Back to Search
Start Over
Enhancing Fingrams to deal with precise fuzzy systems
Enhancing Fingrams to deal with precise fuzzy systems
- Source :
- Fuzzy Sets and Systems. 297:1-25
- Publication Year :
- 2016
- Publisher :
- Elsevier BV, 2016.
-
Abstract
- Interpretability is a highly valued capability of fuzzy systems that turns essential when dealing with human interaction. Precise fuzzy modeling prioritizes performance at the cost of harming interpretability. Fuzzy Inference-grams (Fingrams) permit the graphical representation of fuzzy systems facilitating their comprehension, analysis and interpretation at inference level. We enhance Fingrams to better represent and analyze precise fuzzy systems. A specific metric and new representations handle the particularities of such systems. A new visual artifact allows to discover the set of data instances not covered by a given fuzzy system. A novel visual representation allows to study in detail the elements that are involved in the inference of a single data instance. The potentials of the enhanced methodology are sketched by taking the Fuzzy Unordered Rule Induction Algorithm (FURIA) as an illustrative example of precise fuzzy system. For instance, a highly valuable representation is obtained for the stretching mechanism of FURIA, thus facilitating its comprehensibility.
- Subjects :
- Adaptive neuro fuzzy inference system
Fuzzy classification
Neuro-fuzzy
Logic
Computer science
business.industry
05 social sciences
050301 education
02 engineering and technology
Fuzzy control system
Machine learning
computer.software_genre
Defuzzification
Fuzzy logic
Artificial Intelligence
0202 electrical engineering, electronic engineering, information engineering
Fuzzy set operations
020201 artificial intelligence & image processing
Fuzzy associative matrix
Artificial intelligence
business
0503 education
computer
Subjects
Details
- ISSN :
- 01650114
- Volume :
- 297
- Database :
- OpenAIRE
- Journal :
- Fuzzy Sets and Systems
- Accession number :
- edsair.doi...........48a13280432d11754c719d993694e938
- Full Text :
- https://doi.org/10.1016/j.fss.2015.05.019