Package 'YARF'

The results from the function YARFROC.

Make a plot of the ROC with the AUC shaded in? Default is TRUE.

A YARF model object.

A n* x p matrix where each row is an observation. If NULL (the default), then this parameter is set to the model's training data.

A n x p matrix where each row is an observation. If NULL (the default), then this parameter is set to the model's training data.

TRUE (the default) will only return proximity information for trees where each observation is out-of-bag. If X1 is not NULL, this parameter has no effect on output.

TRUE (the default) will only return proximity information for trees where each observation is out-of-bag. If X2 is not NULL, this parameter has no effect on output.

Print out periodic messages on progress of the computation of the proximity information. The default is TRUE.

The fit model to test against permuted training data.

The covariates to permute as a vector of indices or names. If multiple covariates they are permuted as a single block to not break collinearity. If this parameter is specified as NULL, we test all the covariates by permuting y instead. Default is NULL.

How many different YARF models should be built - one for each different permutation of the variables tests. The is the resolution of the test. The default is 100.

Plot a histogram of the permutation samples with the original sample displayed? Default is TRUE.

The YARF model in which to query

This controls whether or not the function plots the interaction strengths. If so, it plots in order of strength from highest to lowest. Default is TRUE.

Should we only show interactions between different variables? Default is FALSE to show interactions of variables with themselves.

How many first order interactions to plot? All would be p-choose-2 which can be large. The default is 50. Use Inf to plot all p-choose-2.

The YARF model in which to query

The YARF model whose trees are to be illustrated

A vector of the tree indices (a subset of 1...num_trees)

Illustrate the tree only up to this depth. Default is -1 indicating no restriction.

A vector with 3 elements in 0, 1, ..., 255 specifying the RGB color of the background. Default is c(255, 255, 255) for white.

A vector with 3 elements in 0, 1, ..., 255 specifying the RGB color of the lines. Default is c(100, 100, 100) for grey.

A vector with 3 elements in 0, 1, ..., 255 specifying the RGB color of the text. Default is c(0, 0, 0) for black.

The font family of the text in the illustration. This string is passed in as the first argument in a new java.awt.Font object. You can find a list of legal arguments in the Java 8 documentation. Default is arial.

The font size in the illustrations in pixels. Default is 12.

The margin in pixels of the entire illustration (equal on all sides). Default is 70.

Estimate as to the size of the character width in pixels. Differs based on fonts.

The length of half of a split in pixels. Default is 20.

The length of the depth of a split in pixels. Default is 100.

A custom javascript function which creates a special custom message to be printed at split nodes (see below). The default is NULL which prints no special message.

function printAtSplitNode(node){ //node is of type YARF.YARFNode

...

return message; //a string

}

A custom javascript function which creates a special custom message to be printed at leaf nodes (see below). The default is NULL which prints no special message.

function printAtLeaf(node){ //node is of type YARF.YARFNode

...

return message; //a string

}

The image format of the resulting illustration. Legal values are "jpg", "gif", "png". Default is "png".

In the illustrations of split rules, set this to TRUE to use the names in the data frame. and FALSE to use "X_1", "X_2", etc. Default is TRUE.

The name of the file. Note that "_00t.png" will be added where "t" is the tree index and "00" is an appropriate number of leading zeroes. Default is "yarf_mod_tree".

Should the first file be opened by the operating system? Default is FALSE.

The training data frame

The vector of training responses

An already-fit model that you wish to test if there is any predictive power left over. This model's class must respond to the "predict" S3 method. The default is NULL which indicates the standard linear model (missingness not allowed there).

The resolution of the test. See cov_importance_test. The default is 100.

Plot the result as a histogram. Default is TRUE.

Additional parameters to be passed to YARF.

A YARF model object.

A n* x p matrix where each row is an observation you wish to predict for. Here, p is the same length as the original training data.

Other methods passed to predict

A YARF model object.

