Skip to content

Latest commit

 

History

History
73 lines (47 loc) · 2.45 KB

File metadata and controls

73 lines (47 loc) · 2.45 KB

Experimental Protocol

This document describes the evaluation protocol used in SpikingTSF for long-term time-series forecasting.


Datasets

Dataset Domain Frequency Variables
ETTh1 Electricity (hourly) 1h 7
ETTh2 Electricity (hourly) 1h 7
ETTm1 Electricity (15-min) 15min 7
ETTm2 Electricity (15-min) 15min 7

Datasets are sourced from the Autoformer repository. See DATASETS.md for placement instructions.


Forecasting Setting

  • Task: Multivariate long-term forecasting (-M flag)
  • Look-back window (input length): 96
  • Prediction horizons: 96, 192, 336, 720
  • Train / Validation / Test split: Standard ETT protocol as implemented in data_provider/ETT_data_loader.py. Refer to the data loader for exact split boundaries.

Evaluation Metrics

All metrics are computed on the test set:

Metric Description Direction
MSE Mean Squared Error Lower is better
MAE Mean Absolute Error Lower is better
RMSE Root Mean Squared Error Lower is better
Coefficient of Determination Higher is better

Implementation: utils/metrics.py.


Normalization

Per-sample instance normalization is applied in the data loader. This is consistent across all models.


Reproducibility

  • Seeds: 42, 1234, 3407 (3 independent runs per configuration)
  • Reported values: Mean ± standard deviation across 3 seeds
  • Hyperparameters are tuned per model per dataset using Optuna (see hparam/). The best hyperparameter configuration is then used for the 3-seed evaluation.
  • All models use the same look-back window, prediction horizon, and dataset split.

Hyperparameter Policy

  • Hyperparameter tuning is performed independently per model and per dataset.
  • Tuned hyperparameters include: learning rate, weight decay, batch size, number of layers, attention dimension, scheduler, and other model-specific parameters.
  • The best trial from the Optuna study is used for the final 3-seed evaluation.

Reporting

  • Numbers are reported as mean ± std across 3 runs.
  • Results are logged to Output/<ModelName>/<dataset>_pl<horizon>.txt.
  • Do not report results from partial runs or single seeds as final results.
  • Results from this library should not be used in publications without verifying them independently.