CoolFace
Datasetpublic

Real-TSF/TIME-Output

This repository contains the extracted time series features (tsfeatures) for each variate and the detailed forecasting results for every experiment. Note: These files are for building leaderboard and visualization; users do not need to download this directory. features/: Statistical Features (tsfeatures) Each dataset's features are saved to: output/features/{dataset}/{freq}/. This directory stores the computed tsfeatures for the variates in the dataset. The folder contains a CSV file… See the full description on the dataset page: https://huggingface.co/datasets/Real-TSF/TIME-Output.

sourceHugging Faceapache-2.0updated 13h agoView on Hugging Face
0likes2.1kdownloads
README.md29 linesDownload Raw Back to root
1---
2license: apache-2.0
3task_categories:
4- time-series-forecasting  # 这是让系统识别为时间序列任务的关键
5tags:                      # 自定义标签
6- time-series
7- forecasting
8- benchmark
9---
10
11This repository contains the extracted time series features (tsfeatures) for each variate and the detailed forecasting results for every experiment.
12
13**Note**: These files are for building leaderboard and visualization; users do not need to download this directory.
14
15* `features/`: **Statistical Features (tsfeatures)**
16
17    Each dataset's features are saved to: `output/features/{dataset}/{freq}/`.
18    This directory stores the computed tsfeatures for the variates in the dataset. The folder contains a CSV file named either:
19    * `test.csv`: Features calculated on the **test split**.
20    * `full.csv`: Features calculated on the **full dataset**.
21
22* `results/`: **Experimental Results**
23
24    Each experiment is saved to: `output/results/{model}/{dataset}/{freq}/{horizon}/`.
25    Each directory represents a specific model-dataset-frequency-horizon combination and contains the following files:
26    * `config.json`: The experiment settings and hyperparameters.
27    * `metrics.npz`: Window-level forecasting metrics (e.g., MASE, CRPS).
28    * `predictions.npz`: Quantile prediction results for each sliding window.
29