A n* x p data frame where each row is an observation you wish to obtain prediction information. Here, p is the same length as the original training data and the column names match the original training data. If a matrix object is passed in, it is automatically cast to a data frame. This may not always give you what you want. If NULL (the default), then this parameter is set to the model's training data.

If X is NULL, setting this to TRUE (the default) will only return node information for trees where each observation is out-of-bag. If X is not NULL, this parameter has no effect on output.

A YARF model object.

Other methods passed to print

A YARF model object.

A n* x p data frame where each row is an observation. If NULL (the default), then this parameter is set to the model's training data.

Script ...

The YARF model whose trees are to be pruned.

A custom javascript function (see below) which determines the pruning of a node (i.e. deletes the node's children and sets the node to a leaf and then sets a y_hat using the assign function).

function pruneIf(node){ //node is of type YARF.YARFNode

...

return boolean //true if this node is to be pruned }

An array of indices which specify which of the trees to prune. The default is NULL which means all of them.

The model in which to query

Either "trees" which counts the number of trees where each variable was used one (or more) times or "splits" which counts the number of times each variable was used across the whole model.

The model in which to query

Either "trees" which computes the proportion of trees where each variable was used one (or more) times or "splits" which computes the proportion of times each variable was used across the whole model.

The number of cores to use

The first string

The second string

A YARF model object.

Other methods passed to summary

The information returned from compute_raw_proximity_info().

The data frame of training data

The vector of training responses which is either numeric (for regression or factor (for classification).

The data frame of training data where the last column is responses

Other data that is used in the training but the RF doesn't split on it

If TRUE, missingness in the response variable, y, is allowed. If the missings are not handled (somehow) in the custom functions, YARF will crash!! Default is FALSE.

The number of trees in the RF. Default is NULL which sets the value to 500 if fit_until_convergence is set to FALSE and 10,000 if fit_until_convergence is set to TRUE.

If the model is fit asynchronously via the wait parameter being set to TRUE, this number represents a vague contract between the user and the software as the model fitting can be (a) halted by the user or the model can be (b) set to converge automatically (by setting the fit_until_convergence parameter to TRUE). In the latter case, this parameter functions as the maximum number of trees that can be created and thus, it is recommended that this number is very large, hence the default of 10,000.

A list with keys 1, 2, ..., num_trees where each value is the indices of the training data you wish to use for each tree. The default is NULL indicating the default algorithm of sampling 1, ..., n with replacement (i.e. the non-parametric bootstrap default). Needless to say indices specified here will not be part of the out-of-bag collection of indices. You may specify more than enough (i.e. more than num_trees) elements in this list. Only the first num_trees will be used. No warning message will be displayed if you include too many. If this parameter is specified, there is no need to specify n_max_per_tree.

An upper limit on the number of observations used to build each tree. If this parameter is specified, there is no need to specify bootstrap_indices. Default is NULL for the entire dataset, n.

An optional list with keys 1,2,..., num_trees where each value is indices of the training data you wish to use in some custom way for each tree using a custom function. If not custom function is specified which makes use of this, it will be ignored. However, indices specified here will not be part of the out-of-bag collection of indices. If this is not your wish, the elements in the vectors specified in this list's values should be a subset of those in the values of bootstrap_indices. The default is NULL indicating you do not wish to specify any "other" data records.

The number of variables tried at every split. The default is NULL which indicates the out-of-box RF default which is floor(p / 3) for regression and floor(sqrt(p)) for classification. If you want to use all possible features, set this parameter to "all". If you want a custom function, leave this NULL and see next parameter.

The minimum number of observations in a node. YARF will stop splitting at this point. If NULL the out-of-the-box default of 5 for regression and 1 for classification will be used.

