AutoGluon Time Series - Forecasting Quick Start

Via a simple fit() call, AutoGluon can train and tune

• simple forecasting models (e.g., ARIMA, ETS, Theta),

• powerful deep learning models (e.g., DeepAR, Temporal Fusion Transformer),

• tree-based models (e.g., LightGBM),

• an ensemble that combines predictions of other models

to produce multi-step ahead probabilistic forecasts for univariate time series data.

This tutorial demonstrates how to quickly start using AutoGluon to generate hourly forecasts for the M4 forecasting competition dataset.

Loading time series data as a TimeSeriesDataFrame

First, we import some required modules

import pandas as pd
from autogluon.timeseries import TimeSeriesDataFrame, TimeSeriesPredictor

To use autogluon.timeseries, we will only need the following two classes:

• TimeSeriesDataFrame stores a dataset consisting of multiple time series.

• TimeSeriesPredictor takes care of fitting, tuning and selecting the best forecasting models, as well as generating new forecasts.

We load a subset of the M4 hourly dataset as a pandas.DataFrame

df = pd.read_csv("https://autogluon.s3.amazonaws.com/datasets/timeseries/m4_hourly_subset/train.csv")
df.head()

	item_id	timestamp	target
0	H1	1750-01-01 00:00:00	605.0
1	H1	1750-01-01 01:00:00	586.0
2	H1	1750-01-01 02:00:00	586.0
3	H1	1750-01-01 03:00:00	559.0
4	H1	1750-01-01 04:00:00	511.0

AutoGluon expects time series data in long format. Each row of the data frame contains a single observation (timestep) of a single time series represented by

• unique ID of the time series ("item_id") as int or str

• timestamp of the observation ("timestamp") as a pandas.Timestamp or compatible format

• numeric value of the time series ("target")

The raw dataset should always follow this format with at least three columns for unique ID, timestamp, and target value, but the names of these columns can be arbitrary. It is important, however, that we provide the names of the columns when constructing a TimeSeriesDataFrame that is used by AutoGluon. AutoGluon will raise an exception if the data doesn’t match the expected format.

train_data = TimeSeriesDataFrame.from_data_frame(
    df,
    id_column="item_id",
    timestamp_column="timestamp"
)
train_data.head()

		target
item_id	timestamp
H1	1750-01-01 00:00:00	605.0
	1750-01-01 01:00:00	586.0
	1750-01-01 02:00:00	586.0
	1750-01-01 03:00:00	559.0
	1750-01-01 04:00:00	511.0

We refer to each individual time series stored in a TimeSeriesDataFrame as an item. For example, items might correspond to different products in demand forecasting, or to different stocks in financial datasets. This setting is also referred to as a panel of time series. Note that this is not the same as multivariate forecasting — AutoGluon generates forecasts for each time series individually, without modeling interactions between different items (time series).

TimeSeriesDataFrame inherits from pandas.DataFrame, so all attributes and methods of pandas.DataFrame are available in a TimeSeriesDataFrame. It also provides other utility functions, such as loaders for different data formats (see TimeSeriesDataFrame for details).

Training time series models with TimeSeriesPredictor.fit

To forecast future values of the time series, we need to create a TimeSeriesPredictor object.

Models in autogluon.timeseries forecast time series multiple steps into the future. We choose the number of these steps — the prediction length (also known as the forecast horizon) — depending on our task. For example, our dataset contains time series measured at hourly frequency, so we set prediction_length = 48 to train models that forecast up to 48 hours into the future.

We instruct AutoGluon to save trained models in the folder ./autogluon-m4-hourly. We also specify that AutoGluon should rank models according to mean absolute scaled error (MASE), and that data that we want to forecast is stored in the column "target" of the TimeSeriesDataFrame.

predictor = TimeSeriesPredictor(
    prediction_length=48,
    path="autogluon-m4-hourly",
    target="target",
    eval_metric="MASE",
)

predictor.fit(
    train_data,
    presets="medium_quality",
    time_limit=600,
)

