Nixtla - ML Forecast

Nixtla - ML Forecast

import pandas as pd
from statsforecast import StatsForecast

df = pd.read_csv('https://datasets-nixtla.s3.amazonaws.com/air-passengers.csv', parse_dates=['ds'])
df.head()

	unique_id	ds	y
0	AirPassengers	1949-01-01	112
1	AirPassengers	1949-02-01	118
2	AirPassengers	1949-03-01	132
3	AirPassengers	1949-04-01	129
4	AirPassengers	1949-05-01	121

df['unique_id'].value_counts()

unique_id
AirPassengers    144
Name: count, dtype: int64

StatsForecast.plot(df, engine='plotly')

from mlforecast import MLForecast
from mlforecast.target_transforms import Differences
from sklearn.linear_model import LinearRegression

fcst = MLForecast(
    models=LinearRegression(),
    freq='MS',  # our serie has a monthly frequency
    lags=[12],
    target_transforms=[Differences([1])],
)

fcst.fit(df)

MLForecast(models=[LinearRegression], freq=<MonthBegin>, lag_features=['lag12'], date_features=[], num_threads=1)

preds = fcst.predict(12)
preds

	unique_id	ds	LinearRegression
0	AirPassengers	1961-01-01	444.656555
1	AirPassengers	1961-02-01	417.470764
2	AirPassengers	1961-03-01	446.903076
3	AirPassengers	1961-04-01	491.014160
4	AirPassengers	1961-05-01	502.622253
5	AirPassengers	1961-06-01	568.751465
6	AirPassengers	1961-07-01	660.044312
7	AirPassengers	1961-08-01	643.343323
8	AirPassengers	1961-09-01	540.666748
9	AirPassengers	1961-10-01	491.462799
10	AirPassengers	1961-11-01	417.095245
11	AirPassengers	1961-12-01	461.206299

StatsForecast.plot(df, preds, engine='plotly')