A custom javascript function which selects the variables to be greedily searched (see below) The default is NULL which employs the mtry argument. If you specify your function please randomize the order of the returned attributes to arbitrate ties.

function tryVars(node){ //node is of type YARF.YARFNode

...

return int_array //a subset of 0,...,p-1, indices indicating the variables to perform the exhaustive search on

}

A custom javascript function which selects the split values to be greedily searched in feature j. The default is NULL which employs the midpoints of all sorted values.

function tryVals(node, j){ //node is of type YARF.YARFNode and j is the feature number in 0,...,p-1

...

return double_array //a vector of split vals to greedily assess

}

A custom javascript function to be used to calculate nodesize (see below). The default is NULL where nodesize will be calculated as a static constant (see the nodesize argument).

function makeNodeIntoLeaf(node){ //node is of type YARF.YARFNode

...

return boolean //where true makes this node into a leaf

}

A custom cost calculation for a potential node (when considering a split) in Javascript (see below). The default is NULL which means the out-of-the-box default of sum of squared error relative to the sample average (if regression) and sum of entropy (if classification). You may find it convenient to also made a node assignment here. If so, make sure you specify the node assignment function as a blank function (not NULL).

function nodeCost(node){ //node is of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom cost calculation in Javascript for an entire split considering both the putative left and right children nodes (see below). The default is NULL which means the out-of-the-box default for Random Forests which is sum of left and right nodes' costs for regression and average of left and right nodes' cost (relative to the number of observations in each node).

function totalChildrenCost(leftNode, rightNode){ //both nodes are of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom node assignment function in Javascript (see below). This function is run after RF greedily finds the "lowest cost" split. The default is NULL corresponding to the sample average of the node responses in regression or the modal class during classification.

function assignYhatToNode(node){ //node is of type YARF.YARFNode

...

return double //assigned as this node's predicted value ("y_hat")

}

A custom function in Javascript which is executed after a node is given birth to. The default is NULL which implies nothing special is done, the Random Forest default. This is particularly useful to record extra information in the node (e.g. by writing a hash to the other_info field in the node).

function nodeAfterNodeBirth(node){ //node is of type YARF.YARFNode

...

}

Custom Javascript code that are always in scope when running all your custom methods. The default is NULL for no shared scripts.

Use the "missing-incorporated-in-attributes" strategy to fit data with missingness. The default is TRUE.

Replace missing predictor values with their training-column means before fitting.

What rule should be used for missing data seen in testing for nodes that did not see missing data in training? Choices: "RANDOM" and "CONDITIONAL_ON_QUANTILE". "RANDOM" will assign a split rule for missing data at such nodes purely arbitrarily and with equal probability, i.e. 50 will condition this probability on the quantile of the split value amongst the observed (non-missing) data. For example, if the split value lies at the 75th percentile, then with probability .75, this node's rule for missing data will be to go left, and with probability .25 to go right. Note that for either of these options, the randomization happens at training time and thus during prediction the rule will be deterministic. The default is "RANDOM".

Should the YARF model be saved? The default is FALSE as this is costly in processing time and memory. This can only be set to TRUE if wait = TRUE. If TRUE, we will automatically serialize after other operations that add data (such as the OOB evaluation).

Set a random seed for reproducibility.

Should we hang R to wait for the YARF model to complete? The default is TRUE.

Should we also calculate the OOB error? Default is TRUE. Automatically is turned off if wait becomes FALSE.

Default is FALSE. If TRUE, then the wait parameter is set to FALSE regardless of the user-specified value. Then, the Random Forest model is fit until "convergence" as defined below:

After each tree is completed, the OOB cost is computed. If the cost decreases the algorithm does not engage. However, upon the first increase in cost, the model checks for convergence by calculating the average change in cost and the standard deviation of change in cost. If a 95% confidence interval of the average cost is within the window created by plus or minus tolerance (i.e. from zero), then the algorithm has "converged". Upon convergence, the model is stopped and a message is printed to the console. Note that specification of the oob_cost_calculation_script and a proper tolerance level is essential.

It is strongly recommended to set the num_trees parameter large, otherwise the model can halt before it has converged since num_trees represents an upper limit. If left unspecified, num_trees is set to 10,000 when fit_until_convergence set to TRUE.

If the user wishes to view the convergence in real time, we recommend the YARF_progress function. However, this function will lock the console.

