My partner and I have been trying to built a Landslide Susceptibility Map, using a Fuzzy Rule Based approach, with a software called FuzzyCell (Yanar and Akyürek, 2006). We have nine causative factors.
At first, we tried to develop our fuzzy rule base, using all of the linguistic variables: It didn't work at all, because the result didn't have any resemblance to a susceptibility map. On the contrary, when we used rules with four and five variables, trying to make all the variables to interact with each other, the result was very good.
I know the result would be very different by using one of the two approaches: Is it valid if we used the second one? if we worked with rules made of four or five variables, would it be valid?
If the fuzzy rules with fewer variables are those obtained considering the most significant variables for modeling, then, the information content can be considered equal, if is not carried out an analysis for the definition of the most significant variables, then the level of information with more or less variables may be different
- Badges
- Science topic
- Similar topics
- Linguistics