Fuzzy Rule Interpolation Matlab Toolbox

Sparse Fuzzy Model Identification Matlab ToolBox

Download
The current version (FRIT v1.1.9) of the FRI toolbox is available for download free of charge under GNU General Public License .


If you use the toolbox please refer to it as follows:

Z. C. Johanyák, D. Tikk, S. Kovács, K. W. Wong: "Fuzzy Rule Interpolation Matlab Toolbox – FRI Toolbox," Proc. of the IEEE World Congress on Computational Intelligence (WCCI'06), 15th Int. Conf. on Fuzzy Systems (FUZZ-IEEE'06), July 16--21, 2006, Vancouver, BC, Canada, pp. 1427-1433, Omnipress.
[Bibtex]  [ draft pdf ]
The current version (SFMI v1.1.1) of the SFMI toolbox is available for download free of charge under GNU General Public License .


If you use the toolbox please refer to it as follows:

Johanyák, Zs. Cs.: Sparse Fuzzy Model Identification Matlab Toolbox - RuleMaker (SFMI) Toolbox, IEEE 6th International Conference on Computational Cybernetics, November 27-29, 2008, Stara Lesná, Slovakia, pp. 69-74. [draft version]

Online documentation
An online HTML documentation and a dependency graph of FRIT's functions is also available here.


An online HTML documentation and a dependency graph of SFMI's functions is also available here.



Installation instructions
Unzip the files from the file FRIT_ToolBox.zip. It creates a folder named "FRIT_ToolBox_1_1_9". Add this directory and its subfolders to the Matlab search path (File menu/Seth Path...-->Add with Subfolders).
Unzip the files from the file SFMI_ToolBox.zip. It creates a folder named "SFMI_ToolBox_1_1_1". Add this directory and its subfolders to the Matlab search path (File menu/Seth Path...-->Add with Subfolders).

Starting the program
The individual FRI methods can be reached as functions that can be invoked with proper parameters. However, there are two easy-to-use GUI programs that help you to try the methods. They can be started by the command 

>>FRIT
and 
>>TestIt

in the Matlab Command Window. Enjoy the program.

The individual functions can be reached that can be invoked with proper parameters. However, there is an easy-to-use GUI program that helps you trying the methods. They can be started by the command 

>>SFMI

in the Matlab Command Window. Enjoy the program.



Versions

FRIT v1.1.9

FRIT v1.1.9 has been released on 26.03.2009. New features:

  • Some minor corrections in different functions

FRIT v1.1.7

FRIT v1.1.7 has been released on 05.09.2008. New features:

  • The sets are displayed with different colors
  • Minor bug in VEIN corrected

FRIT v1.1.6

FRIT v1.1.6 has been released on 04.08.2008. New features:

  • Some minor corrections in different functions

FRIT v1.1.5

FRIT v1.1.5 has been released on 04.07.2008. New features:

  • New, optimzed defuzzification for piecewise linear fuzzy sets
  • More perspicuous displaying of antecedents and consequents
  • A method to 'normalize' abnormal conclusions with a convex hull
  • The method KHstabilized has been improved to have the ability to extrapolate
  • The method KHstabilized has been optimized for time complexity
  • Some minor corrections in different functions

FRIT v1.1.4

FRIT v1.1.4 has been released on 30.09.2007. New features:

  • Enhancements of the documentation
  • 3D representation of the rules in case of SISO systems with trapezoidal shaped fuzzy sets
  • 3D representation of the rule antecedents in case of 2IN MISO systems with trapezoidal shaped fuzzy sets
  • Several minor corrections in different functions

FRIT v1.1.3

FRIT v1.1.3 has been released on 02.08.2007. New features:

  • Implementation of the method VEIN [Method description].
  • A GUI based system tester is included as well. It can be started by the command TestIt.
  • Several minor corrections in different functions

FRIT v1.1.2

FRIT v1.1.2 has been released on 23.02.2007. New features:

  • Extension of the method FIVE to the case of fuzzy observations.
  • Some improvements related to the ability of handling crisp fuzzy sets (infinite scaling functions) has been added to the original version of FIVE.
  • Some minor corrections in the implementation of the methods FRIPOC.
  • A beta version of a rule extraction and system tuning modul is also included. It can be started by the command RuleMaker.
  • A beta version of a system tester is included as well. It can be started by the command TestIt.
  • Enhancements of the documentation.