This parameter will determine the out-of-bag cost of the RF forest model. Default is NULL. If it is NULL, then the cost will be 1-R^2 for regression models and misclassification error for classification models.

If non-null, an optional custom Javascript function which calculates the cost of a prediction given the true value of the prediction (see below). If is likely similar to cost_single_node_calc_script. It is recommended to share code between them by writing a function included in the shared_scripts argument which can be referenced when calculating OOB results.

function oobCost(y_hat, y){ //y_hat is the predicted value and y is the true value (both are of type double)

...

return double //where a larger number indicates a higher cost to the error between y and y_hat.

}

If this script is specified (i.e. the argument is non-NULL), this will also affect the output of the YARF_update_with_oob_results function. This script can be reset after the model is built if necessary.

This parameter is ignored unless fit_until_convergence is set to TRUE. If so, this controls the tolerance used when assessing the convergence of the RF model. Default is 0.1. If oob_cost_calculation_script is specified, this should also be specified as 0.1 may not be appropriate for a custom oob cost function.

Should we print out messages verbosely during construction? Default is FALSE.

Should we print out messages from Java? Default is FALSE.

The yarf model object such that YARF_update_with_oob_validation_results has been run before on a user-determined proportion of OOB samples as data validation.

The yarf model object

Frequency of messages in seconds. Default is 5 seconds. Minimum is 0.1. NULL indicates it is shown once and then it completes.

If non-null, it will plot a secondary window with only the trees in the last trail_trees number of progress updates defined by time_delay_in_seconds versus error which allows assessing the convergence more closely. Default is 50.

The YARF model with settings you wish to duplicate

The training data as a data frame. Default is NULL indicating the original training data.

The training responses as numeric / integer / factor vector. Default is NULL indicating the original training responses.

A YARF model object.

A n* x p data frame where each row is an observation you wish to predict for. Here, p is the same length as the original training data and new_data has the same column names. If a matrix object is passed in, it is automatically cast to a data frame. This may not always give you what you want.

The yarf model object

Should we print a message to console? Default is TRUE.

The yarf model object

The yarf model object

A custom javascript function which aggregates the predictions in the trees for one observations into one scalar prediction (see below).

function aggregateYhatsIntoOneYhat(y_hats, yarf){ //y_hats is an array of doubles of size num_trees //and yarf provides access to the entire random forest object if needed (of type YARF.YARF)

...

return double //this is the final predicted value aggregated from all tree predictions

}

The yarf model object

The yarf model object

An optional custom Javascript function which calculates the cost of a prediction given the true value of the prediction (see below). If is likely similar to cost_single_node_calc_script. It is recommended to share code between them by writing a function included in shared_scripts which is passed into YARF upon construction and can be referenced when calculating OOB results.

function oobCost(y_hat, y){ //y_hat is the predicted value and y is the true value (both are of type double)

...

return double //where a larger number indicates a higher cost to the error between y and y_hat.

}

The indices to compute OOB performance. This is generally a private argument for use by the validation / test functions.

The yarf model object such that YARF_update_with_oob_validation_results has been run before on a user-determined proportion of OOB samples as data validation.

The yarf model object

The percentage of the data to use for validation data (and the remaining is test data). The default is 80%.

The data frame of training data

The vector of training responses

The data frame of training data where the last column is responses

Other data that is used in the training but the RF doesn't split on it

If TRUE, missingness in the response variable, y, is allowed. If this is not handled in the custom functions, YARF will crash. Default is FALSE.

The number of trees in the RF. Default is NULL which sets the value to 500 if fit_until_convergence is set to FALSE and 10,000 if fit_until_convergence is set to TRUE.

If the model is fit asynchronously via the wait parameter being set to TRUE, this number represents a vague contract between the user and the software as the model fitting can be (a) halted by the user or the model can be (b) set to converge automatically (by setting the fit_until_convergence parameter to TRUE). In the latter case, this parameter functions as the maximum number of trees that can be created and thus, it is recommended that this number is very large, hence the default of 10,000.

