问题描述
我的数据集 (be) 只是一个通用系列，我使用该系列的 2 个滞后值作为预测变量。当我通过 mlr3 运行 ranger 时，我收到一条错误消息，说 pdata$se 中存在缺失值，不知道出了什么问题，数据集中没有缺失值。任何帮助表示赞赏。
library(mlr3verse)
library(ranger)
library(tidyverse)
be <- read_csv("index,series,lag1,lag2
1,7577039.736,3501685.67,3266643.3920000005
2,12670615.333999999,3501685.67
3,15547362.117999999,7577039.736
4,17862429.700000003,12670615.333999999
5,12456382.122,15547362.117999999
6,11776258.268,17862429.700000003
7,7415982.802,12456382.122
8,4909208.992,11776258.268
9,3153256.446,7415982.802
10,3627321.234,4909208.992
11,3598643.3920000005,3153256.446
12,3601685.67,3627321.234
13,6851039.736,3598643.3920000005
14,12010615.333999999,3601685.67
15,14907362.117999999,6851039.736
16,17062069.700000003,12010615.333999999
17,13776351.902,14907362.117999999
18,11200258.268,17062069.700000003
19,7014952.802,13776351.902
20,5149328.592,11200258.268
21,3555456.446,7014952.802
22,3401281.694,5149328.592
23,3600643.512,3555456.446
24,3799885.67,3401281.694
25,6640039.915999999,3600643.512
26,13330615.414,3799885.67
27,15333362.797999999,6640039.915999999
28,17463529.756,13330615.414
29,13974352.982,15333362.797999999
30,11601538.787999999,17463529.756
31,7417952.998,13974352.982
32,5379327.002,11601538.787999999
33,3357456.446,7417952.998
34,3519881.694,5379327.002
35,3976643.3920000005,3357456.446
36,3827685.67,3519881.694
37,7448839.736,3976643.3920000005
38,12072815.333999999,3827685.67
39,14607362.498,7448839.736
40,17278069.700000003,12072815.333999999
41,13984351.762,14607362.498
42,11640258.068,17278069.700000003
43,7374951.602,13984351.762
44,11640258.068
45,3757456.446,7374951.602
46,3773281.694,5149328.592
47,4057799.7260000003,3757456.446
48,3905802.05,3773281.694
49,7600857.22,4057799.7260000003
50,12319199.668,3905802.05
51,14905472.284,7600857.22
52,17630683.714,12319199.668
53,14269747.042,14905472.284
54,11877814.702,17630683.714
55,7525461.166,14269747.042
56,5254417.278,11877814.702
57,3834139.578,7525461.166
58,3850287.79,5254417.278
59,4776868.952,3834139.578
60,6650859.738,3850287.79
61,8932155.812,4776868.952
62,13286680.239999998,6650859.738
63,15266373.356,8932155.812
64,14753582.824000001,13286680.239999998
65,11423115.616,15266373.356
66,9694168.222,14753582.824000001
67,12653617.627999999,11423115.616
68,3537417.068,9694168.222
69,2314654.84,12653617.627999999
70,1903689.762,3537417.068")
n_train <- 48
n_test <- 22
task <- TaskRegr$new(id = "regr_ranger",backend = be,target = "series")
learner <- lrn("regr.ranger",predict_type = "se")
set.seed(1502)
train <- learner$train(task,row_ids = 1:n_train)
test_preds <- learner$predict(task,row_ids = (n_train + 1):(n_train + n_test))
#> Warning in value[[3L]](cond): Calibration Failed with error:
#> Error in approx(x = calib.x,y = calib.y,xout = vars): need at least two non-NA values to interpolate
#> Falling back to non-calibrated variance estimates.
#> Warning in sqrt(infjack$var.hat): NaNs produced
#> Error in check_prediction_data.PredictionDataRegr(pdata): Assertion on 'pdata$se' Failed: Contains missing values (element 21).
由 reprex package (v2.0.0) 于 2021 年 5 月 14 日创建
解决方法
暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！
如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。
小编邮箱:dio#foxmail.com (将#修改为@）