It appears you are concerned that ANFIS may not be suitable for your needs, particularly if you require exact binary outputs (0 or 1). Given that ANFIS Sugeno systems typically produce continuous outputs, it may not be the best choice for strict binary classification tasks, unless you are willing to perform post-processing of the fuzzy output data using the round() function.
x = linspace(1, 100);
fuzzy_out = rand(100, 1);
y = round(fuzzy_out);
figure
stem(x, fuzzy_out); ylim([-0.25, 1.25]), yline(0.5, 'r--')
title('Non-binary fuzzy outputs')
figure
stem(x, y); ylim([-0.25, 1.25]), yline(0.5, 'r--')
title('Exact binary outputs (0 or 1)')
