This function lets the user get a confusion matrix and accuracy, and for for binary classification models: AUC, Precision, Sensitivity, and Specificity, given the expected (tags) values and predicted values (scores).
Usage
model_metrics(
tag,
score,
multis = NA,
abc = TRUE,
thresh = 10,
auto_n = TRUE,
thresh_cm = 0.5,
target = "auto",
type = "test",
model_name = NA,
plots = TRUE,
quiet = FALSE,
subtitle = NA
)Arguments
- tag
Vector. Real known label
- score
Vector. Predicted value or model's result
- multis
Data.frame. Containing columns with each category score (only used when more than 2 categories coexist)
- abc
Boolean. Arrange columns and rows alphabetically when categorical values?
- thresh
Integer. Threshold for selecting binary or regression models: this number is the threshold of unique values we should have in
'tag'(more than: regression; less than: classification)- auto_n
Add
n_before digits when it's categorical and not numerical, even though seems numerical?- thresh_cm
Numeric. Value to splits the results for the confusion matrix. Range of values: (0-1)
- target
Value. Which is your target positive value? If set to
'auto', the target with largestmean(score)will be selected. Change the value to overwrite. Only used when binary categorical model.- type
Character. One of: "train", "test".
- model_name
Character. Model's name for reference.
- plots
Boolean. Create plots objects?
- quiet
Boolean. Quiet all messages, warnings, recommendations?
- subtitle
Character. Subtitle for plots
Value
List. Multiple performance metrics that vary depending on
the type of model (classification or regression). If plot=TRUE,
multiple plots are also returned.
See also
Other Machine Learning:
ROC(),
conf_mat(),
export_results(),
gain_lift(),
h2o_automl(),
h2o_predict_API(),
h2o_predict_MOJO(),
h2o_predict_binary(),
h2o_predict_model(),
h2o_selectmodel(),
impute(),
iter_seeds(),
lasso_vars(),
model_preprocess(),
msplit()
Other Model metrics:
ROC(),
conf_mat(),
errors(),
gain_lift(),
loglossBinary()
Examples
data(dfr) # Results for AutoML Predictions
lapply(dfr, head)
#> $class2
#> tag scores
#> 1 TRUE 0.3155498
#> 2 TRUE 0.8747599
#> 3 TRUE 0.8952823
#> 4 FALSE 0.0436517
#> 5 TRUE 0.2196593
#> 6 FALSE 0.2816101
#>
#> $class3
#> tag score n_1 n_2 n_3
#> 1 n_3 n_2 0.20343865 0.60825062 0.18831071
#> 2 n_2 n_3 0.17856154 0.07657769 0.74486071
#> 3 n_1 n_1 0.50516951 0.40168718 0.09314334
#> 4 n_3 n_2 0.30880713 0.39062151 0.30057135
#> 5 n_2 n_3 0.01956827 0.07069011 0.90974158
#> 6 n_2 n_3 0.07830017 0.15408720 0.76761264
#>
#> $regr
#> tag score
#> 1 11.1333 25.93200
#> 2 30.0708 39.91900
#> 3 26.5500 50.72246
#> 4 31.2750 47.81292
#> 5 13.0000 30.12853
#> 6 26.0000 13.24153
#>
# Metrics for Binomial Model
met1 <- model_metrics(dfr$class2$tag, dfr$class2$scores,
model_name = "Titanic Survived Model",
plots = FALSE
)
#> Target value: TRUE
print(met1)
#> $dictionary
#> [1] "AUC: Area Under the Curve"
#> [2] "ACC: Accuracy"
#> [3] "PRC: Precision = Positive Predictive Value"
#> [4] "TPR: Sensitivity = Recall = Hit rate = True Positive Rate"
#> [5] "TNR: Specificity = Selectivity = True Negative Rate"
#> [6] "Logloss (Error): Logarithmic loss [Neutral classification: 0.69315]"
#> [7] "Gain: When best n deciles selected, what % of the real target observations are picked?"
