## V6 0.364828074 -0.035326316 -0.329501758

## V7 -0.058746523 0.080343071 -0.021596548

## V8 0.483592031 0.111422669 -0.595014699

## V9 -0.155745580 -1.016502806 1.172248386

## V10 -0.060698812 0.625050169 -0.564351357

## V11 . . .

## V12 . . .

## V13 -0.175719655 0.283930678 -0.108211023

## V14 0.196421536 -0.139576235 -0.056845300

## V15 . . .

## V16 -0.037414770 0.040300172 -0.002885402

## V17 0.149438019 -0.129742710 -0.019695308

## V18 . . .

## V19 0.088822086 -0.130605368 0.041783282

## V20 . . .

## V21 0.007141749 -0.002869644 -0.004272105

## V22 0.125997952 -0.016924514 -0.109073438

## V23 0.043024971 -0.026879150 -0.016145821

## V24 0.016862193 0.083686360 -0.100548554

## V25 . . .

## V26 . . .

## V27 . . .

## V28 -0.111429811 -0.069842376 0.181272187

## V29 . . .

## V30 -0.032032333 -0.006590025 0.038622358

最终模型正确分类率
xtabs(~ y_multi_pred_class + y_multi)

1. ## y_multi
2. ## y_multi_pred_class 1 2 3
3. ## 1 84 20 16
4. ## 2 30 136 19
5. ## 3 28 18 149

mean(y_multi == y_multi_pred_class)
## [1] 0.738
交叉验证测试集正确分类率

1. lapply(unique(foldid), function(id) {
2. ## 拟合排除测试集（foldid==id）
3. ## 使用模型拟合最佳lambda测试集Y?hat
4. y_pred <- (predict(fit, newx = x_multi[foldid == id,], type = "class",
5. s = lambda.min))
6. ## 测试集Y
7. y <- y_multi[foldid == id]
8. ## 测试集CCR
9. mean(y == y_pred)
10. }) %>%

## [1] 0.68 0.64 0.76 0.72 0.70 0.66 0.60 0.72 0.62 0.76
## [1] 0.686
二元逻辑回归示例

1. ## # A tibble: 100 x 30
2. ## V1 V2 V3 V4 V5 V6 V7 V8
3. ## <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
4. ## 1 -0.61926135 0.01624409 -0.62606831 0.41268461 0.4944374 -0.4493269 0.6760053 -0.06771419
5. ## 2 1.09427278 0.47257285 -1.33714704 -0.64058126 0.2823199 -0.6093321 0.3547232 -0.62686515
6. ## 3 -0.35670402 0.30121334 0.19056192 0.23402677 0.1698086 1.2291427 1.1628095 0.88024242
7. ## 4 -2.46907012 2.84771447 1.66024352 1.56881297 -0.8330570 -0.5620088 -0.6142455 -1.76529838
8. ## 5 0.56728852 0.88888747 -0.01158671 0.57627526 -0.8689453 -0.3132571 0.6902907 -1.29961200
9. ## 6 0.91292543 0.77350086 0.55836355 -0.53509922 0.3507093 -0.5763021 -0.3882672 0.55518663
10. ## 7 0.09567305 0.14027229 -0.76043921 -0.04935541 1.5740992 -0.1240903 -1.1106276 1.72895452
11. ## 8 1.93420667 -0.71114983 -0.27387147 1.00113828 1.0439012 0.8028893 -0.6035769 -0.51136380
12. ## 9 0.28275701 1.05940570 -0.03944966 0.30277367 -0.9161762 0.6914934 0.6087553 0.30921594
13. ## 10 0.80143492 1.53674274 -1.01230763 -0.38480878 -2.0319100 0.2236314 -1.1628847 -0.52739792
14. ## # ... with 90 more rows, and 22 more variables: V9 <dbl>, V10 <dbl>, V11 <dbl>, V12 <dbl>,
15. ## # V13 <dbl>, V14 <dbl>, V15 <dbl>, V16 <dbl>, V17 <dbl>, V18 <dbl>, V19 <dbl>, V20 <dbl>,
16. ## # V21 <dbl>, V22 <dbl>, V23 <dbl>, V24 <dbl>, V25 <dbl>, V26 <dbl>, V27 <dbl>, V28 <dbl>,
17. ## # V29 <dbl>, V30 <dbl>

as_data_frame(y)

1. ## # A tibble: 100 x 1
2. ## value
3. ## <int>
4. ## 1 0
5. ## 2 1
6. ## 3 1
7. ## 4 0
8. ## 5 1
9. ## 6 0
10. ## 7 0
11. ## 8 0
12. ## 9 1
13. ## 10 1
14. ## # ... with 90 more rows
15. ## 执行岭回归
16. ## 二元逻辑回归
17. family = "binomial",
18. ## “alpha=1”是套索惩罚 ， “alpha=0”是岭惩罚 。