我有一个df：
head(vardata_num,30)
wt ht dist
1 9.00 7.49 5.41
2 8.75 8.03 5.60
3 8.50 7.24 5.30
4 8.50 7.19 5.23
5 8.50 6.94 5.41
6 8.50 7.30 4.52
7 8.50 7.40 2.91
8 8.50 6.97 2.59
9 8.50 7.08 2.08
10 8.50 6.63 1.51
11 8.50 6.73 1.39
12 8.00 6.97 1.76
13 8.00 6.58 1.50
14 8.00 6.90 1.56
15 8.00 6.04 2.43
16 8.00 5.75 2.79
17 8.00 5.72 3.34
18 8.00 5.75 3.53
19 8.00 5.73 3.08
20 7.50 5.45 3.39
21 7.50 5.58 3.34
22 7.50 5.66 4.22
23 7.50 5.71 5.35
24 7.00 5.86 5.99
25 7.00 4.87 6.65
26 7.00 4.87 6.51
27 7.00 4.91 5.34
28 7.00 4.90 4.71
29 7.00 4.83 3.95
30 7.00 4.50 4.9
我正在尝试使用自定义函数使用循环预测这三个变量：
library(mafs)
ts_lapply <- lapply(vardata_num, function(x) select_forecast(ts(x),test_size = 10, horizon = 3, error = "MAPE")) # select_forecast() from `mafs` package
我得到了如下令人敬畏的输出：
ts_lapply [1]＃为一个变量
$wt
$wt$df_models
model ME RMSE MAE MPE MAPE MASE ACF1 best_model runtime_model
1 auto.arima -0.2826189 0.3439279 0.3102743 -4.236451 4.617906 3.149053 0.3624226 tbats 0.25
2 bats -0.1240606 0.1857631 0.1676487 -1.878305 2.479522 1.701510 0.1344866 tbats 1.60
3 croston -0.8903343 0.9183247 0.8903343 -13.164300 13.164300 9.036229 0.4086420 tbats 1.08
4 ets 0.4009141 0.5010641 0.4133541 5.941724 6.126020 4.195235 0.7171674 tbats 0.05
5 meanf -0.3138889 0.3862010 0.3583333 -4.709314 5.322341 3.636816 0.4086420 tbats 0.00
6 naive -0.4250000 0.4808846 0.4250000 -6.339031 6.339031 4.313433 0.4086420 tbats 0.02
7 nnetar -0.9093190 1.0220091 0.9093190 -13.543554 13.543554 9.228909 0.6192314 tbats 1.61
8 rwf -0.4250000 0.4808846 0.4250000 -6.339031 6.339031 4.313433 0.4086420 tbats 0.00
9 rwf_drift -0.3683824 0.4203015 0.3745588 -5.497541 5.582734 3.801493 0.3661817 tbats 0.00
10 snaive -0.4250000 0.4808846 0.4250000 -6.339031 6.339031 4.313433 0.4086420 tbats 0.00
11 splinef 0.3383595 0.4374294 0.3586967 5.014299 5.315592 3.640504 0.7075714 tbats 0.34
12 StructTS 0.2306096 0.3144022 0.2637272 3.405088 3.895718 2.676634 0.6510857 tbats 0.00
13 tbats -0.1240606 0.1857631 0.1676487 -1.878305 2.479522 1.701510 0.1344866 tbats 3.58
14 thetaf -0.4280973 0.4842184 0.4280973 -6.385061 6.385061 4.344868 0.4107693 tbats 0.02
$wt$best_forecast
Point Forecast Lo 80 Hi 80 Lo 95 Hi 95
182 6.435371 6.171671 6.699071 6.032077 6.838666
183 6.353675 5.929582 6.777769 5.705081 7.002270
184 6.288319 5.692546 6.884092 5.377163 7.199475
$wt$df_comparison
time forecasted observed
1 0171-12-18 7.183704 7.25
2 0172-12-18 7.098355 7.25
3 0173-12-18 7.030424 6.75
4 0174-12-18 6.976302 6.75
5 0175-12-19 6.933147 6.75
6 0176-12-18 6.898715 6.75
7 0177-12-18 6.871227 6.75
8 0178-12-18 6.849274 6.75
9 0179-12-18 6.831736 6.75
10 0180-12-18 6.817721 6.50
我想从这些输出列表中提取一些信息，以便可以提供简明的摘要。表格形式的输出格式如下所示：
var no_obs best_model pt_fcast1 pt_fcast2 pt_fcast3
wt 181 tbats 6.435371 6.353675 6.288319
ht 150 croston 6.924046 6.963400 6.660900
dist 240 nnetar 0.8252435 1.5224686 2.2196936
在这里需要一些帮助，以便如何制作这张桌子!!!!