i have the following question about the fuzzy decision trees: I have a training data(10 columns, 1,000,000 rows), how can a decision tree(e.g ID3) "work" with fuzzy sets? This training set has to be "fuzzified" before the FuzzyID3 algorithm uses it?. I have worked with decision trees algorithms but never with fuzzy logic, i'm studying fuzzy logic to understand how can i "wrap" fuzzy logic with ID3 algorithm.
These questions arise from: I need to give a prediction, where some variables have influence in others(some are independent too), that's the reason why i think that fuzzy-ID3 can "solve" that problem. Another important question, some variables that are not in the training data(no historic register) have a clearly influence in variables in the training data. Is there any way to include this variables to the final prediction?. Let me give a situation: This prediction is based in the historical data, but clearly don't include the actual real-world situation, for example, a prediction of a "forest fire(a dumb example)", this consider only the historical data, and indeed the prediction could be good, but did not consider that it's been raining for 1 week...
Sorry for my bad english and my poor knowledges about fuzzy logic. Thanks for reading my questions!
