University:
Whitman College
Course:
MATH-350 | Foundations of Machine Learning
Academic year:

2022
Views:

90

Pages:

2
Author:

foltmedrcwq

Contents Sample RBF script Compute the RBF "Training" by using the Pseudoinverse Test the model Sample RBF script Example of a bad choice of parameters. Here, we'll use a Gaussian with very small width (1/10) tfunc=1; % Data is one dimensional in this case x=linspace(-3,3,100); y=sin(3*x); % Split into training set and testing set: temp=randperm(100); Xtrain=x(temp(1:10)); Ytrain=y(temp(1:10)); Xtest=x(temp(11:end)); Ytest=y(temp(11:end)); Compute the RBF %Choose the centers at random from the data numcenters=13; temp=randperm(100); Centers=x(temp(1:numcenters)); "Training" by using the Pseudoinverse A=edm(Xtrain',Centers'); Phi=rbf1(A,tfunc,1/10); alpha=pinv(Phi)*Ytrain'; Test the model For plotting purposes, we'll test the model on all the data A=edm(x',Centers'); Phi=rbf1(A,tfunc,1/10); Yout=Phi*alpha; plot(Xtrain,Ytrain,'k*',x,y,'g-',x,Yout,'b-'); axis([-3 3 -5 5]) title('Bad Parameters for the Gaussian');

Description

Related Documents

Free up your schedule!

Take 5 seconds to unlock