.Date.numeric
##
##
## ######################### Warning from 'xts' package ##########################
## # #
## # The dplyr lag() function breaks how base R's lag() function is supposed to #
## # work, which breaks lag(my_xts). Calls to lag(my_xts) that you type or #
## # source() into this session won't work correctly. #
## # #
## # Use stats::lag() to make sure you're not using dplyr::lag(), or you can add #
## # conflictRules('dplyr', exclude = 'lag') to your .Rprofile to stop #
## # dplyr from breaking base R's lag() function. #
## # #
## # Code in packages is not affected. It's protected by R's namespace mechanism #
## # Set `options(xts.warn_dplyr_breaks_lag = FALSE)` to suppress this warning. #
## # #
## ###############################################################################
##
## Attaching package: 'xts'
##
## The following objects are masked from 'package:dplyr':
##
## first, last
##
##
## Attaching package: 'PerformanceAnalytics'
##
## The following object is masked from 'package:graphics':
##
## legend
##
## Loading required package: quantmod
## Loading required package: TTR
##
## Attaching package: 'TTR'
##
## The following object is masked from 'package:dials':
##
## momentum
##
## Registered S3 method overwritten by 'quantmod':
## method from
## as.zoo.data.frame zoolibrary(readr)
data <- read_csv("C:/Users/Jstan/OneDrive/Desktop/Intermediate Data/PSUDAT3100/00_data/WA_Fn-UseC_-HR-Employee-Attrition.csv")## Rows: 1470 Columns: 35
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (9): Attrition, BusinessTravel, Department, EducationField, Gender, Job...
## dbl (26): Age, DailyRate, DistanceFromHome, Education, EmployeeCount, Employ...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.data## # A tibble: 1,470 × 35
## Age Attrition BusinessTravel DailyRate Department DistanceFromHome
## <dbl> <chr> <chr> <dbl> <chr> <dbl>
## 1 41 Yes Travel_Rarely 1102 Sales 1
## 2 49 No Travel_Frequently 279 Research & Deve… 8
## 3 37 Yes Travel_Rarely 1373 Research & Deve… 2
## 4 33 No Travel_Frequently 1392 Research & Deve… 3
## 5 27 No Travel_Rarely 591 Research & Deve… 2
## 6 32 No Travel_Frequently 1005 Research & Deve… 2
## 7 59 No Travel_Rarely 1324 Research & Deve… 3
## 8 30 No Travel_Rarely 1358 Research & Deve… 24
## 9 38 No Travel_Frequently 216 Research & Deve… 23
## 10 36 No Travel_Rarely 1299 Research & Deve… 27
## # ℹ 1,460 more rows
## # ℹ 29 more variables: Education <dbl>, EducationField <chr>,
## # EmployeeCount <dbl>, EmployeeNumber <dbl>, EnvironmentSatisfaction <dbl>,
## # Gender <chr>, HourlyRate <dbl>, JobInvolvement <dbl>, JobLevel <dbl>,
## # JobRole <chr>, JobSatisfaction <dbl>, MaritalStatus <chr>,
## # MonthlyIncome <dbl>, MonthlyRate <dbl>, NumCompaniesWorked <dbl>,
## # Over18 <chr>, OverTime <chr>, PercentSalaryHike <dbl>, …factors_vec <- data %>% select(Education, EnvironmentSatisfaction, JobInvolvement, JobSatisfaction, PerformanceRating, RelationshipSatisfaction, WorkLifeBalance) %>% names()
data_clean <- data %>%
