This section aims to explain how to use the applications. Before following the guidance, we need to prepare the following applications:

  • An application used for constructing an FRBS model. We provide the R package "frbs" that can be used to build the model. Furthermore, "frbs" allows to convert FRBS models to the frbsPMML format. It is available on CRAN at
  • Download an executable file of the "frbsJpmml" package. It is available on the Download menu.

In order to construct an FRBS model and export to frbsPMML, the following are steps that should be taken into account:

  1. Constructing an FRBS model: For generating an FRBS model, readers can refer to steps in the "frbs" manual. In "frbs", we provide more than 10 learning methods, such as ANFIS, SLAVE, etc. We can also construct an FRBS model using knowledge from human experts instead of data.
  2. Exporting the FRBS model to frbsPMML format: After obtaining the model, we need to export it to frbsPMML format. When using "frbs", there are two commands used for exporting to frbsPMML format:
    • "frbsPMML()": It is a main function used to convert a model to the frbsPMML format and then display it in the R environment. It can be useful when we want to check the format.
    • "write.frbsPMML()": It is used to write and save a model in the frbsPMML format to a file. The file has the extension .frbsPMML, and can be opened by any editors since basically it contains an XML text.   
  3. Importing frbsPMML format and prediction new data: In order to do prediction new data, we need to import the .frbsPMML file back to an FRBS model. Moreover, we also need to provide a file containing our testing data. For doing these jobs, we provide two following applications:
    • "frbs": For converting an frbsPMML format to an FRBS model, we execute "read.frbsPMML()". After obtaining an FRBS model as an R object, we call "predict()". For more detail, users can read frbs' manual.
    • "frbsJpmml": In this application, reading and prediction are done at the same time. So, we just need to supply two files: a frbsPMML file containing an FRBS model in frbsPMML format and a testing file. Then, we execute the executable file "frbsJpmml.jar".

In the following, we present one example and its detailed description. We are using the gas furnace dataset that is already included in the "frbs" package.

1. Constructing an FRBS model for the gas furnace dataset. In the R environment, we type the following commands:


 2. Exporting the FRBS model to frbsPMML format. In the R environment, we type the following commands:


3. Importing and prediction using "frbs". In the R environment, we type the following commands:


We can also import and predict using "frbsJpmml". We just execute the following on command prompt:



Readers can also download scripts, data, and output from the following links: