MATLAB: It is possible to train the data set of 100×15 in ANFIS

anfiserror in large data set training of anfisMATLABtraining anfis

Dear all,

I am a very new user for ANFIS classifier.So, may I know it is possible to train 100×15 size of data set in .mat file. I trained this training data set ,and i occured error like " create very large rule based system" . May I know what 's the problem is? Thanks all.

Best Answer

This issue is a fundamental limitation and not a bug.

The following Web site provides information about the "curse of dimensionality":

http://www.cs.nthu.edu.tw/~jang/anfisfaq.htm

Question 5 in this page, titled "How come GENFIS1 generates a huge number of fuzzy rules? What is the 'curse of dimensionality?'" is relevant for this issue.

One way to reduce the number of rules generated for a large dataset is to use the GENFIS2 method could be like:

in = train_data(:,1:37);
out = train_data(:,38);
fismat = genfis2(in,out,0.8); % generate an initial FIS with 80 rules instead of 728
fis = anfis(train_data,fismat); % takes long time but avoids 'out of memory' error

Additionally, for large datasets, it is not recommend that we execute multiple trainings in a loop. To manage memory usage, I suggest using a separate script file for each training and executing "clear classes" at the MATLAB prompt between executing each of the scripts.

Related Solutions

MATLAB: Error with Fuzzy Tree Model

There were a couple of issues that needed fixing.

Every FIS needs to have its Inputs, Outputs and MFs defined separately.
FIS3 would not have two MFs at its inputs. All it gets from FIS1 or FIS2 is a number. It decides what that number means based on the MF defined on its (FIS3's) inputs. So, yes, in your case you would have to redefine FIS1 and FIS2 output MFs as the MFs for the two inputs to FIS3.
The error that you see - “Error using FuzzyInferenceSystem/addRule (line 1148). Invalid input membership function name in rule description.” is because the inputs of FIS3 are instantiated, by default, to have 3 MFs called 'mf1', 'mf2', 'mf3' for both the inputs. You can see this by typing the following in the command window after executing the script until the line that errors out:

fis3.Inputs(1).MembershipFunctions
fis3.Inputs(2).MembershipFunctions

They have these default member functions because fis3 was created in your script with the following command:

fis3 = mamfis('Name','fis3','NumInputs',2,'NumOutputs',1);

Here's the code with a couple of changes. You should be able to adapt this for your use:

% Fuzzy model 1
fis1 = mamfis('Name',"tipper1");
fis1 = addInput(fis1,[0 10],'Name',"service");
fis1 = addMF(fis1,"service","gaussmf",[1.5 0],'Name',"poor");
fis1 = addMF(fis1,"service","gaussmf",[1.5 5],'Name',"good");
fis1 = addMF(fis1,"service","gaussmf",[1.5 10],'Name',"excellent");
fis1 = addInput(fis1,[0 10],'Name',"food");
fis1 = addMF(fis1,"food","trapmf",[-2 0 1 3],'Name',"rancid");
fis1 = addMF(fis1,"food","trapmf",[7 9 10 12],'Name',"delicious");
fis1 = addOutput(fis1,[0 30],'Name',"tip1");
fis1 = addMF(fis1,"tip1","trimf",[0 5 10],'Name',"cheap");
fis1 = addMF(fis1,"tip1","trimf",[10 15 20],'Name',"average");
fis1 = addMF(fis1,"tip1","trimf",[20 25 30],'Name',"generous");
rulefis1a = "service==poor | food==rancid => tip1=cheap";
rulefis1b = "service==good => tip1=average";
rulefis1c = "service==excellent | food==delicious => tip1=generous";
rulefis1d = [rulefis1a rulefis1b rulefis1c];
fis1b = addRule(fis1,rulefis1d);
 
%Fuzzy model 2
fis2 = mamfis('Name',"tipper2");
fis2 = addInput(fis2,[0 10],'Name',"service");
fis2 = addMF(fis2,"service","gaussmf",[1.5 0],'Name',"poor");
fis2 = addMF(fis2,"service","gaussmf",[1.5 5],'Name',"good");
fis2 = addMF(fis2,"service","gaussmf",[1.5 10],'Name',"excellent");
fis2 = addInput(fis2,[0 10],'Name',"food");
fis2 = addMF(fis2,"food","trapmf",[-2 0 1 3],'Name',"rancid");
fis2 = addMF(fis2,"food","trapmf",[7 9 10 12],'Name',"delicious");
fis2 = addOutput(fis2,[0 30],'Name',"tip2");
fis2 = addMF(fis2,"tip2","trimf",[0 5 10],'Name',"cheap");
fis2 = addMF(fis2,"tip2","trimf",[10 15 20],'Name',"average");
fis2 = addMF(fis2,"tip2","trimf",[20 25 30],'Name',"generous");
rulefis2a = "service==poor | food==rancid => tip2=cheap";
rulefis2b = "service==good => tip2=average";
rulefis2c = "service==excellent | food==delicious => tip2=generous";
rulefis2d = [rulefis2a rulefis2b rulefis2c];
fis2b = addRule(fis2,rulefis2d);
%Fuzzy model 3
fis3 = mamfis('Name','fis3');
fis3 = addInput(fis3,[0 30],'Name',"tip1");
fis3 = addMF(fis3,"tip1","trimf",[0 5 10],'Name',"cheap");
fis3 = addMF(fis3,"tip1","trimf",[10 15 20],'Name',"average");
fis3 = addMF(fis3,"tip1","trimf",[20 25 30],'Name',"generous");
fis3 = addInput(fis3,[0 30],'Name',"tip2");
fis3 = addMF(fis3,"tip2","trimf",[0 5 10],'Name',"cheap");
fis3 = addMF(fis3,"tip2","trimf",[10 15 20],'Name',"average");
fis3 = addMF(fis3,"tip2","trimf",[20 25 30],'Name',"generous");
fis3.Outputs(1).Name = "final_tip";
fis3 = addMF(fis3,"final_tip","trapmf",[-2 0 1 3],'Name',"low");
fis3 = addMF(fis3,"final_tip","trapmf",[7 9 10 12],'Name',"high");
rulefis3a = "tip1==cheap | tip2==average => final_tip=low";
rulefis3b = "tip1==average => final_tip=low";
rulefis3c = "tip1==generous | tip2==generous => final_tip=high";
rulefis3d = [rulefis3a rulefis3b rulefis3c];
fis3b = addRule(fis3,rulefis3d);
con1 = ["tipper1/tip1" "fis3/tip1"];
con2 = ["tipper2/tip2" "fis3/tip2"];
aggTree = fistree([fis1b fis2b fis3],[con1;con2]);
aggTree.FIS(3) = addRule(aggTree.FIS(3), rulefis3d)
output = evalfis(aggTree,[0.2 0.25 0.3 0.2])

MATLAB: Need help with genfis2 warning: Rank deficient

You have duplicates (or near duplicates) in your rows.

Best Answer

Related Solutions

MATLAB: Error with Fuzzy Tree Model

MATLAB: Need help with genfis2 warning: Rank deficient

Related Question