These functions calculate the recall() of a measurement system for finding relevant documents compared to reference results (the truth regarding relevance). Highly related functions are precision() and f_meas().

recall(data, ...)

# S3 method for data.frame
recall(data, truth, estimate, estimator = NULL, na_rm = TRUE, ...)

recall_vec(truth, estimate, estimator = NULL, na_rm = TRUE, ...)



Either a data.frame containing the truth and estimate columns, or a table/matrix where the true class results should be in the columns of the table.


Not currently used.


The column identifier for the true class results (that is a factor). This should be an unquoted column name although this argument is passed by expression and supports quasiquotation (you can unquote column names). For _vec() functions, a factor vector.


The column identifier for the predicted class results (that is also factor). As with truth this can be specified different ways but the primary method is to use an unquoted variable name. For _vec() functions, a factor vector.


One of: "binary", "macro", "macro_weighted", or "micro" to specify the type of averaging to be done. "binary" is only relevant for the two class case. The other three are general methods for calculating multiclass metrics. The default will automatically choose "binary" or "macro" based on estimate.


A logical value indicating whether NA values should be stripped before the computation proceeds.


A tibble with columns .metric, .estimator, and .estimate and 1 row of values.

For grouped data frames, the number of rows returned will be the same as the number of groups.

For recall_vec(), a single numeric value (or NA).


The recall (aka sensitivity) is defined as the proportion of relevant results out of the number of samples which were actually relevant. When there are no relevant results, recall is not defined and a value of NA is returned.

When the denominator of the calculation is 0, recall is undefined. This happens when both # true_positive = 0 and # false_negative = 0 are true, which mean that there were no true events. When computing binary recall, a NA value will be returned with a warning. When computing multiclass recall, the individual NA values will be removed, and the computation will procede, with a warning.

Relevant Level

There is no common convention on which factor level should automatically be considered the "event" or "positive" result. In yardstick, the default is to use the first level. To change this, a global option called yardstick.event_first is set to TRUE when the package is loaded. This can be changed to FALSE if the last level of the factor is considered the level of interest by running: options(yardstick.event_first = FALSE). For multiclass extensions involving one-vs-all comparisons (such as macro averaging), this option is ignored and the "one" level is always the relevant result.


Macro, micro, and macro-weighted averaging is available for this metric. The default is to select macro averaging if a truth factor with more than 2 levels is provided. Otherwise, a standard binary calculation is done. See vignette("multiclass", "yardstick") for more information.


Suppose a 2x2 table with notation:


The formulas used here are:

$$recall = A/(A+C)$$ $$precision = A/(A+B)$$ $$F_meas_\beta = (1+\beta^2) * precision * recall/((\beta^2 * precision)+recall)$$

See the references for discussions of the statistics.


Buckland, M., & Gey, F. (1994). The relationship between Recall and Precision. Journal of the American Society for Information Science, 45(1), 12-19.

Powers, D. (2007). Evaluation: From Precision, Recall and F Factor to ROC, Informedness, Markedness and Correlation. Technical Report SIE-07-001, Flinders University

See also

Other class metrics: accuracy(), bal_accuracy(), detection_prevalence(), f_meas(), j_index(), kap(), mcc(), npv(), ppv(), precision(), sens(), spec()

Other relevance metrics: f_meas(), precision()


# Two class data("two_class_example") recall(two_class_example, truth, predicted)
#> # A tibble: 1 x 3 #> .metric .estimator .estimate #> <chr> <chr> <dbl> #> 1 recall binary 0.880
# Multiclass library(dplyr) data(hpc_cv) hpc_cv %>% filter(Resample == "Fold01") %>% recall(obs, pred)
#> # A tibble: 1 x 3 #> .metric .estimator .estimate #> <chr> <chr> <dbl> #> 1 recall macro 0.548
# Groups are respected hpc_cv %>% group_by(Resample) %>% recall(obs, pred)
#> # A tibble: 10 x 4 #> Resample .metric .estimator .estimate #> <chr> <chr> <chr> <dbl> #> 1 Fold01 recall macro 0.548 #> 2 Fold02 recall macro 0.541 #> 3 Fold03 recall macro 0.634 #> 4 Fold04 recall macro 0.570 #> 5 Fold05 recall macro 0.550 #> 6 Fold06 recall macro 0.540 #> 7 Fold07 recall macro 0.531 #> 8 Fold08 recall macro 0.584 #> 9 Fold09 recall macro 0.568 #> 10 Fold10 recall macro 0.537
# Weighted macro averaging hpc_cv %>% group_by(Resample) %>% recall(obs, pred, estimator = "macro_weighted")
#> # A tibble: 10 x 4 #> Resample .metric .estimator .estimate #> <chr> <chr> <chr> <dbl> #> 1 Fold01 recall macro_weighted 0.726 #> 2 Fold02 recall macro_weighted 0.712 #> 3 Fold03 recall macro_weighted 0.758 #> 4 Fold04 recall macro_weighted 0.712 #> 5 Fold05 recall macro_weighted 0.712 #> 6 Fold06 recall macro_weighted 0.697 #> 7 Fold07 recall macro_weighted 0.675 #> 8 Fold08 recall macro_weighted 0.721 #> 9 Fold09 recall macro_weighted 0.673 #> 10 Fold10 recall macro_weighted 0.699
# Vector version recall_vec(two_class_example$truth, two_class_example$predicted)
#> [1] 0.879845
# Making Class2 the "relevant" level options(yardstick.event_first = FALSE) recall_vec(two_class_example$truth, two_class_example$predicted)
#> [1] 0.7933884
options(yardstick.event_first = TRUE)