#Address factors imported as numeric
mutate(across(all_of(factors_vec), as.factor)) %>%
# Drop zero-variance variables
select(-c(Over18, EmployeeCount, StandardHours)) %>%
mutate(across(where(is.character), factor))
glimpse(data_clean)## Rows: 1,470
## Columns: 32
## $ Age <dbl> 41, 49, 37, 33, 27, 32, 59, 30, 38, 36, 35, 2…
## $ Attrition <fct> Yes, No, Yes, No, No, No, No, No, No, No, No,…
## $ BusinessTravel <fct> Travel_Rarely, Travel_Frequently, Travel_Rare…
## $ DailyRate <dbl> 1102, 279, 1373, 1392, 591, 1005, 1324, 1358,…
## $ Department <fct> Sales, Research & Development, Research & Dev…
## $ DistanceFromHome <dbl> 1, 8, 2, 3, 2, 2, 3, 24, 23, 27, 16, 15, 26, …
## $ Education <fct> 2, 1, 2, 4, 1, 2, 3, 1, 3, 3, 3, 2, 1, 2, 3, …
## $ EducationField <fct> Life Sciences, Life Sciences, Other, Life Sci…
## $ EmployeeNumber <dbl> 1, 2, 4, 5, 7, 8, 10, 11, 12, 13, 14, 15, 16,…
## $ EnvironmentSatisfaction <fct> 2, 3, 4, 4, 1, 4, 3, 4, 4, 3, 1, 4, 1, 2, 3, …
## $ Gender <fct> Female, Male, Male, Female, Male, Male, Femal…
## $ HourlyRate <dbl> 94, 61, 92, 56, 40, 79, 81, 67, 44, 94, 84, 4…
## $ JobInvolvement <fct> 3, 2, 2, 3, 3, 3, 4, 3, 2, 3, 4, 2, 3, 3, 2, …
## $ JobLevel <dbl> 2, 2, 1, 1, 1, 1, 1, 1, 3, 2, 1, 2, 1, 1, 1, …
## $ JobRole <fct> Sales Executive, Research Scientist, Laborato…
## $ JobSatisfaction <fct> 4, 2, 3, 3, 2, 4, 1, 3, 3, 3, 2, 3, 3, 4, 3, …
## $ MaritalStatus <fct> Single, Married, Single, Married, Married, Si…
## $ MonthlyIncome <dbl> 5993, 5130, 2090, 2909, 3468, 3068, 2670, 269…
## $ MonthlyRate <dbl> 19479, 24907, 2396, 23159, 16632, 11864, 9964…
## $ NumCompaniesWorked <dbl> 8, 1, 6, 1, 9, 0, 4, 1, 0, 6, 0, 0, 1, 0, 5, …
## $ OverTime <fct> Yes, No, Yes, Yes, No, No, Yes, No, No, No, N…
## $ PercentSalaryHike <dbl> 11, 23, 15, 11, 12, 13, 20, 22, 21, 13, 13, 1…
## $ PerformanceRating <fct> 3, 4, 3, 3, 3, 3, 4, 4, 4, 3, 3, 3, 3, 3, 3, …
## $ RelationshipSatisfaction <fct> 1, 4, 2, 3, 4, 3, 1, 2, 2, 2, 3, 4, 4, 3, 2, …
## $ StockOptionLevel <dbl> 0, 1, 0, 0, 1, 0, 3, 1, 0, 2, 1, 0, 1, 1, 0, …
## $ TotalWorkingYears <dbl> 8, 10, 7, 8, 6, 8, 12, 1, 10, 17, 6, 10, 5, 3…
## $ TrainingTimesLastYear <dbl> 0, 3, 3, 3, 3, 2, 3, 2, 2, 3, 5, 3, 1, 2, 4, …
## $ WorkLifeBalance <fct> 1, 3, 3, 3, 3, 2, 2, 3, 3, 2, 3, 3, 2, 3, 3, …
## $ YearsAtCompany <dbl> 6, 10, 0, 8, 2, 7, 1, 1, 9, 7, 5, 9, 5, 2, 4,…
## $ YearsInCurrentRole <dbl> 4, 7, 0, 7, 2, 7, 0, 0, 7, 7, 4, 5, 2, 2, 2, …
## $ YearsSinceLastPromotion <dbl> 0, 1, 0, 3, 2, 3, 0, 0, 1, 7, 0, 0, 4, 1, 0, …
## $ YearsWithCurrManager <dbl> 5, 7, 0, 0, 2, 6, 0, 0, 8, 7, 3, 8, 3, 2, 3, …set.seed(1234)
data_split <- initial_split(data_clean, strata = "Attrition")
train_tbl <- training(data_split)
test_tbl <- testing(data_split)recipe_obj <- recipe(Attrition ~ ., data = train_tbl) %>%
