Time Series Forecasting Module Development Tutorial¶

I. Overview¶

Time series forecasting aims to predict the possible values or states at a future point in time or within a future time period by analyzing patterns, trends, periodicity, and other characteristics in historical data. This helps enterprises and organizations make more accurate decisions, optimize resource allocation, reduce risks, and seize potential market opportunities. These time series data typically originate from various sensors, economic activities, social behaviors, and other real-world application scenarios. For example, stock prices, temperature changes, website traffic, sales data, and the like are all typical examples of time series data.

II. Supported Model List¶

Model Name	Model Download Link	mse	mae	Model Size (M)	Introduce
DLinear	Inference Model/Trained Model	0.382	0.394	76k	Simple structure, high efficiency and easy-to-use time series prediction model
Nonstationary	Inference Model/Trained Model	0.600	0.515	60.3M	Based on the transformer structure, targeted optimization of long-term time series prediction models for non-stationary time series
PatchTST	Inference Model/Trained Model	0.385	0.397	2.2M	High-precision long-term time series prediction model that takes into account both local patterns and global dependencies
TiDE	Inference Model/Trained Model	0.405	0.412	34.9M	High-precision model suitable for handling multivariate, long-term time series prediction problems
TimesNet	Inference Model/Trained Model	0.417	0.431	5.2M	Through multi-period analysis, TimesNet is a highly adaptable high-precision time series analysis model

Test Environment Description:

Performance Test Environment
Test Dataset: ETTH1 test dataset.
Hardware Configuration:
- GPU: NVIDIA Tesla T4
- CPU: Intel Xeon Gold 6271C @ 2.60GHz
- Other Environments: Ubuntu 20.04 / cuDNN 8.6 / TensorRT 8.5.2.2
Inference Mode Description

Mode	GPU Configuration	CPU Configuration	Acceleration Technology Combination
Normal Mode	FP32 Precision / No TRT Acceleration	FP32 Precision / 8 Threads	PaddleInference
High-Performance Mode	Optimal combination of pre-selected precision types and acceleration strategies	FP32 Precision / 8 Threads	Pre-selected optimal backend (Paddle/OpenVINO/TRT, etc.)

III. Quick Integration¶

❗ Before quick integration, please install the PaddleX wheel package. For detailed instructions, refer to the PaddleX Local Installation Guide

Just a few lines of code can complete the inference of the Time Series Forecasting module, allowing you to easily switch between models under this module. You can also integrate the model inference of the the Time Series Forecasting module into your project. Before running the following code, please download the demo csv to your local machine.

from paddlex import create_model
model = create_model(model_name="DLinear")
output = model.predict("ts_fc.csv", batch_size=1)
for res in output:
    res.print(json_format=False)
    res.save_to_csv(save_path="./output/")
    res.save_to_json(save_path="./output/res.json")

After running, the result obtained is:

{'res': {'input_path': 'ts_fc.csv', 'forecast':                            OT
date
2018-06-26 20:00:00  9.586131
2018-06-26 21:00:00  9.379762
2018-06-26 22:00:00  9.252275
2018-06-26 23:00:00  9.249993
2018-06-27 00:00:00  9.164998
...                       ...
2018-06-30 15:00:00  8.830340
2018-06-30 16:00:00  9.291553
2018-06-30 17:00:00  9.097666
2018-06-30 18:00:00  8.905430
2018-06-30 19:00:00  8.993793

[96 rows x 1 columns]}}

The meanings of the parameters in the running results are as follows: - input_path: Indicates the path to the input time-series file for prediction. - forecast: Indicates the time-series forecast result. You can save the forecast results to a CSV file using res.save_to_csv() or to a JSON file using res.save_to_json().

Descriptions of related methods, parameters, etc., are as follows:

The create_model method instantiates a time-series forecasting model (using DLinear as an example). Specific descriptions are as follows:

Parameter	Description	Type	Options	Default Value
`model_name`	The name of the model	`str`	All model names supported by PaddleX	None
`model_dir`	The storage path of the model	`str`	None	None

The model_name must be specified. After specifying model_name, PaddleX's built-in model parameters are used by default. If model_dir is specified, the user-defined model is used.
The predict() method of the time-series forecasting model is called for inference prediction. The parameters of the predict() method are input and batch_size, with specific descriptions as follows:

Parameter	Description	Type	Options	Default Value
`input`	Data to be predicted, supporting multiple input types	`Python Var`/`str`/`list`	Python variable, such as time-series data represented by `pandas.DataFrame` File path, such as the local path of a time-series file: `/root/data/ts.csv` URL link, such as the network URL of a time-series file: Example Local directory, which must contain data files for prediction, such as the local path: `/root/data/` List, elements of the list must be of the above types, such as `[pandas.DataFrame, pandas.DataFrame]`, `[\"/root/data/ts1.csv\", \"/root/data/ts2.csv\"]`, `[\"/root/data1\", \"/root/data2\"]`, `[{\"ts\": \"/root/data1\"}, {\"ts\": \"/root/data2/ts.csv\"}]`	None
`batch_size`	Batch size	`int`	Any integer greater than 0	1

The prediction results are processed for each sample, with the prediction result being of type dict, and support operations such as printing, saving as a csv file, and saving as a json file:

Method	Description	Parameter	Type	Explanation	Default Value
`print()`	Print the result to the terminal	`format_json`	`bool`	Whether to format the output content with `JSON` indentation	`True`
		`indent`	`int`	Specify the indentation level to beautify the output `JSON` data, making it more readable. Only effective when `format_json` is `True`	4
		`ensure_ascii`	`bool`	Control whether non-`ASCII` characters are escaped to `Unicode`. When set to `True`, all non-`ASCII` characters will be escaped; `False` retains the original characters. Only effective when `format_json` is `True`	`False`
`save_to_json()`	Save the result as a `json` file	`save_path`	`str`	The file path for saving. When a directory is provided, the saved file name matches the input file name	None
		`indent`	`int`	Specify the indentation level to beautify the output `JSON` data, making it more readable. Only effective when `format_json` is `True`	4
		`ensure_ascii`	`bool`	Control whether non-`ASCII` characters are escaped to `Unicode`. When set to `True`, all non-`ASCII` characters will be escaped; `False` retains the original characters. Only effective when `format_json` is `True`	`False`
`save_to_csv()`	Save the result as a time-series `csv` file	`save_path`	`str`	The file path for saving. When a directory is provided, the saved file name matches the input file name	None

Additionally, it also supports obtaining the visualized time-series with results and the prediction results via attributes, as follows:

Attribute	Description
`json`	Get the prediction result in `json` format
`csv`	Get the time-series prediction result in `csv` format

For more information on using PaddleX's single-model inference API, refer to the PaddleX Single Model Python Script Usage Instructions.

IV. Custom Development¶

If you seek higher accuracy, you can leverage PaddleX's custom development capabilities to develop better Time Series Forecasting models. Before developing a Time Series Forecasting model with PaddleX, ensure you have installed PaddleClas plugin for PaddleX. The installation process can be found in the custom development section of the PaddleX Local Installation Tutorial.

4.1 Dataset Preparation¶

Before model training, you need to prepare a dataset for the task. PaddleX provides data validation functionality for each module. Only data that passes validation can be used for model training. Additionally, PaddleX provides demo datasets for each module, which you can use to complete subsequent development. If you wish to use private datasets for model training, refer to PaddleX Time Series Forecasting Task Module Data Preparation Tutorial.

4.1.1 Demo Data Download¶

You can download the demo dataset to a specified folder using the following commands:

wget https://paddle-model-ecology.bj.bcebos.com/paddlex/data/ts_dataset_examples.tar -P ./dataset
tar -xf ./dataset/ts_dataset_examples.tar -C ./dataset/

4.1.2 Data Validation¶

Data validation can be completed with a single command:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=check_dataset \
    -o Global.dataset_dir=./dataset/ts_dataset_examples

After executing the above command, PaddleX will verify the dataset and collect basic information about it. Once the command runs successfully, a message saying Check dataset passed ! will be printed in the log. The verification results will be saved in ./output/check_dataset_result.json, and related outputs will be stored in the ./output/check_dataset directory, including visual examples of sample images and a histogram of sample distribution.

👉 Verification Result Details (click to expand)

The specific content of the verification result file is:

{
  "done_flag": true,
  "check_pass": true,
  "attributes": {
    "train_samples": 12194,
    "train_table": [
      [
        "date",
        "HUFL",
        "HULL",
        "MUFL",
        "MULL",
        "LUFL",
        "LULL",
        "OT"
      ],
      [
        "2016-07-01 00:00:00",
        5.827000141143799,
        2.009000062942505,
        1.5989999771118164,
        0.4620000123977661,
        4.203000068664552,
        1.3400000333786009,
        30.5310001373291
      ],
      [
        "2016-07-01 01:00:00",
        5.692999839782715,
        2.075999975204468,
        1.4919999837875366,
        0.4259999990463257,
        4.142000198364259,
        1.371000051498413,
        27.78700065612793
      ]
    ],
    "val_samples": 3484,
    "val_table": [
      [
        "date",
        "HUFL",
        "HULL",
        "MUFL",
        "MULL",
        "LUFL",
        "LULL",
        "OT"
      ],
      [
        "2017-11-21 02:00:00",
        12.994000434875488,
        4.889999866485597,
        10.055999755859377,
        2.878000020980835,
        2.559000015258789,
        1.2489999532699585,
        4.7129998207092285
      ],
      [
        "2017-11-21 03:00:00",
        11.92199993133545,
        4.554999828338623,
        9.097000122070312,
        3.0920000076293945,
        2.559000015258789,
        1.2790000438690186,
        4.8540000915527335
      ]
    ]
  },
  "analysis": {
    "histogram": ""
  },
  "dataset_path": "./dataset/ts_dataset_examples",
  "show_type": "csv",
  "dataset_type": "TSDataset"
}

The verification results above indicate that check_pass being True means the dataset format meets the requirements. Explanations for other indicators are as follows:

attributes.train_samples: The number of training samples in this dataset is 12194;
attributes.val_samples: The number of validation samples in this dataset is 3484;
attributes.train_sample_paths: A list of relative paths to the top 10 rows of training samples in this dataset;
attributes.val_sample_paths: A list of relative paths to the top 10 rows of validation samples in this dataset;

4.1.3 Dataset Format Conversion/Dataset Splitting (Optional) (Click to Expand)¶

👉 Details on Format Conversion/Dataset Splitting (Click to Expand)

After completing dataset verification, you can convert the dataset format or re-split the training/validation ratio by modifying the configuration file or appending hyperparameters.

(1) Dataset Format Conversion

Time Series Forecasting supports converting xlsx and xls format datasets to the required format.

Parameters related to dataset verification can be set by modifying the CheckDataset fields in the configuration file. Example explanations for some parameters in the configuration file are as follows:

CheckDataset:
convert:
enable: Whether to enable dataset format conversion, supporting xlsx and xls format conversion, default is False;
src_dataset_type: If dataset format conversion is enabled, the source dataset format needs to be set, default is null.

Modify the paddlex/configs/modules/ts_forecast/DLinear.yaml configuration as follows:

......
CheckDataset:
  ......
  convert:
    enable: True
    src_dataset_type: null
  ......

Then execute the command:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=check_dataset \
    -o Global.dataset_dir=./dataset/ts_forecast_to_convert

Of course, the above parameters also support being set by appending command-line arguments. For a LabelMe format dataset, the command is:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=check_dataset \
    -o Global.dataset_dir=./dataset/ts_forecast_to_convert \
    -o CheckDataset.convert.enable=True \

(2) Dataset Splitting

Parameters for dataset splitting can be set by modifying the CheckDataset fields in the configuration file. Example explanations for some parameters in the configuration file are as follows:

CheckDataset:
split:
enable: Whether to enable re-splitting the dataset, set to True to perform dataset splitting, default is False;
train_percent: If re-splitting the dataset, set the percentage of the training set, which should be an integer between 0 and 100, ensuring the sum with val_percent is 100;

For example, if you want to re-split the dataset with a 90% training set and a 10% validation set, modify the configuration file as follows:

......
CheckDataset:
  ......
  split:
    enable: True
    train_percent: 90
    val_percent: 10
  ......

Then execute the command:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=check_dataset \
    -o Global.dataset_dir=./dataset/ts_dataset_examples

After dataset splitting, the original annotation files will be renamed to xxx.bak in the original path.