A list with keys 1, 2, ..., num_trees where each value is the indices of the training data you wish to use for each tree. The default is NULL indicating the default algorithm of sampling 1, ..., n with replacement (i.e. the non-parametric bootstrap default). Needless to say indices specified here will not be part of the out-of-bag collection of indices. You may specify more than enough (i.e. more than num_trees) elements in this list. Only the first num_trees will be used. No warning message will be displayed if you include too many.

An upper limit on the number of observations used to build each tree. If this parameter is specified, there is no need to specify bootstrap_indices. Default is NULL for the entire dataset, n.

An optional list with keys 1,2,..., num_trees where each value is indices of the training data you wish to use in some custom way for each tree using a custom function. If not custom function is specified which makes use of this, it will be ignored. However, indices specified here will not be part of the out-of-bag collection of indices. If this is not your wish, the elements in the vectors specified in this list's values should be a subset of those in the values of bootstrap_indices. The default is NULL indicating you do not wish to specify any "other" data records.

The minimum number of observations in a node. YARF will stop splitting at this point. If NULL the out-of-the-box default of 5 for regression and 1 for classification will be used.

A custom JavaScript function that selects candidate variables at each split.

A custom javascript function which selects the split values to be greedily searched in feature j. The default is NULL which employs the midpoints of all sorted values.

function tryVals(node, j){ //node is of type YARF.YARFNode and j is the feature number in 0,...,p-1

...

return double_array //a vector of split vals to greedily assess

}

A custom javascript function to be used to calculate nodesize (see below). The default is NULL where nodesize will be calculated as a static constant (see the nodesize argument).

function makeNodeIntoLeaf(node){ //node is of type YARF.YARFNode

...

return boolean //where true makes this node into a leaf

}

A custom cost calculation for a potential node (when considering a split) in Javascript (see below). The default is NULL which means the out-of-the-box default of sum of squared error relative to the sample average (if regression) and sum of entropy (if classification). You may find it convenient to also made a node assignment here. If so, make sure you specify the node assignment function as a blank function (not NULL).

function nodeCost(node){ //node is of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom cost calculation in Javascript for an entire split considering both the putative left and right children nodes (see below). The default is NULL which means the out-of-the-box default for Random Forests which is sum of left and right nodes' costs for regression and average of left and right nodes' cost (relative to the number of observations in each node).

function totalChildrenCost(leftNode, rightNode){ //both nodes are of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom node assignment function in Javascript (see below). This function is run after RF greedily finds the "lowest cost" split. The default is NULL corresponding to the sample average of the node responses in regression or the modal class during classification.

function assignYhatToNode(node){ //node is of type YARF.YARFNode

...

return double //assigned as this node's predicted value ("y_hat")

}

A custom function in Javascript which is executed after a node is given birth to. The default is NULL which implies nothing special is done, the Random Forest default. This is particularly useful to record extra information in the node (e.g. by writing a hash to the other_info field in the node).

function nodeAfterNodeBirth(node){ //node is of type YARF.YARFNode

...

}

Custom Javascript code that are always in scope when running all your custom methods. The default is NULL for no shared scripts.

Use the "missing-incorporated-in-attributes" strategy to fit data with missingness. The default is TRUE.

Replace missing predictor values with their training-column means before fitting.

Should the YARF model be saved? The default is FALSE as this is costly in processing time and memory. This can only be set to TRUE if wait = TRUE. If TRUE, we will automatically serialize after other operations that add data (such as the OOB evaluation).

Set a random seed for reproducibility.

Should we hang R to wait for the YARF model to complete? The default is TRUE.

Should we also calculate the OOB error? Default is TRUE. Automatically is turned off if wait becomes FALSE.

Fit trees until the out-of-bag convergence criterion is met.

This parameter will determine the out-of-bag cost of the RF forest model. Default is NULL. If it is NULL, then the cost will be 1-R^2 for regression models and misclassification error for classification models.

