microeco (version 0.7.1)

trans_classifier: Create trans_classifier object for machine-learning-based model prediction.

Description

This class is a wrapper for methods of machine-learning-based classification models.

Author(s): Felipe Mansoldo and Chi Liu

Methods

Public methods

• trans_classifier$new()

• trans_classifier$cal_feature_sel()

• trans_classifier$cal_split()

• trans_classifier$set_trainControl()

• trans_classifier$cal_train()

• trans_classifier$cal_feature_imp()

• trans_classifier$cal_predict()

• trans_classifier$plot_confusionMatrix()

• trans_classifier$cal_ROC()

• trans_classifier$plot_ROC()

• trans_classifier$clone()

Method new()

Create the trans_classifier object.

Usage

trans_classifier$new(
  dataset = NULL,
  x.predictors = "all",
  y.response = NULL,
  n.cores = 1
)

Arguments

dataset

the object of microtable Class.

x.predictors

default "all"; character string or data.frame; a character string represents selecting the corresponding data from microtable$taxa_abund; data.frame represents other customized data. See the following available options and description:

'all'

use all the taxa stored in microtable$taxa_abund

'Genus'

use Genus level table in microtable$taxa_abund, or other specific taxonomic rank

other input

must be a data.frame; It should be have the same format with the data.frame in microtable$taxa_abund, i.e. rows are features; cols are samples with same names in sample_table

y.response

default NULL; the response variable in sample_table.

n.cores

default 1; the CPU thread used.

Returns

data_feature and data_response in the object.

Examples

\donttest{
data(dataset)
t1 <- trans_classifier$new(
		dataset = dataset, 
		x.predictors = "all",
		y.response = "Group")
}

Method cal_feature_sel()

Perform feature selection.

Usage

trans_classifier$cal_feature_sel(
  boruta.maxRuns = 300,
  boruta.pValue = 0.01,
  boruta.repetitions = 4,
  ...
)

Arguments

boruta.maxRuns

default 300; maximal number of importance source runs; passed to the maxRuns parameter in Boruta function of Boruta package.

boruta.pValue

default 0.01; p value passed to the pValue parameter in Boruta function of Boruta package.

boruta.repetitions

default 4; repetition runs for the feature selection.

...

parameters pass to Boruta function of Boruta package.

Returns

optimized data_feature in the object.

Examples

\donttest{
t1$cal_feature_sel(boruta.maxRuns = 300, boruta.pValue = 0.01)
}

Method cal_split()

Split data for training and testing.

Usage

trans_classifier$cal_split(prop.train = 3/4)

Arguments

prop.train

default 3/4; the ratio of the dataset used for the training.

Returns

optimized data_feature in the object.

Examples

\donttest{
t1$cal_split(prop.train = 3/4)
}

Method set_trainControl()

Set trainControl for the following training.

Usage

trans_classifier$set_trainControl(method = "repeatedcv", ...)

Arguments

method

default 'repeatedcv'; the method used in trainControl function of caret package.

...

parameters pass to trainControl function of caret package.

Returns

trainControl in the object.

Examples

\dontrun{
t1$set_trainControl()
}

Method cal_train()

Run the training.

Usage

trans_classifier$cal_train(
  method = "rf",
  metric = "Accuracy",
  max.mtry = 2,
  max.ntree = 200,
  ...
)

Arguments

method

default "rf"; representing the random forest method; see method in train function of caret package.

metric

default "Accuracy"; see metric in train function of caret package.

max.mtry

default 2; maximum mtry.

max.ntree

default 200; Number of trees to grow; pass to the ntree parameter of randomForest function in randomForest package.

...

parameters pass to train function of caret package.

Returns

res_train in the object.

Examples

\dontrun{
t1$cal_train()
}

Method cal_feature_imp()

Get feature importance from the training model.

Usage

trans_classifier$cal_feature_imp(...)

Arguments

...

parameters pass to the evaluating function; If "rf" used, pass to randomForest::importance.

Returns

res_feature_imp in the object. One row for each predictor variable. The column(s) are different importance measures.

Examples

\dontrun{
t1$cal_feature_imp()
}

Method cal_predict()

Run the prediction.

Usage

trans_classifier$cal_predict(positive_class = NULL)

Arguments

positive_class

default NULL; see positive parameter in confusionMatrix function of caret package.

Returns

res_confusion_fit and res_confusion_stats stored in the object.

Examples

\dontrun{
t1$cal_predict()
}

Method plot_confusionMatrix()

Plot the cross-tabulation of observed and predicted classes with associated statistics.

Usage

trans_classifier$plot_confusionMatrix(
  plot_confusion = TRUE,
  plot_statistics = TRUE
)

Arguments

plot_confusion

default TRUE; whether plot the confusion matrix.

plot_statistics

default TRUE; whether plot the statistics.

Returns

ggplot object in the object.

Examples

\dontrun{
t1$plot_confusionMatrix()
}

Method cal_ROC()

Get ROC curve data and the performance data.

Usage

trans_classifier$cal_ROC(...)

Arguments

...

parameters pass to plot.performance function of ROCR package.

Returns

a list including res_perf, all_auc_perf and all_perf_table stored in the object.

Examples

\dontrun{
t1$cal_ROC()
}

Method plot_ROC()

Plot ROC curve.

Usage

trans_classifier$plot_ROC(
  color_values = RColorBrewer::brewer.pal(8, "Dark2"),
  ...
)

Arguments

color_values

default RColorBrewer::brewer.pal(8, "Dark2"); colors used in the plot.

...

parameters pass to geom_line function of ggplot2 package.

Returns

ggplot2 object.

Examples

\dontrun{
t1$plot_ROC(size = 1, alpha = 0.7)
}

Method clone()

The objects of this class are cloneable with this method.

Usage

trans_classifier$clone(deep = FALSE)

Arguments

deep

Whether to make a deep clone.

Examples

# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$new`
## ------------------------------------------------

# }
# NOT RUN {
data(dataset)
t1 <- trans_classifier$new(
		dataset = dataset, 
		x.predictors = "all",
		y.response = "Group")
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_feature_sel`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_feature_sel(boruta.maxRuns = 300, boruta.pValue = 0.01)
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_split`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_split(prop.train = 3/4)
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$set_trainControl`
## ------------------------------------------------

# }
# NOT RUN {
t1$set_trainControl()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_train`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_train()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_feature_imp`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_feature_imp()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_predict`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_predict()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$plot_confusionMatrix`
## ------------------------------------------------

# }
# NOT RUN {
t1$plot_confusionMatrix()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$cal_ROC`
## ------------------------------------------------

# }
# NOT RUN {
t1$cal_ROC()
# }
# NOT RUN {
## ------------------------------------------------
## Method `trans_classifier$plot_ROC`
## ------------------------------------------------

# }
# NOT RUN {
t1$plot_ROC(size = 1, alpha = 0.7)
# }