logo

AutoTS

wangzf / 2022-04-26

目录

AutoTS 简介

AutoTS is a time series package for Python designed for rapidly deploying high-accuracy forecasts at scale.

所有模型都支持预测多变量（多个时间序列）输出，还支持概率（上限/下限）预测。 大多数模型可以轻松扩展到数万甚至数十万个输入序列。许多模型还支持传入用户定义的外生回归变量

AutoTS 使用

AutoTS 安装

$ pip install autots

简单示例

from autots.datasets import load_monthly
from autots import AutoTS

df_long = load_monthly(long = True)

model = AutoTS(
    forecast_length = 3,
    frequency = "inter",
    ensemble = "simple",
    max_generations = 5,
    num_validations = 2,
)

model = model.fit(
    df_long, 
    date_col = "datetime",
    value_col = "value",
    id_col = "series_id",
)

# best model info
print(model)

数据格式

仅支持宽数据的低阶 API

from autots import AutoTS
from autots import load_hourly, load_daily, load_weekly, load_yearly, load_live_daily

# data
long = False
df = load_daily(long = long)


# model
model = AutoTS(
    forecast_length = 21,
    frequency = "infer",
    prediction_interval = 0.9,
    ensemble = None,
    model_list = "fast",  # "superfast", "default", "fast_parallel"
    transformer_list = "fast",  # "superfast"
    drop_most_recent = 1,
    max_generations = 4,
    num_validations = 2,
    validation_method = "backwards",
)
model = model.fit(
    df,
    date_col = "datetime" if long else None,
    value_col = "value" if long else None,
    id_col = "series_id" if long else None,
)

# prediction
prediction = model.predict()

# plot a sample
prediction.plot(
    model.df_wide_metric,
    series = model.df_wide_numeric.columns[0],
    start_date = "2019-01-01",
)

# best model
print(model)

# point forecast datetime
forecast_df = prediction.forecast
# upper and lower forecast
forecasts_up, forecasts_low = prediction.upper_forecast, prediction.lower_forecast

# accuracy of all tried model results
model_results = model.results()

# aggregated from cross validation
validation_results = model.results("validation")

快速和大数据

使用适当的模型列表，尤其是预定义的列表

查看预定义列表:

from autots.models.model_list import model_lists

参考