Content deleted Content added
m Fixing broken anchor: 2022-10-21 #Takagi-Sugeno-Kang (TSK)→Fuzzy logic#Takagi–Sugeno–Kang (TSK) |
Split the paragraph into a list for better readability |
||
| (2 intermediate revisions by 2 users not shown) | |||
Line 1:
{{Short description|Type of artificial neural network}}
An '''adaptive neuro-fuzzy inference system''' or '''adaptive network-based fuzzy inference system''' ('''ANFIS''') is a kind of [[artificial neural network]] that is based on [[Fuzzy logic#Takagi–Sugeno–Kang (TSK)|Takagi–Sugeno fuzzy inference system]]. The technique was developed in the early 1990s.<ref>{{cite conference |last=Jang |first=Jyh-Shing R |title=Fuzzy Modeling Using Generalized Neural Networks and Kalman Filter Algorithm |year=1991|conference=Proceedings of the 9th National Conference on Artificial Intelligence, Anaheim, CA, USA, July 14–19 |volume=2 |pages=762–767 |url=http://www.aaai.org/Papers/AAAI/1991/AAAI91-119.pdf }}</ref><ref>{{cite journal |last=Jang |first=J.-S.R. |s2cid=14345934 |year=1993 |title=ANFIS: adaptive-network-based fuzzy inference system |journal=IEEE Transactions on Systems, Man, and Cybernetics |volume=23 |issue=3 |pages=665–685 |doi=10.1109/21.256541 }}</ref> Since it integrates both neural networks and [[fuzzy logic]] principles, it has potential to capture the benefits of both in a single [[:wikt:framework|framework]].
Line 4 ⟶ 5:
==ANFIS architecture==
It is possible to identify two parts in the network structure, namely premise and consequence parts. In more details, the architecture is composed by five layers.
# The first layer takes the input values and determines the [[Membership function (mathematics)|membership functions]] belonging to them. It is commonly called fuzzification layer. The membership degrees of each function are computed by using the premise parameter set, namely {a,b,c}. # The second layer is responsible of generating the firing strengths for the rules. Due to its task, the second layer is denoted as "rule layer". # The role of the third layer is to normalize the computed firing strengths, by dividing each value for the total firing strength. # The fourth layer takes as input the normalized values and the consequence parameter set {p,q,r}. # The values returned by this layer are the defuzzificated ones and those values are passed to the last layer to return the final output.<ref name="KarabogaKaya2018">{{cite journal|last1=Karaboga|first1=Dervis|last2=Kaya|first2=Ebubekir|title=Adaptive network based fuzzy inference system (ANFIS) training approaches: a comprehensive survey|journal=Artificial Intelligence Review|volume=52|issue=4|pages=2263–2293|year=2018|issn=0269-2821|doi=10.1007/s10462-017-9610-2|s2cid=40548050}}</ref> ===Fuzzification layer===
The first layer of an ANFIS network describes the difference to a vanilla neural network. Neural networks in general are operating with a [[data pre-processing]] step, in which the [[Feature (machine learning)|features]] are converted into normalized values between 0 and 1. An ANFIS neural network doesn't need a [[sigmoid function]], but it's doing the preprocessing step by converting numeric values into fuzzy values.<ref>{{cite journal |doi=10.1109/72.159060 |pmid=18276470 |year=1992 |publisher=Institute of Electrical and Electronics Engineers (IEEE) |volume=3 |number=5 |pages=714–723 |author=J.-S.R. Jang |title=Self-learning fuzzy controllers based on temporal backpropagation |journal=IEEE Transactions on Neural Networks }}</ref>
Here is an example: Suppose, the network gets as input the distance between two points in the 2d space. The distance is measured in pixels and it can have values from 0 up to 500 pixels. Converting the numerical values into [[
==References==
| |||