> For the complete documentation index, see [llms.txt](https://skywateryang.gitbook.io/timeseriesanalysis101/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://skywateryang.gitbook.io/timeseriesanalysis101/7.-ji-yu-shen-du-xue-xi-de-shi-jian-xu-lie-fen-xi-fang-fa/7.1-lstm-chang-duan-qi-ji-yi-wang-luo/7.1.2-shi-yong-darts-tiao-yong.md).

# 7.1.2 使用Darts调用

```python
from darts import TimeSeries
from darts.dataprocessing.transformers import Scaler
from darts.models import RNNModel
from darts.metrics import mape
from darts.utils.timeseries_generation import datetime_attribute_timeseries
import warnings
warnings.filterwarnings("ignore")
import logging
logging.disable(logging.CRITICAL)
```

```python
df = pd.read_csv('data\Alcohol_Sales.csv')
```

```python
series = TimeSeries.from_dataframe(df, 'DATE', 'S4248SM144NCEN')
```

```python
# 划分训练集和测试集
train, val = series.split_after(pd.Timestamp('20170101'))

transformer = Scaler()
train_transformed = transformer.fit_transform(train)
val_transformed = transformer.transform(val)
series_transformed = transformer.transform(series)

# 创建年和月的协变量序列
year_series = datetime_attribute_timeseries(pd.date_range(start=series.start_time(), freq=series.freq_str, periods=400),
                                             attribute='year', one_hot=False)
year_series = Scaler().fit_transform(year_series)
month_series = datetime_attribute_timeseries(year_series, attribute='month', one_hot=True)
covariates = year_series.stack(month_series)
cov_train, cov_val = covariates.split_after(pd.Timestamp('20170101'))
```

```python
my_model = RNNModel(
    model='LSTM',
    hidden_dim=100,
    dropout=0,
    batch_size=1,
    n_epochs=100,
    optimizer_kwargs={'lr': 1e-3},
    model_name='Alcohol_RNN',
    log_tensorboard=True,
    random_state=42,
    training_length=12,
    input_chunk_length=12,
    force_reset=True
)
```

```python
my_model.fit(train_transformed, covariates=cov_train, val_series=val_transformed, val_covariates=cov_val, verbose=True)
```

```python
def eval_model(model, lag):
    pred_series = model.predict(n=24,covariates=covariates)
    plt.figure(figsize=(8,5))
    series_transformed[-lag:].plot(label='actual')
    pred_series.plot(label='forecast')
    plt.title('MAPE: {:.2f}%'.format(mape(pred_series, val_transformed)))
    plt.legend();

eval_model(my_model,0)
eval_model(my_model,30)
```

![](https://3993849477-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-Mhv1ams1lvf_fMUn_is%2F-MiFh_8zxcPT2WmQwTZR%2F-MiFhiATYMr3QqoGp8Az%2F8_7.png?alt=media\&token=af40e7ff-4710-4300-83cf-5e2100b6aa09)

![](https://3993849477-files.gitbook.io/~/files/v0/b/gitbook-legacy-files/o/assets%2F-Mhv1ams1lvf_fMUn_is%2F-MiFh_8zxcPT2WmQwTZR%2F-MiFhjDInIN3G6llIDEX%2F8_8.png?alt=media\&token=80c08820-0d2d-4c38-ad9f-71e8d718e0da)