# Remove zero variance variables
step_zv(all_predictors()) # Initialize h2o
h2o.init()##
## H2O is not running yet, starting it now...
##
## Note: In case of errors look at the following log files:
## C:\Users\Jstan\AppData\Local\Temp\RtmpOETykb\file6420765b1841/h2o_Jstan_started_from_r.out
## C:\Users\Jstan\AppData\Local\Temp\RtmpOETykb\file6420655eb86/h2o_Jstan_started_from_r.err
##
##
## Starting H2O JVM and connecting: . Connection successful!
##
## R is connected to the H2O cluster:
## H2O cluster uptime: 4 seconds 65 milliseconds
## H2O cluster timezone: America/New_York
## H2O data parsing timezone: UTC
## H2O cluster version: 3.44.0.3
## H2O cluster version age: 4 months and 18 days
## H2O cluster name: H2O_started_from_R_Jstan_fhp551
## H2O cluster total nodes: 1
## H2O cluster total memory: 1.94 GB
## H2O cluster total cores: 8
## H2O cluster allowed cores: 8
## H2O cluster healthy: TRUE
## H2O Connection ip: localhost
## H2O Connection port: 54321
## H2O Connection proxy: NA
## H2O Internal Security: FALSE
## R Version: R version 4.3.1 (2023-06-16 ucrt)## Warning in h2o.clusterInfo():
## Your H2O cluster version is (4 months and 18 days) old. There may be a newer version available.
## Please download and install the latest version from: https://h2o-release.s3.amazonaws.com/h2o/latest_stable.htmlsplit.h2o <- h2o.splitFrame(as.h2o(train_tbl), ratios = c(0.85), seed = 1235)##
|
| | 0%
|
|======================================================================| 100%train_h2o <- split.h2o[[1]]
valid_h2o <- split.h2o[[2]]
test_h2o <- as.h2o(test_tbl)##
|
| | 0%
|
|======================================================================| 100%y <- "Attrition"
x <- setdiff(names(train_tbl), y)
models_h2o <- h2o.automl(
x = x,
y = y,
training_frame = train_h2o,
validation_frame = valid_h2o,
leaderboard_frame = test_h2o,
max_runtime_secs = 30,
nfolds = 5,
seed = 1523)##
|
| | 0%
|
|=== | 4%
## 13:13:42.594: User specified a validation frame with cross-validation still enabled. Please note that the models will still be validated using cross-validation only, the validation frame will be used to provide purely informative validation metrics on the trained models.
## 13:13:42.615: AutoML: XGBoost is not available; skipping it.
|
|======= | 11%
|
|============== | 20%
|
|=================== | 27%
|
|========================= | 36%
|
|============================== | 43%
|
|=================================== | 51%
|
|========================================== | 59%
|
|============================================== | 66%
|
|=================================================== | 73%
|
|======================================================== | 80%
|
|============================================================= | 88%
|
|================================================================== | 95%
|
|======================================================================| 100%models_h2o %>% typeof()## [1] "S4"models_h2o %>% slotNames()## [1] "project_name" "leader" "leaderboard" "event_log"
