No rules fired for Output 1. Defuzzified output value set to its mean range value 2.3873.

Someone knows the cause of this warning? and how to solve it?

Warning: No rules fired for Output 1. Defuzzified output value set to its mean range value 2.3873.

> In throwWarning (line 18)

In evalfis (line 93)

Here is my code,

subset = xlsread ('entrenamientoDrastic.csv', 1, 'A1:G420660');

nit = xlsread('nt2006.csv', 1, 'A1:WQ684');

nit = transpose(nit);

nit = reshape(nit,[], 1);

subset(:,7) = ((subset(:, 7)).*nit)/145;

subset = [subset nit];

[subset, ia, ic] = unique(subset,'rows','stable');

%rng('default');

% Cross varidation (train: 70%, test: 30%)

cv = cvpartition(size(subset,1),'HoldOut',0.3);

idx = cv.test;

% Separate to training and test data

train = subset(~idx,:);

test = subset(idx,:);

train_input = train(:, 1:6);

train_output = train(:,7);

test_input = test(:, 1:6);

test_output = test(:, 7);

opt = genfisOptions('FCMClustering','FISType','mamdani');

opt.NumClusters = 10;

opt.Exponent = 8.5;

%opt.Verbose = 2.1;

opt.MaxNumIteration = 100;

opt.MinImprovement = 1e-5;

fis = genfis(train_input,train_output,opt);

% Evaluate FIS

nit = test(:,8);

actY = evalfis(test_input,fis);

% Calculate RMSE

del = actY - test_output;

rmse = sqrt(mean(del.^2));

[rho,pval] = corr(nit,actY,'Type','Spearman');

The rmse gives an Inf value.

Answer 1

[Sam Chak]

Hi @Eliane Fernández

I replicated the Warning message in the example below. It is easier to understand why the warning occurs by examining the Rule Inference diagram. When the input is 0, it does not belong to any of the input fuzzy sets, and therefore, no rule is triggered. By default, the FIS will automatically compute the mean output range value.

fis = mamfis;

% Fuzzy Input #1
fis = addInput(fis, [-1 +1], 'Name', 'in1');
fis = addMF(fis, 'in1', 'linzmf', [-1 -1/4], 'Name', 'N');
fis = addMF(fis, 'in1', 'linsmf', [+1/4 +1], 'Name', 'P');

% Fuzzy Output
fis = addOutput(fis,  [-2.3873*2 2.3873*4], 'Name', 'out');     % <-- output range
fis = addMF(fis, 'out', 'trimf', [-2.3873*2 0 2.3873*2], 'Name', 'N');
fis = addMF(fis, 'out', 'trimf', [0 +2.3873*2 2.3873*4], 'Name', 'P');

% Fuzzy Rules
rules = [
    "in1==N  => out=N"
    "in1==P  => out=P"
    ];
fis = addRule(fis, rules);

evalfis(fis, 0)

Warning: No rules fired for Output 1. Defuzzified output value set to its mean range value 2.3873.

ans = 2.3873

figure
plotrule(fis)