The above parameters also support setting through appending command line arguments:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml  \
    -o Global.mode=check_dataset \
    -o Global.dataset_dir=./dataset/ts_dataset_examples \
    -o CheckDataset.split.enable=True \
    -o CheckDataset.split.train_percent=90 \
    -o CheckDataset.split.val_percent=10

4.2 Model Training¶

Model training can be completed with just one command. Here, we use the Time Series Forecasting model (DLinear) as an example:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=train \
    -o Global.dataset_dir=./dataset/ts_dataset_examples

You need to follow these steps:

Specify the .yaml configuration file path for the model (here it's DLinear.yaml,When training other models, you need to specify the corresponding configuration files. The relationship between the model and configuration files can be found in the PaddleX Model List (CPU/GPU)).
Set the mode to model training: -o Global.mode=train
Specify the training dataset path: -o Global.dataset_dir

Other related parameters can be set by modifying the Global and Train fields in the .yaml configuration file, or adjusted by appending parameters in the command line. For example, to train using the first two GPUs: -o Global.device=gpu:0,1; to set the number of training epochs to 10: -o Train.epochs_iters=10. For more modifiable parameters and their detailed explanations, refer to the PaddleX TS Configuration Parameters Documentation.

👉 More Details (Click to Expand)

During model training, PaddleX automatically saves model weight files, with the default path being output. To specify a different save path, use the -o Global.output field in the configuration file.
PaddleX abstracts the concepts of dynamic graph weights and static graph weights from you. During model training, both dynamic and static graph weights are produced, and static graph weights are used by default for model inference.
After model training, all outputs are saved in the specified output directory (default is ./output/), typically including:
train_result.json: Training result record file, including whether the training task completed successfully, produced weight metrics, and related file paths.
train.log: Training log file, recording model metric changes, loss changes, etc.
config.yaml: Training configuration file, recording the hyperparameters used for this training session.
best_accuracy.pdparams.tar, scaler.pkl, .checkpoints, .inference: Model weight-related files, including Model weight-related files, including network parameters, optimizers, and network architecture.

4.3 Model Evaluation¶

After model training, you can evaluate the specified model weights on the validation set to verify model accuracy. Using PaddleX for model evaluation requires just one command:

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=evaluate \
    -o Global.dataset_dir=./dataset/ts_dataset_examples

Similar to model training, follow these steps:

Specify the .yaml configuration file path for the model (here it's DLinear.yaml).
Set the mode to model evaluation: -o Global.mode=evaluate
Specify the validation dataset path: -o Global.dataset_dir

Other related parameters can be set by modifying the Global and Evaluate fields in the .yaml configuration file. For more details, refer to the PaddleX TS Configuration Parameters Documentation.

👉 More Details (Click to Expand)

When evaluating the model, you need to specify the model weight file path. Each configuration file has a default weight save path. If you need to change it, simply append the command line parameter, e.g., -o Evaluate.weight_path=./output/best_model/best_model.pdparams.

After model evaluation, the following outputs are typically produced:

evaluate_result.json: Records the evaluation results, specifically whether the evaluation task completed successfully and the model's evaluation metrics, including mse and mae.

4.4 Model Inference and Integration¶

After model training and evaluation, you can use the trained model weights for inference predictions or Python integration.

4.4.1 Model Inference¶

To perform inference predictions via the command line, use the following command:

Before running the following code, please download the demo csv to your local machine.

python main.py -c paddlex/configs/modules/ts_forecast/DLinear.yaml \
    -o Global.mode=predict \
    -o Predict.model_dir="./output/inference" \
    -o Predict.input="ts_fc.csv"

Similar to model training and evaluation, the following steps are required:

Specify the .yaml configuration file path of the model (here it's DLinear.yaml)
Set the mode to model inference prediction: -o Global.mode=predict
Specify the model weights path: -o Predict.model_dir="./output/best_accuracy/inference"

Specify the input data path: -o Predict.inputh="..." Other related parameters can be set by modifying the fields under Global and Predict in the .yaml configuration file. For details, refer to PaddleX Common Model Configuration File Parameter Description.

Alternatively, you can use the PaddleX wheel package for inference, easily integrating the model into your own projects.

4.4.2 Model Integration¶

The model can be directly integrated into the PaddleX pipeline or into your own projects.

Pipeline Integration

The Time Series Forecasting module can be integrated into PaddleX pipelines such as the Time Series Forecasting Pipeline (ts_fc). Simply replace the model path to update the Time Series Forecasting module's model.

Module Integration

The weights you produce can be directly integrated into the Time Series Forecasting module. You can refer to the Python sample code in Quick Integration and just replace the model with the path to the model you trained.

You can also use the PaddleX high-performance inference plugin to optimize the inference process of your model and further improve efficiency. For detailed procedures, please refer to the PaddleX High-Performance Inference Guide.