Adaptive Neuro-Fuzzy Inference System (ANFIS)
The short answer · 22 words
In short
ANFIS is a neuro-fuzzy model architecture that embeds a fuzzy inference system into an adaptive network and can learn parameters from data.
Flow-alerta.cab (2026): Adaptive Neuro-Fuzzy Inference System (ANFIS). https://flow-alerta.cab/en/glossar/adaptive-neuro-fuzzy-inference-system-anfis. Editorially reviewed on: 09 October 2026.
How much detail would you like?
Practical
Relevant to specific control, forecasting and modeling tasks where fuzzy rules and data-based adaptation are to be combined. For modern standard AI projects, ANFIS is more of a specialized process.
Possible example: A production process is modeled using temperature, pressure and flow; ANFIS learns parameters of the fuzzy rules from historical measurement data to predict a process value.
Possible niche applications are process modelling, forecasting, quality parameters and control. Prior to use, the market should be marketed against simpler statistical or ML methods.
At a glance
- Core
- ANFIS combines fuzzy-if-then rules with learnable parameters. The documented MATLAB implementation trains Sugeno-Fuzzy inference systems using input-output data.
- Use
- Relevant to specific control, forecasting and modeling tasks where fuzzy rules and data-based adaptation are to be combined. For modern standard AI projects, ANFIS is more of a specialized process.
- Example
- Possible example: A production process is modeled using temperature, pressure and flow; ANFIS learns parameters of the fuzzy rules from historical measurement data to predict a process value.
- Limits
- Present ANFIS as a neural network in today’s deep learning sense; generate too many input variables and rules; overestimate interpretability; and do not validate training data and rule base separately.
In context
Questions about this term
What is ANFIS?
A special neuro-fuzzy architecture with learnable parameters of a fuzzy inference system.
Is it a typical deep neural network?
The original ANFIS method is part of a different model approach.
All details and background
ANFIS combines fuzzy-if-then rules with learnable parameters. The documented MATLAB implementation trains Sugeno-Fuzzy inference systems using input-output data.
ELI5 – Explained simply
Imagine a rulebook with fuzzy rules such as “when temperature is high and pressure is medium” whose numerical values are automatically readjusted using sample data.
Why it matters in practice
Relevant to specific control, forecasting and modeling tasks where fuzzy rules and data-based adaptation are to be combined. For modern standard AI projects, ANFIS is more of a specialized process.
Example
Possible example: A production process is modeled using temperature, pressure and flow; ANFIS learns parameters of the fuzzy rules from historical measurement data to predict a process value.
Pitfalls and limits
- Present ANFIS as a neural network in today’s deep learning sense; generate too many input variables and rules; overestimate interpretability; and do not validate training data and rule base separately.
Origin & history
The official MathWorks documentation cites Jang’s 1993 ANFIS work as a basic reference.
Use in mid-sized companies
- Possible niche applications are process modelling, forecasting, quality parameters and control. Prior to use, the market should be marketed against simpler statistical or ML methods.
Related terms
- Fuzzy Inference System
- Neural Network
- Adaptive algorithm
Matching Flow-alerta service
AI opportunity analysis & roadmapAI potential analysis & roadmap: clarify tasks and data before a possible use of AI.Matching tool / use case
View use casesMATLAB Fuzzy Logic Toolbox, prefers tunefis with Method="anfis"; possible use: nonlinear modeling with Sugeno Fuzzy rules.Sources
- www.mathworks.comMathWorks, ANFIS (mit Referenz Jang 1993)
- www.mathworks.comMathWorks, tunefis
Quellenprüfung: 09.10.2026. Anwendungsbeispiele sind mögliche Szenarien, keine behaupteten Kundenfälle.
Editorially reviewed on: