12760811
Descripción
Mapa Mental por Rafael Salgado, actualizado hace más de 1 año
|
Creado por Rafael Salgado
hace alrededor de 6 años
|
|
ML intro
- By Data
- Supervised: given training set is labeled
- Unsupervised: no right or wrong answer provided in training set
- Supervised: given training set is labeled
- By Result
- Categorization
- Continuous
- Categorization
- Linear
regression
- Gradient descent: small step iterations until reaching a minima
Nota:
- Find minimum of cost function J by iterating mini steps to the minimal point
- By number of features "n"
- Single variable: find parameters theta0 and theta1
- Multivariate: find parameters theta0... thetan
- Single variable: find parameters theta0 and theta1
- By function: select best function to match training set
- Linear
- Polynomial: square root
- Linear
- J(ThetaMatrix) = sum ((h(x) - y)^2)/2m
- h(x): hypothesis function. h(x) = x0*theta0 + x1*theta1
- x0 is equal to 1 most of the time!
- x0 is equal to 1 most of the time!
- y: output or target variable
- m: number of training sets
- h(x): hypothesis function. h(x) = x0*theta0 + x1*theta1
- Normal
(analytical)
- Gradient descent: small step iterations until reaching a minima
- Logistic regression
- Gradient descent
- Advanced optimization
methods
- functions in
Matlab/Octave)
- fminunc
- functions in
Matlab/Octave)
- Details
- Used for classification models
- Used for classification models
- Gradient descent
- find parameters Theta to
minimize cost
function J
¿Quieres crear tus propios Mapas Mentales gratis con GoConqr? Más información.