expCreateFromTable

Create an exreport experiment from a tabular representation

Usage

expCreateFromTable(data, output, name, parameters = list(),
  respectOrder = FALSE)

Arguments

data
Input tabular data satisfying the previous constraints.
output
String indicating the name of the output that the table values represent.
name
A string which will identify the experiment in the report.
parameters
A list of strings containing the names and values for the static configuration of the algorithm. The name of each element of the list will correspond with the name of a parameter and the element with the value asigned.
respectOrder
A logical parameter which indicates if the order of the elements of the method and problem columns must be respected by appearance or ordered alphabeticaly. It affects to the look of data representations.

Value

A new exreport experiment object.

Description

Create an exreport experiment object from a tabular representation. The input data must be a table having methods as rows and problems as columns. The values in such table correspond to a particular output. The resulting experiment can be characterized with static parameters.

Examples

# We generate a data frame where the methods are rows and the problems columns # from the wekaExperiment problem. (This is only an example, normally you # would prefer to load a proper experiment and process it.) library(reshape2) df <- dcast(wekaExperiment[wekaExperiment$featureSelection=="no",], method ~ problem, value.var="accuracy", fun.aggregate = mean) # We can create it and parametrice accordingly: expCreateFromTable(df, output="accuracy", name="weka")
#Experiment name: weka #method: J48, NaiveBayes, OneR, RandomForest #problem: anneal, audiology, balance-scale, car, glass, horsecolic, hypothyroid, ionosphere, liver-disorders, lymph, primary-tumor, soybean, vehicle, vote, vowel #parameters: *) *) #outputs: accuracy
# Optionally we can set a fixed value for parameters, and ordered by appearance: expCreateFromTable(df, output="accuracy", name="weka", parameters=list(featureSelection = "no"), respectOrder=TRUE)
#Experiment name: weka #method: J48, NaiveBayes, OneR, RandomForest #problem: anneal, audiology, balance-scale, car, glass, horsecolic, hypothyroid, ionosphere, liver-disorders, lymph, primary-tumor, soybean, vehicle, vote, vowel #parameters: *) featureSelection [no] #outputs: accuracy

Back to top

Built by staticdocs. Styled with bootstrap.