## [5] "modeling_steps" "training_info"models_h2o@leaderboard## model_id auc logloss
## 1 StackedEnsemble_BestOfFamily_1_AutoML_1_20240509_131342 0.8457389 0.3108670
## 2 GBM_1_AutoML_1_20240509_131342 0.8393743 0.3276879
## 3 StackedEnsemble_BestOfFamily_4_AutoML_1_20240509_131342 0.8377023 0.3313649
## 4 StackedEnsemble_BestOfFamily_3_AutoML_1_20240509_131342 0.8368932 0.3242535
## 5 StackedEnsemble_BestOfFamily_2_AutoML_1_20240509_131342 0.8368393 0.3221989
## 6 GLM_1_AutoML_1_20240509_131342 0.8366775 0.3216285
## aucpr mean_per_class_error rmse mse
## 1 0.6521905 0.2149676 0.3012776 0.09076818
## 2 0.6011589 0.2246764 0.3128686 0.09788676
## 3 0.6640050 0.1996764 0.3087477 0.09532513
## 4 0.6612023 0.2406958 0.3041739 0.09252174
## 5 0.6574361 0.2101133 0.3029567 0.09178274
## 6 0.6330270 0.2557443 0.3074904 0.09455036
##
## [25 rows x 7 columns]best_model <- models_h2o@leader
best_model## Model Details:
## ==============
##
## H2OBinomialModel: stackedensemble
## Model ID: StackedEnsemble_BestOfFamily_1_AutoML_1_20240509_131342
## Model Summary for Stacked Ensemble:
## key value
## 1 Stacking strategy cross_validation
## 2 Number of base models (used / total) 2/2
## 3 # GBM base models (used / total) 1/1
## 4 # GLM base models (used / total) 1/1
## 5 Metalearner algorithm GLM
## 6 Metalearner fold assignment scheme Random
## 7 Metalearner nfolds 5
## 8 Metalearner fold_column NA
## 9 Custom metalearner hyperparameters None
##
##
## H2OBinomialMetrics: stackedensemble
## ** Reported on training data. **
##
## MSE: 0.07074738
## RMSE: 0.2659838
## LogLoss: 0.2523481
## Mean Per-Class Error: 0.167144
## AUC: 0.9123509
## AUCPR: 0.7992585
## Gini: 0.8247018
##
## Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold:
## No Yes Error Rate
## No 746 35 0.044814 =35/781
## Yes 44 108 0.289474 =44/152
## Totals 790 143 0.084673 =79/933
##
## Maximum Metrics: Maximum metrics at their respective thresholds
## metric threshold value idx
## 1 max f1 0.362001 0.732203 115
## 2 max f2 0.267316 0.759097 149
## 3 max f0point5 0.516904 0.830078 74
## 4 max accuracy 0.516904 0.922830 74
## 5 max precision 0.948780 1.000000 0
## 6 max recall 0.004867 1.000000 391
## 7 max specificity 0.948780 1.000000 0
## 8 max absolute_mcc 0.516904 0.691492 74
## 9 max min_per_class_accuracy 0.206982 0.842105 183
## 10 max mean_per_class_accuracy 0.267316 0.854492 149
## 11 max tns 0.948780 781.000000 0
## 12 max fns 0.948780 151.000000 0
## 13 max fps 0.000723 781.000000 399
## 14 max tps 0.004867 152.000000 391
## 15 max tnr 0.948780 1.000000 0
## 16 max fnr 0.948780 0.993421 0
## 17 max fpr 0.000723 1.000000 399
## 18 max tpr 0.004867 1.000000 391
##
## Gains/Lift Table: Extract with `h2o.gainsLift(<model>, <data>)` or `h2o.gainsLift(<model>, valid=<T/F>, xval=<T/F>)`
## H2OBinomialMetrics: stackedensemble
## ** Reported on validation data. **
##
## MSE: 0.07848421
## RMSE: 0.2801503
## LogLoss: 0.2783257
## Mean Per-Class Error: 0.2044755
## AUC: 0.866014
## AUCPR: 0.6889672
## Gini: 0.732028
##
## Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold:
