glmnet包使用笔记

Logistic Regression

family = "binomial"

library(glmnet)
data(BinomialExample)
x <- BinomialExample$x
y <- BinomialExample$y
fit <- glmnet(x, y, family = "binomial")
predict(fit, newx = x[1:5,], type = "class", s = c(0.05, 0.01))

beta 为不同?下，自变量的系数

图片alt

> as.matrix(fit$beta)
    s0         s1         s2          s3         s4         s5
V1   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V2   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V3   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V4   0 -0.0795432 -0.1525939 -0.21325350 -0.2696943 -0.3229869
V5   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V6   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V7   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000
V8   0  0.0000000  0.0000000  0.00000000  0.0000000  0.0000000

s for specifiying the value(s) of ? at which to extract coefficients/predictions.

图片alt

不同的98个Lambda取值下，的df（自由度），%Dev （percent deviance explained）

> print(fit)
Call:  glmnet(x = x, y = y, family = "binomial") 
   Df  %Dev   Lambda
1   0  0.00 0.240500
2   1  2.90 0.219100
......
97 30 99.89 0.000032
98 30 99.90 0.000029