setfis

(To be removed) Set fuzzy system properties

setfis will be removed in a future release. Set fuzzy inference system properties using dot notation instead. For more information, see Compatibility Considerations.

Syntax

fis = setfis(fis,fisPropName,fisPropVal)
fis = setfis(fis,varType,varIndex,varPropName,varPropVal)
fis = setfis(fis,varType,varIndex,'mf',mfIndex,mfPropName,mfPropVal)

Description

The command setfis can be called with three, five, or seven input arguments, depending on whether you want to set a property of the entire FIS structure, for a particular variable belonging to that FIS structure, or for a particular membership function belonging to one of those variables. The arguments are:

• fis — FIS structure in the MATLAB^® workspace.

• varType — Variable type, specified as either 'input' or 'output'.

• varIndex — Variable index, specified as a positive integer.

• mfIndex — Membership function index, specified as a positive integer.

• fisPropName — FIS property you want to set, specified as one of the following:

  • 'name'

  • 'type'

  • 'andmethod'

  • 'ormethod'

  • 'impmethod'

  • 'aggmethod'

  • 'defuzzmethod'

• fisPropVal — New value of the FIS property you want to set, specified as a character vector or string.

• varPropName — Variable property you want to set, specified as either 'name' or 'range'.

• varPropVal — New value of the variable property you want to set, specified as a character vector or string (for 'name'), or a two-element row vector (for 'range').

• mfPropName — Membership function property you want to set, specified as either 'name', 'type', or 'params'.

• mfPropVal — New value of the membership function property you want to set, specified as a character vector or string (for 'name' or 'type'), or a numerical row vector (for 'params').

Examples

Set Fuzzy Inference System Properties

Load a fuzzy inference system.

fis = readfis('tipper');

Set the defuzzification method to the bisector method.

fis = setfis(fis,'defuzzmethod','bisector');

View the defuzzification method of the updated FIS.

getfis(fis,'defuzzmethod')

ans = 
'bisector'

Set Variable Properties in FIS

Load fuzzy inference system.

fis = readfis('tipper');

Set the name of the first input variable to 'help'.

fis = setfis(fis,'input',1,'name','help');

View the name of the variable in the updated system.

getfis(fis,'input',1,'name')

ans = 
'help'

Set Membership Function Properties in FIS

Load a fuzzy inference system.

fis = readfis('tipper');

Change the type of the second membership function of the first input variable to a triangular membership function.

fis = setfis(fis,'input',1,'mf',2,'type','trimf');

When changing the type of a membership function, you must also set the parameters accordingly. To convert the original Gaussian membership function parameters to triangular membership function parameters, use the mf2mf command.

gaussParams = getfis(fis,'input',1,'mf',2,'params');
triParams = mf2mf(gaussParams,'gaussmf','trimf');

Set the membership function parameters to the converted values.

fis = setfis(fis,'input',1,'mf',2,'params',triParams);

View the updated membership function properties.

getfis(fis,'input',1,'mf',2)

ans = struct with fields:
      Name: 'good'
      Type: 'trimf'
    params: [1.4680 5 8.5320]

Compatibility Considerations

setfis will be removed

Not recommended starting in R2018b

setfis will be removed in a future release. Set fuzzy inference system properties using dot notation instead. There are differences between these approaches that require updates to your code.

Update Code

This table shows some typical usages of setfis for setting fuzzy inference system properties and how to update your code to use dot notation instead.

If your code has this form:	Use this code instead:
fis = setfis(fis,'andmethod','prod')	fis.AndMethod = 'prod'
fis = setfis(fis,'input',1,... 'name','service')	fis.Inputs(1).Name = "service"
fis = setfis(fis,'input',2,... 'mf',1,... params,[5 10 15])	fis.Inputs(2).MembershipFunctions(1).Parameters = ... [5 10 15]

Previously, fuzzy inference systems were represented as structures. Now, fuzzy inference systems are represented as objects. Fuzzy inference system object properties have different names than the corresponding structure fields. For more information on fuzzy inference system objects, see mamfis and sugfis.