Datasets:

ServiceNow
/

context-is-key

Tasks:

Time Series Forecasting

Modalities:

Formats:

Languages:

Size:

ArXiv:

Libraries:

License:

Dataset card Data Studio Files Files and versions

marcotet commited on Jul 11

Commit

744e9a2

·

verified ·

1 Parent(s): 8ac49a8

Update README.md

Files changed (1) hide show

README.md +6 -1

README.md CHANGED Viewed

@@ -104,4 +104,9 @@ Users of the benchmark should only gives the *background*, *scenario*, *constrai
 together with the timestamps of *future_time*.
 The other features are there to compute the RCPRS metric and classification of the tasks.
-Note: to convert *past_time* and *future_time* to Pandas DataFrame, use the following snipet: `pd.read_json(StringIO(entry["past_time"]))`.

 together with the timestamps of *future_time*.
 The other features are there to compute the RCPRS metric and classification of the tasks.
+Note: to convert *past_time* and *future_time* to Pandas DataFrame, use the following snipet: `pd.read_json(StringIO(entry["past_time"]))`.
+## Computing the RCPRS metric
+Code to compute the RCPRS metric is available in the [`compute_rcrps_with_hf_dataset.py`](https://huggingface.co/datasets/ServiceNow/context-is-key/blob/main/compute_rcrps_with_hf_dataset.py) script inside this dataset repository.
+Please look at the `__main__` section of the script to see an example on how to use it.