## No Yes Error Rate
## No 136 7 0.048951 =7/143
## Yes 9 16 0.360000 =9/25
## Totals 145 23 0.095238 =16/168
##
## Maximum Metrics: Maximum metrics at their respective thresholds
## metric threshold value idx
## 1 max f1 0.410193 0.666667 22
## 2 max f2 0.140760 0.677419 54
## 3 max f0point5 0.557493 0.740741 13
## 4 max accuracy 0.557493 0.910714 13
## 5 max precision 0.972597 1.000000 0
## 6 max recall 0.009026 1.000000 155
## 7 max specificity 0.972597 1.000000 0
## 8 max absolute_mcc 0.513884 0.620966 19
## 9 max min_per_class_accuracy 0.175808 0.790210 49
## 10 max mean_per_class_accuracy 0.140760 0.801119 54
## 11 max tns 0.972597 143.000000 0
## 12 max fns 0.972597 24.000000 0
## 13 max fps 0.002717 143.000000 167
## 14 max tps 0.009026 25.000000 155
## 15 max tnr 0.972597 1.000000 0
## 16 max fnr 0.972597 0.960000 0
## 17 max fpr 0.002717 1.000000 167
## 18 max tpr 0.009026 1.000000 155
##
## Gains/Lift Table: Extract with `h2o.gainsLift(<model>, <data>)` or `h2o.gainsLift(<model>, valid=<T/F>, xval=<T/F>)`
## H2OBinomialMetrics: stackedensemble
## ** Reported on cross-validation data. **
## ** 5-fold cross-validation on training data (Metrics computed for combined holdout predictions) **
##
## MSE: 0.09410337
## RMSE: 0.3067627
## LogLoss: 0.3272577
## Mean Per-Class Error: 0.2418079
## AUC: 0.832081
## AUCPR: 0.6284665
## Gini: 0.664162
##
## Confusion Matrix (vertical: actual; across: predicted) for F1-optimal threshold:
## No Yes Error Rate
## No 727 54 0.069142 =54/781
## Yes 63 89 0.414474 =63/152
## Totals 790 143 0.125402 =117/933
##
## Maximum Metrics: Maximum metrics at their respective thresholds
## metric threshold value idx
## 1 max f1 0.329737 0.603390 114
## 2 max f2 0.169543 0.651214 204
## 3 max f0point5 0.490390 0.678879 65
## 4 max accuracy 0.490390 0.888532 65
## 5 max precision 0.943267 1.000000 0
## 6 max recall 0.001498 1.000000 398
## 7 max specificity 0.943267 1.000000 0
## 8 max absolute_mcc 0.329737 0.529345 114
## 9 max min_per_class_accuracy 0.171534 0.769737 202
## 10 max mean_per_class_accuracy 0.184402 0.774665 195
## 11 max tns 0.943267 781.000000 0
## 12 max fns 0.943267 151.000000 0
## 13 max fps 0.001051 781.000000 399
## 14 max tps 0.001498 152.000000 398
## 15 max tnr 0.943267 1.000000 0
## 16 max fnr 0.943267 0.993421 0
## 17 max fpr 0.001051 1.000000 399
## 18 max tpr 0.001498 1.000000 398
##
## Gains/Lift Table: Extract with `h2o.gainsLift(<model>, <data>)` or `h2o.gainsLift(<model>, valid=<T/F>, xval=<T/F>)`
## Cross-Validation Metrics Summary:
## mean sd cv_1_valid cv_2_valid cv_3_valid cv_4_valid
## accuracy 0.882697 0.038071 0.917098 0.910995 0.871508 0.822581
## auc 0.828122 0.046002 0.866695 0.803746 0.851724 0.859774
## err 0.117303 0.038071 0.082902 0.089005 0.128492 0.177419
## err_count 21.800000 6.833740 16.000000 17.000000 23.000000 33.000000
## f0point5 0.659094 0.081791 0.743802 0.720000 0.664557 0.536481
## cv_5_valid
## accuracy 0.891304
## auc 0.758669
## err 0.108696
## err_count 20.000000
## f0point5 0.630631
##
## ---
## mean sd cv_1_valid cv_2_valid cv_3_valid
## precision 0.675339 0.109435 0.782609 0.750000 0.677419
## r2 0.304412 0.046335 0.359537 0.317880 0.314642
## recall 0.627024 0.085125 0.620690 0.620690 0.617647
## residual_deviance 121.894005 7.082528 109.750520 126.416504 123.152390
## rmse 0.307452 0.016203 0.285964 0.296383 0.324735
## specificity 0.931105 0.054799 0.969512 0.962963 0.931034
## cv_4_valid cv_5_valid
## precision 0.500000 0.666667
## r2 0.297836 0.232165
## recall 0.757576 0.518518
## residual_deviance 122.740060 127.410545
## rmse 0.320117 0.310062
## specificity 0.836601 0.955414#h2o.getModel("GBM_grid_1_AutoML_1_20240417_173906_model_1") %>%
#h2o.saveModel("h2o_models/")
