Implementing perceptron using while and for
1 Ansicht (letzte 30 Tage)
Ältere Kommentare anzeigen
Hello, so i'm trying to implement a perceptron in Matlab in order to seperate the letters I and O by using a while and a for loop. I run the code but it wont come to an end. Any ideas?
I've attached the Excel file from where I import my data. In the first column is the number of data, in the second the number of pixels, the third is the variance on the horizontal axis and the final column is the classes, where with 1 is the letter I and with -1 the letter O.
Here's the script i use to call the function and initialize my data:
clear
clc
data=xlsread("O_I_classification data.xlsx");
x1=data(:,2); %Ορισμός της μεταβλητής X1
x2=data(:,3); %Ορισμός της μεταβλητής Χ2
arithmos_dedomenwn=data(:,1); %Ως arithmos_dedomenwn ορίζεται η πρώτη στήλη των δεδομένων που περιέχει τον αριθμό των δεδομένων
y=data(:,4); %Ως y ορίζεται η τελευταία στήλη που περιλαμβάνει τις κλάσεις (δεδομένα εξόδου).
[w1,w2,b,counter]= Ask31(x1,x2,y);
And here's the function I've managed to write:
function [w1,w2,b,counter]= Ask31(x1,x2,y)
x_k=[x1,x2];
x_k_ones=[x_k ones(59,1)];
row=size(x_k_ones);
W=rand(3,1)';
counter=0;
for i=1:row
counter=counter+1;
z=sum(x_k_ones(i,:).*W);
y_problepsi=sign(z);
W(i,1)=W(i,1)+0.7*(y(i,1)-y_problepsi(i,1)).*x_k_ones(i,1);
while sum(abs(y-y_problepsi))~=0
delta_w=0.7*(y-y_problepsi).*x_k_ones;
W=W+delta_w;
if isequal(y,y_problepsi)
break
end
end
end
w1=W(1,1);
w2=W(1,2);
b=W(1,3);
end
0 Kommentare
Antworten (1)
Geoff Hayes
am 14 Apr. 2022
@Nick Vasilakis - your while loop looks like
while sum(abs(y-y_problepsi))~=0
delta_w=0.7*(y-y_problepsi).*x_k_ones;
W=W+delta_w;
if isequal(y,y_problepsi)
break
end
end
yet no where in the code do the values for the condition, y and y_problepsi change. This could be the reason why you are getting stuck in this loop. I recommend reviewing how the W is defined and whether it should be used to update y or y_problepsi.
1 Kommentar
Siehe auch
Kategorien
Mehr zu Sequence and Numeric Feature Data Workflows finden Sie in Help Center und File Exchange
Community Treasure Hunt
Find the treasures in MATLAB Central and discover how the community can help you!
Start Hunting!