[coursera] linear regression that is 1st learning algorithm

1st learning algorithm -> linear regression

[review]

supervised learning : given a "right answer" 

 - classification problem : 0 or 1 discrete value

 - regression problem : "predict" real-valued output

1.  linear regression with one variable

"exponentiation" ? 거듭제곱, 누승법 ... 

Training set -> learning algorithm -> h (hypothesis)

x -> h -> estimated value (y)   :   h maps from x's to y's

2.  mathematical definition of the linear regression with one variable

how to choose the theta 0 and theta 1?

y = ax + b

theta 0 = b value

theta 1 = a value

"abbreviation" ? 약어, 축약어 ...

J(theta 0, theta 1) = cost function = squared error function

==============================================================

ⓒ 2016 Coursera Inc. All rights reserved.

Created by: Stanford University.

Taught by: Andrew Ng, Associate Professor, Stanford University; Chief Scientist, 
Baidu; Chairman and Co-founder, Coursera

Website: www.coursera.org/learn/machine-learning

==============================================================