#best_model <- h2o.loadModel("h2o_models/GBM_grid_1_AutoML_1_20240417_173906_model_1")predictions <- h2o.predict(best_model, newdata = test_h2o)##
|
| | 0%
|
|======================================================================| 100%predictions_tbl <- predictions %>%
as_tibble()
predictions_tbl %>%
bind_cols(test_tbl)## # A tibble: 369 × 35
## predict No Yes Age Attrition BusinessTravel DailyRate Department
## <fct> <dbl> <dbl> <dbl> <fct> <fct> <dbl> <fct>
## 1 No 0.690 0.310 59 No Travel_Rarely 1324 Research &…
## 2 No 0.933 0.0670 35 No Travel_Rarely 809 Research &…
## 3 No 0.961 0.0392 34 No Travel_Rarely 1346 Research &…
## 4 No 0.785 0.215 22 No Non-Travel 1123 Research &…
## 5 No 0.976 0.0236 53 No Travel_Rarely 1219 Sales
## 6 No 0.978 0.0222 24 No Non-Travel 673 Research &…
## 7 No 0.873 0.127 21 No Travel_Rarely 391 Research &…
## 8 No 0.932 0.0679 34 Yes Travel_Rarely 699 Research &…
## 9 No 0.997 0.00306 53 No Travel_Rarely 1282 Research &…
## 10 Yes 0.169 0.831 32 Yes Travel_Frequently 1125 Research &…
## # ℹ 359 more rows
## # ℹ 27 more variables: DistanceFromHome <dbl>, Education <fct>,
## # EducationField <fct>, EmployeeNumber <dbl>, EnvironmentSatisfaction <fct>,
## # Gender <fct>, HourlyRate <dbl>, JobInvolvement <fct>, JobLevel <dbl>,
## # JobRole <fct>, JobSatisfaction <fct>, MaritalStatus <fct>,
## # MonthlyIncome <dbl>, MonthlyRate <dbl>, NumCompaniesWorked <dbl>,
## # OverTime <fct>, PercentSalaryHike <dbl>, PerformanceRating <fct>, …?h2o.performance## starting httpd help server ... doneperformance_h2o <- h2o.performance(best_model, newdata = test_h2o)
typeof(performance_h2o)## [1] "S4"slotNames(performance_h2o)## [1] "algorithm" "on_train" "on_valid" "on_xval" "metrics"performance_h2o@metrics## $model
## $model$`__meta`
## $model$`__meta`$schema_version
## [1] 3
##
## $model$`__meta`$schema_name
## [1] "ModelKeyV3"
##
## $model$`__meta`$schema_type
## [1] "Key<Model>"
##
##
## $model$name
## [1] "StackedEnsemble_BestOfFamily_1_AutoML_1_20240509_131342"
##
## $model$type
## [1] "Key<Model>"
##
## $model$URL
## [1] "/3/Models/StackedEnsemble_BestOfFamily_1_AutoML_1_20240509_131342"
##
##
## $model_checksum
## [1] "-5820552871344264680"
##
## $frame
## $frame$name
## [1] "test_tbl_sid_9840_3"
##
##
## $frame_checksum
## [1] "-54414925341915866"
##
## $description
## NULL
##
## $scoring_time
## [1] 1.715275e+12
##
## $predictions
## NULL
##
## $MSE
## [1] 0.09076818
##
## $RMSE
## [1] 0.3012776
##
## $nobs
## [1] 369
##
## $custom_metric_name
## NULL
##
## $custom_metric_value
## [1] 0
##
## $r2
## [1] 0.3333826
##
## $logloss
## [1] 0.310867
##
## $AUC
## [1] 0.8457389
##
## $pr_auc
## [1] 0.6521905
##
## $Gini
## [1] 0.6914779
##
## $mean_per_class_error
## [1] 0.2149676
##
## $domain
## [1] "No" "Yes"
##
## $cm
## $cm$`__meta`
## $cm$`__meta`$schema_version
## [1] 3
##
## $cm$`__meta`$schema_name
## [1] "ConfusionMatrixV3"
##
## $cm$`__meta`$schema_type
## [1] "ConfusionMatrix"
##
##
## $cm$table
## Confusion Matrix: Row labels: Actual class; Column labels: