According to what I understood, linear regression predicts a result that can have continuous values, while logistic regression predicts a result that is discrete. It seems to me that logistic regression is similar to a classification problem. So why is this called regression?
There is also a related question: what is the difference between linear regression and logistic regression?
There is a strict relationship between linear regression and logistic regression.
With linear regression, you are looking for parameters k i:
h = k 0+ & Sigma; k i& dot; X i= K t & dot; X
, :
h = g (K t & dot; X)
g :
g
g (w) = 1/(1 + e -w)
:
h = 1/(1 + e -K t & dot; X)
K .
, h , x :
h
x
P (Y = 1) = 1/(1 + e -K t & dot; X)
0,5, "".
0,5, :
g (w) > 0,5
w = K t & dot; X & ge; 0
K t & dot; X = 0
- .
. ( Andrew Ng).
http://www.holehouse.org/mlclass/06_Logistic_Regression.html .
. / . . y .
y = c+ x1 * w1 + x2 * w2 + x3 * w3 +..... + xn * wn
- c, w1, w2,..., wn, y.
, . , logistic/sigmoid, y.
y = (c + x1 * w1 + x2 * w2 + x3 * w3 +.... + xn * wn)
y = 1/1 + e [- (c + x1 * w1 + x2 * w2 + x3 * w3 +.... + xn * wn)]
, , , .
, , . , y x, p (y = 1) x, y = 0 y = 1 (0,5).