Beginning AutoGluon training... Time limit = 600s
AutoGluon will save models to '/home/ci/autogluon/docs/tutorials/timeseries/autogluon-m4-hourly'
=================== System Info ===================
AutoGluon Version:  1.2b20250107
Python Version:     3.11.9
Operating System:   Linux
Platform Machine:   x86_64
Platform Version:   #1 SMP Tue Sep 24 10:00:37 UTC 2024
CPU Count:          8
GPU Count:          1
Memory Avail:       28.71 GB / 30.95 GB (92.8%)
Disk Space Avail:   213.32 GB / 255.99 GB (83.3%)
===================================================
Setting presets to: medium_quality
Fitting with arguments:
{'enable_ensemble': True,
 'eval_metric': MASE,
 'hyperparameters': 'light',
 'known_covariates_names': [],
 'num_val_windows': 1,
 'prediction_length': 48,
 'quantile_levels': [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9],
 'random_seed': 123,
 'refit_every_n_windows': 1,
 'refit_full': False,
 'skip_model_selection': False,
 'target': 'target',
 'time_limit': 600,
 'verbosity': 2}
Inferred time series frequency: 'h'
Provided train_data has 148060 rows, 200 time series. Median time series length is 700 (min=700, max=960).
Provided data contains following columns:
target: 'target'
AutoGluon will gauge predictive performance using evaluation metric: 'MASE'
This metric's sign has been flipped to adhere to being higher_is_better. The metric score can be multiplied by -1 to get the metric value.
===================================================
Starting training. Start time is 2025-01-07 02:27:25
Models that will be trained: ['Naive', 'SeasonalNaive', 'RecursiveTabular', 'DirectTabular', 'ETS', 'Theta', 'Chronos[bolt_small]', 'TemporalFusionTransformer']
Training timeseries model Naive. Training for up to 66.4s of the 598.0s of remaining time.
-6.6629       = Validation score (-MASE)
0.15    s     = Training runtime
1.93    s     = Validation (prediction) runtime
Training timeseries model SeasonalNaive. Training for up to 74.5s of the 596.0s of remaining time.
-1.2169       = Validation score (-MASE)
0.14    s     = Training runtime
0.20    s     = Validation (prediction) runtime
Training timeseries model RecursiveTabular. Training for up to 85.1s of the 595.6s of remaining time.
-0.9339       = Validation score (-MASE)
11.22   s     = Training runtime
1.28    s     = Validation (prediction) runtime
Training timeseries model DirectTabular. Training for up to 97.2s of the 583.1s of remaining time.
-1.2921       = Validation score (-MASE)
8.83    s     = Training runtime
0.54    s     = Validation (prediction) runtime
Training timeseries model ETS. Training for up to 114.7s of the 573.7s of remaining time.
-1.9661       = Validation score (-MASE)
0.13    s     = Training runtime
25.57   s     = Validation (prediction) runtime
Training timeseries model Theta. Training for up to 137.0s of the 548.0s of remaining time.
-2.1425       = Validation score (-MASE)
0.13    s     = Training runtime
23.09   s     = Validation (prediction) runtime
Training timeseries model Chronos[bolt_small]. Training for up to 174.9s of the 524.8s of remaining time.
-0.8121       = Validation score (-MASE)
1.84    s     = Training runtime
1.19    s     = Validation (prediction) runtime
Training timeseries model TemporalFusionTransformer. Training for up to 260.8s of the 521.6s of remaining time.
-2.5410       = Validation score (-MASE)
69.33   s     = Training runtime
0.26    s     = Validation (prediction) runtime
Fitting simple weighted ensemble.
Ensemble weights: {'Chronos[bolt_small]': 0.74, 'DirectTabular': 0.04, 'ETS': 0.02, 'RecursiveTabular': 0.19}
-0.7915       = Validation score (-MASE)
1.04    s     = Training runtime
28.57   s     = Validation (prediction) runtime
Training complete. Models trained: ['Naive', 'SeasonalNaive', 'RecursiveTabular', 'DirectTabular', 'ETS', 'Theta', 'Chronos[bolt_small]', 'TemporalFusionTransformer', 'WeightedEnsemble']
Total runtime: 147.59 s
Best model: WeightedEnsemble
Best model score: -0.7915

<autogluon.timeseries.predictor.TimeSeriesPredictor at 0x7f3191ef74d0>

Here we used the "medium_quality" presets and limited the training time to 10 minutes (600 seconds). The presets define which models AutoGluon will try to fit. For medium_quality presets, these are simple baselines (Naive, SeasonalNaive), statistical models (ETS, Theta), tree-based models based on LightGBM (RecursiveTabular, DirectTabular), a deep learning model TemporalFusionTransformer, and a weighted ensemble combining these. Other available presets for TimeSeriesPredictor are "fast_training", "high_quality" and "best_quality". Higher quality presets will usually produce more accurate forecasts but take longer to train.