Predicted class
## No Yes Error Rate
## No 274 35 0.1133 = 35 / 309
## Yes 19 41 0.3167 = 19 / 60
## Totals 293 76 0.1463 = 54 / 369
##
##
## $thresholds_and_metric_scores
## Metrics for Thresholds: Binomial metrics as a function of classification thresholds
## threshold f1 f2 f0point5 accuracy precision recall specificity
## 1 0.948260 0.032787 0.020747 0.078125 0.840108 1.000000 0.016667 1.000000
## 2 0.940206 0.064516 0.041322 0.147059 0.842818 1.000000 0.033333 1.000000
## 3 0.940102 0.095238 0.061728 0.208333 0.845528 1.000000 0.050000 1.000000
## 4 0.893476 0.125000 0.081967 0.263158 0.848238 1.000000 0.066667 1.000000
## 5 0.831028 0.153846 0.102041 0.312500 0.850949 1.000000 0.083333 1.000000
## absolute_mcc min_per_class_accuracy mean_per_class_accuracy tns fns fps tps
## 1 0.118299 0.016667 0.508333 309 59 0 1
## 2 0.167527 0.033333 0.516667 309 58 0 2
## 3 0.205458 0.050000 0.525000 309 57 0 3
## 4 0.237568 0.066667 0.533333 309 56 0 4
## 5 0.265973 0.083333 0.541667 309 55 0 5
## tnr fnr fpr tpr idx
## 1 1.000000 0.983333 0.000000 0.016667 0
## 2 1.000000 0.966667 0.000000 0.033333 1
## 3 1.000000 0.950000 0.000000 0.050000 2
## 4 1.000000 0.933333 0.000000 0.066667 3
## 5 1.000000 0.916667 0.000000 0.083333 4
##
## ---
## threshold f1 f2 f0point5 accuracy precision recall
## 364 0.001661 0.283019 0.496689 0.197889 0.176152 0.164835 1.000000
## 365 0.001315 0.282353 0.495868 0.197368 0.173442 0.164384 1.000000
## 366 0.001124 0.281690 0.495050 0.196850 0.170732 0.163934 1.000000
## 367 0.000865 0.281030 0.494234 0.196335 0.168022 0.163488 1.000000
## 368 0.000809 0.280374 0.493421 0.195822 0.165312 0.163043 1.000000
## 369 0.000570 0.279720 0.492611 0.195313 0.162602 0.162602 1.000000
## specificity absolute_mcc min_per_class_accuracy mean_per_class_accuracy tns
## 364 0.016181 0.051645 0.016181 0.508091 5
## 365 0.012945 0.046130 0.012945 0.506472 4
## 366 0.009709 0.039895 0.009709 0.504854 3
## 367 0.006472 0.032530 0.006472 0.503236 2
## 368 0.003236 0.022971 0.003236 0.501618 1
## 369 0.000000 0.000000 0.000000 0.500000 0
## fns fps tps tnr fnr fpr tpr idx
## 364 0 304 60 0.016181 0.000000 0.983819 1.000000 363
## 365 0 305 60 0.012945 0.000000 0.987055 1.000000 364
## 366 0 306 60 0.009709 0.000000 0.990291 1.000000 365
## 367 0 307 60 0.006472 0.000000 0.993528 1.000000 366
## 368 0 308 60 0.003236 0.000000 0.996764 1.000000 367
## 369 0 309 60 0.000000 0.000000 1.000000 1.000000 368
##
## $max_criteria_and_metric_scores
## Maximum Metrics: Maximum metrics at their respective thresholds
## metric threshold value idx
## 1 max f1 0.296958 0.602941 75
## 2 max f2 0.260290 0.665635 82
## 3 max f0point5 0.606458 0.705128 23
## 4 max accuracy 0.606458 0.891599 23
## 5 max precision 0.948260 1.000000 0
## 6 max recall 0.018508 1.000000 301
## 7 max specificity 0.948260 1.000000 0
## 8 max absolute_mcc 0.467279 0.541270 38
## 9 max min_per_class_accuracy 0.158723 0.760518 119
## 10 max mean_per_class_accuracy 0.260290 0.793608 82
## 11 max tns 0.948260 309.000000 0
## 12 max fns 0.948260 59.000000 0
## 13 max fps 0.000570 309.000000 368
## 14 max tps 0.018508 60.000000 301
## 15 max tnr 0.948260 1.000000 0
## 16 max fnr 0.948260 0.983333 0
## 17 max fpr 0.000570 