Input weights in jfuzzylite

[Unknown]

Hi Juan,

Would you mind explaining how weighting inputs may be applied to jfuzzylite ?

For example, having 3 inputs [input1, input2, input3], how would I set up jfuzzylite to weight “input3” more than the others?

example: input1 with weight 0.2, input2 with weight 0.2, input3 with weight 0.6.

The problem I am solving has multiple inputs and multiple outputs, in case this makes any difference in setting up the weights.

Thanks in advance.

PS. I am using TakagiSugeno fuzzy method.

[Juan Rada-Vilela (admin)]

Hi,

thanks for your post.

Assuming that a weight is be a multiplier in [0.0,1.0], and the input values are only positive (including zero), you would just need to multiply the weight by the input value at the moment of setting the input value for the variable. For example, variable.setValue(inputValue * weight);. If you want something a bit fancier, you can normalise the set of weights as it is done in in Proportional.java, thus requiring weights to be any positive value not necessarily in [0.0,1.0].

Cheers.

[Unknown]

Hi Juan,

Thank you for your prompt answer.

Kind regards

[Author]

Posts