{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

FBFN_OLS

# FBFN_OLS - KartavyaNeema NeuroFuzzySystem (FBFN FBFN(contd...

This preview shows pages 1–9. Sign up to view the full content.

Kartavya Neema

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Neuro Fuzzy System
Fuzzy Basis Function Network (FBFN)

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
FBFN (contd.) y Fuzzy basis function is defined by: y In FBFN, a fuzzy system is represented as a series expansion of fuzzy basis functions (FBFs). Each FBF corresponds to one fuzzy logic rule.
Initial FBF Determination y Fuzzy based function: y Where p is: y Membership function is: y I chose 4 fuzzy rules (M). Thus, 4 membership for each input y We choose a = 1. σ = range of input/ no. of fuzzy rules. X (center) to be the input points in the given input output pair.

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
FBF as Linear regression y In order to train the network, FBF is treated as a special case of linear regression y Where d(t) is the system output, e(t) is the error
FBF as Linear regression y In order to use OLS, Matrix form is given: y Now, can we take the data and start the training to determine weights ( θ )

This preview has intentionally blurred sections. Sign up to view the full version.

View Full Document
Orthogonal Least Square (OLS) y An orthogonal least squares (OLS) learning algorithm is utilized to determine the significant FBFs [i.e. Fuzzy
This is the end of the preview. Sign up to access the rest of the document.

{[ snackBarMessage ]}