1.000000 368
## 18 max tpr 0.018508 1.000000 301
##
## $gains_lift_table
## Gains/Lift Table: Avg response rate: 16.26 %, avg score: 16.72 %
## group cumulative_data_fraction lower_threshold lift cumulative_lift
## 1 1 0.01084011 0.851011 6.150000 6.150000
## 2 2 0.02168022 0.759438 4.612500 5.381250
## 3 3 0.03252033 0.710368 6.150000 5.637500
## 4 4 0.04065041 0.669865 6.150000 5.740000
## 5 5 0.05149051 0.653368 6.150000 5.826316
## 6 6 0.10027100 0.493950 3.075000 4.487838
## 7 7 0.15176152 0.348824 1.942105 3.624107
## 8 8 0.20054201 0.298783 2.391667 3.324324
## 9 9 0.30081301 0.189414 0.831081 2.493243
## 10 10 0.40108401 0.123223 0.498649 1.994595
## 11 11 0.50135501 0.080567 0.498649 1.695405
## 12 12 0.59891599 0.054548 0.683333 1.530543
## 13 13 0.69918699 0.035283 0.166216 1.334884
## 14 14 0.79945799 0.020677 0.332432 1.209153
## 15 15 0.89972900 0.009679 0.332432 1.111446
## 16 16 1.00000000 0.000570 0.000000 1.000000
## response_rate score cumulative_response_rate cumulative_score
## 1 1.000000 0.930511 1.000000 0.930511
## 2 0.750000 0.795514 0.875000 0.863012
## 3 1.000000 0.729192 0.916667 0.818406
## 4 1.000000 0.682241 0.933333 0.791173
## 5 1.000000 0.661355 0.947368 0.763843
## 6 0.500000 0.570087 0.729730 0.669583
## 7 0.315789 0.408511 0.589286 0.581005
## 8 0.388889 0.322687 0.540541 0.518171
## 9 0.135135 0.230009 0.405405 0.422117
## 10 0.081081 0.146515 0.324324 0.353217
## 11 0.081081 0.099645 0.275676 0.302502
## 12 0.111111 0.067385 0.248869 0.264203
## 13 0.027027 0.043611 0.217054 0.232567
## 14 0.054054 0.025474 0.196610 0.206593
## 15 0.054054 0.015610 0.180723 0.185309
## 16 0.000000 0.004749 0.162602 0.167204
## capture_rate cumulative_capture_rate gain cumulative_gain
## 1 0.066667 0.066667 515.000000 515.000000
## 2 0.050000 0.116667 361.250000 438.125000
## 3 0.066667 0.183333 515.000000 463.750000
## 4 0.050000 0.233333 515.000000 474.000000
## 5 0.066667 0.300000 515.000000 482.631579
## 6 0.150000 0.450000 207.500000 348.783784
## 7 0.100000 0.550000 94.210526 262.410714
## 8 0.116667 0.666667 139.166667 232.432432
## 9 0.083333 0.750000 -16.891892 149.324324
## 10 0.050000 0.800000 -50.135135 99.459459
## 11 0.050000 0.850000 -50.135135 69.540541
## 12 0.066667 0.916667 -31.666667 53.054299
## 13 0.016667 0.933333 -83.378378 33.488372
## 14 0.033333 0.966667 -66.756757 20.915254
## 15 0.033333 1.000000 -66.756757 11.144578
## 16 0.000000 1.000000 -100.000000 0.000000
## kolmogorov_smirnov
## 1 0.066667
## 2 0.113430
## 3 0.180097
## 4 0.230097
## 5 0.296764
## 6 0.417638
## 7 0.475566
## 8 0.556634
## 9 0.536408
## 10 0.476375
## 11 0.416343
## 12 0.379450
## 13 0.279612
## 14 0.199676
## 15 0.119741
## 16 0.000000
##
## $residual_deviance
## [1] 229.4198
##
## $null_deviance
## [1] 327.6419
##
## $AIC
## [1] 235.4198
##
## $loglikelihood
## [1] 0
##
## $null_degrees_of_freedom
## [1] 368
##
## $residual_degrees_of_freedom
## [1] 366h2o.auc(performance_h2o)## [1] 0.8457389h2o.confusionMatrix(performance_h2o)## Confusion Matrix (vertical: actual; across: predicted) for max f1 @ threshold = 0.296958386752964:
## No Yes Error Rate
## No 274 35 0.113269 =35/309
## Yes 19 41 0.316667 =19/60
## Totals 293 76 0.146341 =54/369