项目描述
utilsforecast
安装
PyPI
pip install utilsforecast
Conda
conda install -c conda-forge utilsforecast
如何使用
生成合成数据
from utilsforecast.data import generate_series
series = generate_series(3, with_trend=True, static_as_categorical=False)
series
|
unique_id |
ds |
y |
| 0 |
0 |
2000-01-01 |
0.422133 |
| 1 |
0 |
2000-01-02 |
1.501407 |
| 2 |
0 |
2000-01-03 |
2.568495 |
| 3 |
0 |
2000-01-04 |
3.529085 |
| 4 |
0 |
2000-01-05 |
4.481929 |
| ... |
... |
... |
... |
| 481 |
2 |
2000-06-11 |
163.914625 |
| 482 |
2 |
2000-06-12 |
166.018479 |
| 483 |
2 |
2000-06-13 |
160.839176 |
| 484 |
2 |
2000-06-14 |
162.679603 |
| 485 |
2 |
2000-06-15 |
165.089288 |
486行 × 3列
绘图
from utilsforecast.plotting import plot_series
fig = plot_series(series, plot_random=False, max_insample_length=50, engine='matplotlib')
fig.savefig('imgs/index.png', bbox_inches='tight')
预处理
from utilsforecast.preprocessing import fill_gaps
serie = series[series['unique_id'].eq(0)].tail(10)
# drop some points
with_gaps = serie.sample(frac=0.5, random_state=0).sort_values('ds')
with_gaps
|
unique_id |
ds |
y |
| 213 |
0 |
2000-08-01 |
18.543147 |
| 214 |
0 |
2000-08-02 |
19.941764 |
| 216 |
0 |
2000-08-04 |
21.968733 |
| 220 |
0 |
2000-08-08 |
19.091509 |
| 221 |
0 |
2000-08-09 |
20.220739 |
fill_gaps(with_gaps, freq='D')
|
unique_id |
ds |
y |
| 0 |
0 |
2000-08-01 |
18.543147 |
| 1 |
0 |
2000-08-02 |
19.941764 |
| 2 |
0 |
2000-08-03 |
NaN |
| 3 |
0 |
2000-08-04 |
21.968733 |
| 4 |
0 |
2000-08-05 |
NaN |
| 5 |
0 |
2000-08-06 |
NaN |
| 6 |
0 |
2000-08-07 |
NaN |
| 7 |
0 |
2000-08-08 |
19.091509 |
| 8 |
0 |
2000-08-09 |
20.220739 |
评估
from functools import partial
import numpy as np
from utilsforecast.evaluation import evaluate
from utilsforecast.losses import mape, mase
valid = series.groupby('unique_id').tail(7).copy()
train = series.drop(valid.index)
rng = np.random.RandomState(0)
valid['seas_naive'] = train.groupby('unique_id')['y'].tail(7).values
valid['rand_model'] = valid['y'] * rng.rand(valid['y'].shape[0])
daily_mase = partial(mase, seasonality=7)
evaluate(valid, metrics=[mape, daily_mase], train_df=train)
|
unique_id |
指标 |
seas_naive |
rand_model |
| 0 |
0 |
mape |
0.024139 |
0.440173 |
| 1 |
1 |
mape |
0.054259 |
0.278123 |
| 2 |
2 |
mape |
0.042642 |
0.480316 |
| 3 |
0 |
mase |
0.907149 |
16.418014 |
| 4 |
1 |
mase |
0.991635 |
6.404254 |
| 5 |
2 |
mase |
1.013596 |
11.365040 |
下载文件
下载适用于您平台的文件。如果您不确定选择哪个，请了解更多关于安装包的信息。
源代码发行版
编译后的发行版