Inside fit(), AutoGluon will train as many models as possible within the given time limit. Trained models are then ranked based on their performance on an internal validation set. By default, this validation set is constructed by holding out the last prediction_length timesteps of each time series in train_data.

Generating forecasts with TimeSeriesPredictor.predict

We can now use the fitted TimeSeriesPredictor to forecast the future time series values. By default, AutoGluon will make forecasts using the model that had the best score on the internal validation set. The forecast always includes predictions for the next prediction_length timesteps, starting from the end of each time series in train_data.

predictions = predictor.predict(train_data)
predictions.head()

Model not specified in predict, will default to the model with the best validation score: WeightedEnsemble

		mean	0.1	0.2	0.3	0.4	0.5	0.6	0.7	0.8	0.9
item_id	timestamp
H1	1750-01-30 04:00:00	622.191633	598.352106	606.819926	612.956287	617.685092	622.191633	626.836042	631.409826	636.792800	644.542990
	1750-01-30 05:00:00	563.195006	537.393103	546.954576	553.249304	558.428323	563.195006	567.606607	572.418778	578.155968	586.702938
	1750-01-30 06:00:00	521.683860	494.428072	504.041315	511.148911	516.678754	521.683860	526.831450	532.170287	538.339738	547.689556
	1750-01-30 07:00:00	490.505614	461.721751	471.825645	479.072630	485.098704	490.505614	496.272718	502.215465	508.974322	519.732682
	1750-01-30 08:00:00	465.746852	435.351515	445.591065	453.185296	459.950196	465.746852	471.472651	477.611101	485.099102	495.941271

AutoGluon produces a probabilistic forecast: in addition to predicting the mean (expected value) of the time series in the future, models also provide the quantiles of the forecast distribution. The quantile forecasts give us an idea about the range of possible outcomes. For example, if the "0.1" quantile is equal to 500.0, it means that the model predicts a 10% chance that the target value will be below 500.0.

We will now visualize the forecast and the actually observed values for one of the time series in the dataset. We plot the mean forecast, as well as the 10% and 90% quantiles to show the range of potential outcomes.

import matplotlib.pyplot as plt

# TimeSeriesDataFrame can also be loaded directly from a file
test_data = TimeSeriesDataFrame.from_path("https://autogluon.s3.amazonaws.com/datasets/timeseries/m4_hourly_subset/test.csv")

# Plot 4 randomly chosen time series and the respective forecasts
predictor.plot(test_data, predictions, quantile_levels=[0.1, 0.9], max_history_length=200, max_num_item_ids=4);

Evaluating the performance of different models

We can view the performance of each model AutoGluon has trained via the leaderboard() method. We provide the test data set to the leaderboard function to see how well our fitted models are doing on the unseen test data. The leaderboard also includes the validation scores computed on the internal validation dataset.

Note the test data includes both the forecast horizon (last prediction_length values of each time series) as well as the historic data (all except the last prediction_last values).

In AutoGluon leaderboards, higher scores always correspond to better predictive performance. Therefore our MASE scores are multiplied by -1, such that higher “negative MASE”s correspond to more accurate forecasts.

# The test score is computed using the last
# prediction_length=48 timesteps of each time series in test_data
predictor.leaderboard(test_data)

Additional data provided, testing on additional data. Resulting leaderboard will be sorted according to test score (`score_test`).

	model	score_test	score_val	pred_time_test	pred_time_val	fit_time_marginal	fit_order
0	WeightedEnsemble	-0.697903	-0.791454	34.101403	28.572223	1.040225	9
1	Chronos[bolt_small]	-0.725739	-0.812070	0.900767	1.189491	1.842149	7
2	RecursiveTabular	-0.862797	-0.933874	1.346205	1.280111	11.215119	3
3	SeasonalNaive	-1.022854	-1.216909	0.192244	0.198851	0.141757	2
4	DirectTabular	-1.605700	-1.292127	0.622435	0.537270	8.826765	4
5	ETS	-1.806131	-1.966061	31.222695	25.565351	0.133521	5
6	Theta	-1.905365	-2.142531	23.010331	23.087366	0.131717	6
7	TemporalFusionTransformer	-2.272206	-2.540983	0.335873	0.263074	69.328701	8
8	Naive	-6.696079	-6.662942	0.180932	1.929274	0.147607	1

Summary

We used autogluon.timeseries to make probabilistic multi-step forecasts on the M4 Hourly dataset. Check out Forecasting Time Series - In Depth to learn about the advanced capabilities of AutoGluon for time series forecasting.