If non-null, an optional custom Javascript function which calculates the cost of a prediction given the true value of the prediction (see below). If is likely similar to cost_single_node_calc_script. It is recommended to share code between them by writing a function included in the shared_scripts argument which can be referenced when calculating OOB results.

function oobCost(y_hat, y){ //y_hat is the predicted value and y is the true value (both are of type double)

...

return double //where a larger number indicates a higher cost to the error between y and y_hat.

}

If this script is specified (i.e. the argument is non-NULL), this will also affect the output of the YARF_update_with_oob_results function. This script can be reset after the model is built if necessary.

This parameter is ignored unless fit_until_convergence is set to TRUE. If so, this controls the tolerance used when assessing the convergence of the RF model. Default is 0.1. If oob_cost_calculation_script is specified, this should also be specified as 0.1 may not be appropriate for a custom oob cost function.

Should we print out messages verbosely during construction? Default is FALSE.

Should we print out messages from Java? Default is FALSE.

The data frame of training data

The vector of training responses

The data frame of training data where the last column is responses

Other data that is used in the training but the RF doesn't split on it

If TRUE, missingness in the response variable, y, is allowed. If this is not handled in the custom functions, YARF will crash. Default is FALSE.

A list with keys 1, 2, ..., num_trees where each value is the indices of the training data you wish to use for each tree. The default is NULL indicating the default algorithm of sampling 1, ..., n with replacement (i.e. the non-parametric bootstrap default). Needless to say indices specified here will not be part of the out-of-bag collection of indices. You may specify more than enough (i.e. more than num_trees) elements in this list. Only the first num_trees will be used. No warning message will be displayed if you include too many.

An optional list with keys 1,2,..., num_trees where each value is indices of the training data you wish to use in some custom way for each tree using a custom function. If not custom function is specified which makes use of this, it will be ignored. However, indices specified here will not be part of the out-of-bag collection of indices. If this is not your wish, the elements in the vectors specified in this list's values should be a subset of those in the values of bootstrap_indices. The default is NULL indicating you do not wish to specify any "other" data records.

The number of variables tried at every split. The default is NULL which indicates the out-of-box RF default which is floor(p / 3) for regression and floor(sqrt(p)) for classification. If you want a custom function, leave this NULL and see next parameter.

The minimum number of observations in a node. YARF will stop splitting at this point. If NULL the out-of-the-box default of 5 for regression and 1 for classification will be used.

A custom javascript function which selects the variables to be greedily searched (see below) The default is NULL which employs the mtry argument. If you specify your function please randomize the order of the returned attributes to arbitrate ties.

function tryVars(node){ //node is of type YARF.YARFNode

...

return int_array //a subset of 0,...,p-1, indices indicating the variables to perform the exhaustive search on

}

A custom javascript function which selects the split values to be greedily searched in feature j. The default is NULL which employs the midpoints of all sorted values.

function tryVals(node, j){ //node is of type YARF.YARFNode and j is the feature number in 0,...,p-1

...

return double_array //a vector of split vals to greedily assess

}

A custom javascript function to be used to calculate nodesize (see below). The default is NULL where nodesize will be calculated as a static constant (see the nodesize argument).

function makeNodeIntoLeaf(node){ //node is of type YARF.YARFNode

...

return boolean //where true makes this node into a leaf

}

A custom cost calculation for a potential node (when considering a split) in Javascript (see below). The default is NULL which means the out-of-the-box default of sum of squared error relative to the sample average (if regression) and sum of entropy (if classification). You may find it convenient to also made a node assignment here. If so, make sure you specify the node assignment function as a blank function (not NULL).

function nodeCost(node){ //node is of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom cost calculation in Javascript for an entire split considering both the putative left and right children nodes (see below). The default is NULL which means the out-of-the-box default for Random Forests which is sum of left and right nodes' costs for regression and average of left and right nodes' cost (relative to the number of observations in each node).

function totalChildrenCost(leftNode, rightNode){ //both nodes are of type YARF.YARFNode

...