Release 1.1.1

R 1.1.1 has been released on 12.11.2006. New features:

  • New XML-based data file format for the FIS and observation data. Thus it is not required the presence of the Matlab Fuzzy Logic Toolbox for the read of the observation and system (FIS) data.
  • The old data files still can be loaded if you dispose of the for required FLTB files: getfis.m, parsrule.m, setfis.m and showrule.m
  • The results of the calculation can be saved in XML format
  • The preferences of the interpolation methods are shown in an edit box in the main window
  • The preferences can be set in a dialog box invoked by the menu item Interpolation/Preferences
  • Enhancements of the documentation

Release 1.1.0

R 1.1.0 has been released on 30.10.2006. It contains a new GUI and a new FRI technique called LESFRI (Fuzzy Rule Interpolation by the Least Squares Method - Johanyák & Kovács).

Release 1.0.1

R 1.0.1 has been released on 31.08.2006. It contains some minor corrections of the method VKK and the graphical representation. The method FRIPOC (Fuzzy rule interpolation based on polar cuts - Johanyák & Kovács) has been also included.

Release 1.0.0

Version R 1.0 of the Fuzzy Rule Interpolation Toolbox has been released on 06.06.2006. It is a collection of Matlab functions implementing interpolation based fuzzy inference techniques. It supports the FRI methods KH, stabilized version of the KH, MACI, IMUL, CRF, FIVE, VKK, GM with SCM, FERI, and FPL, and GM with FEAT-p, FERI, and FPL.

SFMI v1.1.1

Version 1.0.1 of the Fuzzy RuleMaker Matlab Toolbox has been released on 26.03.2009. New features:

  • The name of the model identification toolbox has benn changed to Sparse Fuzzy Model Identification ToolBox - SFMI ToolBox
  • Some minor corrections in different functions

RuleMaker v1.0.1

Version 1.0.1 of the Fuzzy RuleMaker Matlab Toolbox has been released on 08.09.2008. New features:

  • 3D representation of the antecedent space if the number of input dimensions is 2 or 3.
  • 3D representation of the rule base if the number of input dimensions is 1 or 2.

RuleMaker v1.0.0

Version 1.0.0 of the Fuzzy RuleMaker Matlab Toolbox has been released on 28.09.2007. It implements three automatic fuzzy system generation methods

  • ACP - a method based on fuzzy clustering and a "hill climbing" type parameter identification algorithm
  • RBE-DSS - a method based on creation of a raw (initial) fuzzy system with two rules describing the minimum and maximum output, and a "hill climbing" type parameter identification algorithm combined with the iterative extension of the rule base. The new fuzzy sets are determined using predetermined (default) set shapes [ Method description].
  • RBE-SI - a method based on creation of a raw (initial) fuzzy system with two rules describing the minimum and maximum output, and a "hill climbing" type parameter identification algorithm combined with the iterative extension of the rule base. The new fuzzy sets are determined using set interpolation [ Method description].




Installation Instructions
Download
Starting the program
Online documentation
FRIT
FRIT v1.1.9
FRIT v1.1.7
FRIT v1.1.6
FRIT v1.1.5
FRIT v1.1.4
FRIT v1.1.3
FRIT v1.1.2
Release 1.1.1
Release 1.1.0
Release 1.0.1
Release 1.0.0
SFMI
SFMI v1.1.1
RuleMaker v1.0.1
RuleMaker v1.0.0