The c-machine-learning from parthopdas

c-machine-learning's Introduction

Coursera: Machine Learning

Regression vs Classification:

	Regression	Classification
Predicted values	Continuous	Discrete
Intuition	Predictions are points on a hyper-plane.	Predictions are points in hyper-volumes divided by a hyper-plane
Linear vs Polynomial	Hyper-plane vs Hyper-surface	Hyper-volumes on 'either side' of Hyper-plane vs Hyper-surface
E.g. for $n = 1$, output is	Straight line in $\mathbb{R}^2$	Areas on either side of a straight line in $\mathbb{R}^2$
Linear vs Polynomial	Hyper-plane vs Curved Hyper-surface in $\mathbb{R}^{n+1}$	Hyper-volumes on 'either side' of Hyper-plane vs Curved Hyper-surface in in $\mathbb {R} ^{n+1}$

Octave cheatsheet:

Command	Action
pinv/inv	matrix (psuedo)inverse
format long/short	Precision of double display
sprintf/printf/disp	Format strings/Display value
1.0:0.1:2.9	Range
zeros/ones/rand/eye(r, c)	Commonly used matrices
size, length	Dimensions of matrix
who/whos	List objects
save/load	serialization of data
A(r, c)/A(2, :)/A([1 4], :)	read/write cells/rows/columns
[A, [...]]	append row/columns
A(:)	unroll
[A B]/[A; B]	Concat column/row wise
/.	matrix multiply/matrix element-wise multiply
[val ind] = max/max	Value/index of max
elementwise compare op
find(compare op)
sum/prod/floor/ceil
flip/fliplr/flipup	Flip stuff
plot
hold on/off	Reuse/Create new figure
clear
close
print -dpng	print the plot in a bmp
figure	select figure
subplot	picture - in - picture
axis	change axis
imagesc	visualize matrix as grid of colors
colorbar, colormap
,	chain commands
;	dont print command output
timeit/tick-tock	Time steps

Recommend Projects