線性回歸 Linear Regression
阿新 • • 發佈:2017-08-16
gen for 技術 sed 2.x nbsp log ear tro
.
什麽是線性回歸?
y = f(x), a function that representing the relationship between input x and output y. The Function we used here is the general Linear Function.
`
常用的線性回歸函數
1.多項式函數(Polynominal Function)
(graph from wikipedia)
Comments:
1. most common function, fitting to most cases in life.
2. y is ranged from
3. x is ranged from
4. one x only has one y. (say, otherwise, for , one x has 2 different y)
2.sigmoid 函數(Sigmoid Function)
(graph from wikipedia)
Comments:
1. y is ranged from 0 to 1 (so it can represent the pobability, that‘s why the logistic regression uses this function to fitting the relationship between the prob and input)
2.x is ranged from
2.RBF 函數(Radial Basis Function)
(graph from wikipedia)
Comments:
1. radially symmetric
2. wide used (e.g. Gaussian Prob Density)
2017-08-15
線性模型的效果受線性公式影響, 應當選擇和當前應用項目匹配的函數。
線性回歸 Linear Regression