return double //where a higher number indicates a higher cost

}

A custom node assignment function in Javascript (see below). This function is run after RF greedily finds the "lowest cost" split. The default is NULL corresponding to the sample average of the node responses in regression or the modal class during classification.

function assignYhatToNode(node){ //node is of type YARF.YARFNode

...

return double //assigned as this node's predicted value ("y_hat")

}

A custom function in Javascript which is executed after a node is given birth to. The default is NULL which implies nothing special is done, the Random Forest default. This is particularly useful to record extra information in the node (e.g. by writing a hash to the other_info field in the node).

function nodeAfterNodeBirth(node){ //node is of type YARF.YARFNode

...

}

Custom Javascript code that are always in scope when running all your custom methods. The default is NULL for no shared scripts.

Use the "missing-incorporated-in-attributes" strategy to fit data with missingness. The default is TRUE.

Replace missing predictor values with their training-column means before fitting.

Should the YARF model be saved? The default is FALSE as this is costly in processing time and memory. This can only be set to TRUE if wait = TRUE. If TRUE, we will automatically serialize after other operations that add data (such as the OOB evaluation).

Set a random seed for reproducibility.

Should we also calculate the OOB error? Default is TRUE. Automatically is turned off if wait becomes FALSE.

This parameter will determine the out-of-bag cost of the RF forest model. Default is NULL. If it is NULL, then the cost will be 1-R^2 for regression models and misclassification error for classification models.

If non-null, an optional custom Javascript function which calculates the cost of a prediction given the true value of the prediction (see below). If is likely similar to cost_single_node_calc_script. It is recommended to share code between them by writing a function included in the shared_scripts argument which can be referenced when calculating OOB results.

function oobCost(y_hat, y){ //y_hat is the predicted value and y is the true value (both are of type double)

...

return double //where a larger number indicates a higher cost to the error between y and y_hat.

}

If this script is specified (i.e. the argument is non-NULL), this will also affect the output of the YARF_update_with_oob_results function. This script can be reset after the model is built if necessary.

Should we print out messages verbosely during construction? Default is FALSE.

Should we print out messages from Java? Default is FALSE.

The training data as a data frame.

The training responses as numeric / integer / factor vector.

The test data as a data frame (optional, default NULL for not specified).

The maximum number of iterations before stopping (default is 10, as in package missForest).

The algorithm quits after imputation errors increase once. Default is TRUE but if not, the algorithm continues all maxiter runs and then returns the final imputations.

Prints out messages by iteration. Default is TRUE.

Settings to pass to YARF during the iterations.

The covariate data frame

The binary response vector

The rate to plot on the x-axis of the ROC like plot as a 3-character case-insensitive string. The default is "FPR" but it canbe any of the following:

"tpr" - true positive rate (AKA "sensitivity" or "recall") i.e the number predicted positive among those truly positive "fnr" - false negative rate (AKA "miss rate") i.e the number predicted negative among those truly positive "fpr" - false positive rate (AKA "fall-out") i.e. the number predicted positive among those truly negative "tnr" - true negative rate (AKA "specificity") i.e. the number predicted negative among those truly negative "ppv" - positive predictive value i.e. the number truly positive among those predicted positive "fdr" - false discovery rate i.e. the number truly negative among those predicted positive "fomr" - false omission rate i.e. the number truly positive among those predicted negative "npv" - negative predictive value i.e. the number truly negative among those predicted negative

The rate to plot on the y-axis of the ROC like plot as a 3-character string. The default is "TPR" but it can be any of those listed in the documentation for the x_axis parameter. Must be different from x-axis.

Relative multiplier controlling the resampled class proportions.

Smallest positive-class proportion considered while tracing the curve.

Target spacing between adjacent metric values.

Recursion stops when candidate class proportions differ by less than this value.

Whether to make an additional pass emphasizing resolution on the y-axis.

Whether to draw intermediate and final ROC-like plots.

Whether to print progress information.

Additional arguments passed to YARF construction. This cannot include bootstrap_indicies.