#> [8] "Lift: When best n deciles selected, how much better than random is?"
#>
#> $confusion_matrix
#> Pred
#> Real FALSE TRUE
#> FALSE 9 156
#> TRUE 68 35
#>
#> $gain_lift
#> # A tibble: 10 × 10
#> percentile value random target total gain optimal lift response score
#> <fct> <fct> <dbl> <int> <int> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 TRUE 10.4 26 28 25.2 27.2 142. 25.2 93.3
#> 2 2 TRUE 20.5 24 27 48.5 53.4 137. 23.3 79.7
#> 3 3 TRUE 30.2 19 26 67.0 78.6 122. 18.4 46.2
#> 4 4 TRUE 39.9 12 26 78.6 100 97.0 11.7 31.5
#> 5 5 TRUE 50 9 27 87.4 100 74.8 8.74 19.8
#> 6 6 TRUE 60.1 5 27 92.2 100 53.5 4.85 13.3
#> 7 7 TRUE 69.8 3 26 95.1 100 36.4 2.91 9.31
#> 8 8 TRUE 79.9 2 27 97.1 100 21.6 1.94 7.49
#> 9 9 TRUE 89.9 3 27 100 100 11.2 2.91 6.14
#> 10 10 TRUE 100 0 27 100 100 0 0 3.19
#>
#> $metrics
#> AUC ACC PRC TPR TNR
#> 1 0.89467 0.16418 0.18325 0.33981 0.054545
#>
# Metrics for Multi-Categorical Model
met2 <- model_metrics(dfr$class3$tag, dfr$class3$score,
multis = subset(dfr$class3, select = -c(tag, score)),
model_name = "Titanic Class Model",
plots = FALSE
)
print(met2)
#> $dictionary
#> [1] "AUC: Area Under the Curve"
#> [2] "ACC: Accuracy"
#> [3] "PRC: Precision = Positive Predictive Value"
#> [4] "TPR: Sensitivity = Recall = Hit rate = True Positive Rate"
#> [5] "TNR: Specificity = Selectivity = True Negative Rate"
#> [6] "Logloss (Error): Logarithmic loss [Neutral classification: 0.69315]"
#> [7] "Gain: When best n deciles selected, what % of the real target observations are picked?"
#> [8] "Lift: When best n deciles selected, how much better than random is?"
#>
#> $confusion_matrix
#> # A tibble: 3 × 4
#> `Real x Pred` n_3 n_1 n_2
#> <fct> <int> <int> <int>
#> 1 n_3 120 11 18
#> 2 n_1 12 43 8
#> 3 n_2 26 15 15
#>
#> $metrics
#> AUC ACC
#> 1 0.7896 0.66418
#>
#> $metrics_tags
#> # A tibble: 3 × 9
#> tag n p AUC order ACC PRC TPR TNR
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 n_3 149 55.6 0.826 1 0.75 0.759 0.805 0.681
#> 2 n_1 63 23.5 0.867 2 0.828 0.623 0.683 0.873
#> 3 n_2 56 20.9 0.675 3 0.75 0.366 0.268 0.877
#>
# Metrics for Regression Model
met3 <- model_metrics(dfr$regr$tag, dfr$regr$score,
model_name = "Titanic Fare Model",
plots = FALSE
)
print(met3)
#> $dictionary
#> [1] "RMSE: Root Mean Squared Error"
#> [2] "MAE: Mean Average Error"
#> [3] "MAPE: Mean Absolute Percentage Error"
#> [4] "MSE: Mean Squared Error"
#> [5] "RSQ: R Squared"
#> [6] "RSQA: Adjusted R Squared"
#>
#> $metrics
#> rmse mae mape mse rsq rsqa
#> 1 20.30881 14.24359 0.07303959 412.4477 0.3169 0.3143
#>