Octave Notes from Coursera online Machine Learning Course

Basic Operation

1. ~= : not equal
2. 2^6
3. && and || or
4. % comment
5. PS1('>> ') change prompt
6. ; not show results
7. a=3
8. a='hi'
9. pi
10. disp(a)
11. sprintf('fsdf %0.2f', a) , string print
12. format long/short

Math

std
sqrt
log
exp
sin, cos

Matrix/Vector

1. A=[1 2;3 4;5 6]
2. v=1:0.1:2 or v=1:6
3. ones, zeros, rand, randn for random normal, eye
4. sqrt
5. hist(w, bins=50)
6. help
7. A(2,:)
8. A([1 3],:) : 1st, 3rd row
9. A=[A, [1;2;3]] : append column
10. size(A) and size(A,1)
11. length(v)
12. C=[A B]
13. C=[A;B]

File/Imput/Save

1. pwd
2. cd
3. ls
4. load ***.dat , load('***.dat')
5. who , whos, =ls
6. save hello.txt v

Data/Matrix

• A*C
• A .* C and A .^ 2 ; elementwise
• abs, log, exp
• transpose A'
• max(a), max(A,[],1), column max, max(A,[],2), row max
• [val, ind] = max(a)
• find(a < 3)
• magic(3)
• [r,c]=find(A>7)
• sum, prod, floor, ceil, sum(A,1), pinv, inv

Plot

• plot(x,y,'rx') : r is color, x marker, ‘-‘ for lines, ‘+’, ‘o’
• hold on
• xlabel('time')
• legend('sin', 'cos')
• title('title')
• print -dpng 'xxx.png' : save to png
• figure(1)
• subplot(1,2,1)
• axis([0.5 1 -1 1])
• imagesc(A), colorbar, colarmap gray
• surf : 3-d
• contour : contour

Clear

clear
clc
close all

Loop

for i = 1:10,
	statement;
end;

indices=1:10;
for i = indices,
...

while i<=5,
...
	break;
end;

if i<=5,
	...
elseif ...,
else
	...
end;

2:end

Function

need to name myfunction.m, and put in the pwd

function [y1, y2]=myfcn(x)
y1=...

addpath('C:fff')

fmincg

submit homework

change to homework directory, after you finish, just type

submit()

Emacs

(autoload 'octave-mode "octave-mod" nil t)
(setq auto-mode-alist
	(cons '("\\.m$" . octave-mode) auto-mode-alist))
(add-hook 'octave-mode-hook
(lambda ()
        (abbrev-mode 1)
        (auto-fill-mode 1)
        (if (eq window-system 'x)
            (font-lock-mode 1))))

C-c Tab l : exec line