Time series forecasting suite using statistical models
Project description
statsforecast
Forecasting using statistical models
Install
pip install statsforecast
How to use
import numpy as np
import pandas as pd
from IPython.display import display, Markdown
from statsforecast import StatsForecast
from statsforecast.models import random_walk_with_drift, seasonal_naive, ses
def display_df(df):
display(Markdown(df.to_markdown()))
rng = np.random.RandomState(0)
serie1 = np.arange(1, 8)[np.arange(100) % 7] + rng.randint(-1, 2, size=100)
serie2 = np.arange(100) + rng.rand(100)
series = pd.DataFrame(
{
'ds': pd.date_range('2000-01-01', periods=serie1.size + serie2.size, freq='D'),
'y': np.hstack([serie1, serie2]),
},
index=pd.Index([0] * serie1.size + [1] * serie2.size, name='unique_id')
)
display_df(pd.concat([series.head(), series.tail()]))
| unique_id | ds | y |
|---|---|---|
| 0 | 2000-01-01 00:00:00 | 0 |
| 0 | 2000-01-02 00:00:00 | 2 |
| 0 | 2000-01-03 00:00:00 | 2 |
| 0 | 2000-01-04 00:00:00 | 4 |
| 0 | 2000-01-05 00:00:00 | 5 |
| 1 | 2000-07-14 00:00:00 | 95.7649 |
| 1 | 2000-07-15 00:00:00 | 96.9441 |
| 1 | 2000-07-16 00:00:00 | 97.75 |
| 1 | 2000-07-17 00:00:00 | 98.3394 |
| 1 | 2000-07-18 00:00:00 | 99.4895 |
fcst = StatsForecast(series, models=[random_walk_with_drift, (seasonal_naive, 7), (ses, 0.1)], freq='D', n_jobs=2)
forecasts = fcst.forecast(5)
display_df(forecasts)
2021-12-08 20:33:06 statsforecast.core INFO: Computing forecasts
2021-12-08 20:33:07 statsforecast.core INFO: Computed forecasts for random_walk_with_drift.
2021-12-08 20:33:07 statsforecast.core INFO: Computed forecasts for seasonal_naive_season_length-7.
2021-12-08 20:33:07 statsforecast.core INFO: Computed forecasts for ses_alpha-0.1.
| unique_id | ds | random_walk_with_drift | seasonal_naive_season_length-7 | ses_alpha-0.1 |
|---|---|---|---|---|
| 0 | 2000-04-10 00:00:00 | 3.0303 | 3 | 3.85506 |
| 0 | 2000-04-11 00:00:00 | 3.06061 | 5 | 3.85506 |
| 0 | 2000-04-12 00:00:00 | 3.09091 | 4 | 3.85506 |
| 0 | 2000-04-13 00:00:00 | 3.12121 | 7 | 3.85506 |
| 0 | 2000-04-14 00:00:00 | 3.15152 | 6 | 3.85506 |
| 1 | 2000-07-19 00:00:00 | 100.489 | 93.0166 | 90.4709 |
| 1 | 2000-07-20 00:00:00 | 101.489 | 94.2307 | 90.4709 |
| 1 | 2000-07-21 00:00:00 | 102.489 | 95.7649 | 90.4709 |
| 1 | 2000-07-22 00:00:00 | 103.489 | 96.9441 | 90.4709 |
| 1 | 2000-07-23 00:00:00 | 104.489 | 97.75 | 90.4709 |
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
statsforecast-0.2.0.tar.gz
(13.4 kB
view hashes)
Built Distribution
Close
Hashes for statsforecast-0.2.0-py3-none-any.whl
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 | c9e4150083795a7d5c962162527bc9a140141cd3957b72c5fb21f27dbc065755 |
|
| MD5 | 266f419fa9dd8a92b98aad6751167774 |
|
| BLAKE2b-256 | bc688f939502e6c22c2d302c8067cb6824c54c974ca484044aa2a892e2c4c6c2 |
