https://github.com/nixtla logo
Join Slack
Powered by
# general
  • v

    Valeriy

    08/15/2022, 12:42 PM
    The link doesn't seem to work?
  • m

    Max (Nixtla)

    08/15/2022, 1:45 PM
    Hi Valery! We have been updating a lot of things in preparation for the next release. Some of the things we changed in the new version is a sklearn like syntax. Here you can find the getting started: https://nixtla.github.io/statsforecast/examples/Getting_Started_with_Auto_Arima_and_ETS.html
  • s

    Slackbot

    08/15/2022, 1:45 PM
    This message was deleted.
    👍 1
    v
    • 2
    • 1
  • n

    Nelson Elías Troncoso Martínez

    08/16/2022, 4:31 PM
    Hi everyone! Is there any way to use AutoARIMAProphet inside the class StatsForecast?
  • s

    Slackbot

    08/16/2022, 5:28 PM
    This message was deleted.
    😢 2
    👍 1
    n
    • 2
    • 1
  • s

    Slackbot

    08/17/2022, 2:48 PM
    This message was deleted.
    👏 2
    👍 2
    🙌 4
    🔥 3
    m
    k
    v
    • 4
    • 11
  • s

    Slackbot

    08/17/2022, 7:14 PM
    This message was deleted.
    v
    • 2
    • 1
  • s

    Slackbot

    08/17/2022, 9:24 PM
    This message was deleted.
    🙌 1
    k
    • 2
    • 1
  • v

    Valeriy

    08/17/2022, 9:24 PM
    How it ended. Lol
    🤣 2
  • k

    Kin Gtz. Olivares

    08/17/2022, 9:25 PM
    it is fun hahaha
  • m

    Max (Nixtla)

    08/17/2022, 11:04 PM
    We made it to the front page of Hacker News. Show some love by upvoting our post: “Show HN: Exp. Smoothing is 32% more accurate and 100x faster than Neural-Prophet”. https://news.ycombinator.com/
    👍 1
    🚀 1
  • s

    Slackbot

    08/18/2022, 9:51 AM
    This message was deleted.
    m
    l
    • 3
    • 7
  • j

    Jason

    08/18/2022, 2:14 PM
    Hello! Hope all is well.
  • s

    Slackbot

    08/18/2022, 2:15 PM
    This message was deleted.
    j
    m
    • 3
    • 5
  • s

    Slackbot

    08/18/2022, 6:06 PM
    This message was deleted.
    👀 3
    f
    g
    • 3
    • 4
  • s

    Slackbot

    08/19/2022, 9:45 AM
    This message was deleted.
    👀 1
    m
    a
    • 3
    • 3
  • s

    Slackbot

    08/23/2022, 11:28 AM
    This message was deleted.
    k
    • 2
    • 1
  • s

    Slackbot

    08/31/2022, 11:35 AM
    This message was deleted.
    👀 1
    f
    t
    • 3
    • 3
  • s

    Slackbot

    09/03/2022, 5:33 AM
    This message was deleted.
    k
    t
    • 3
    • 2
  • p

    Phi Nguyen

    09/08/2022, 4:02 PM
    I am getting an error when trying to run statsforecast (ETS) on the M5 data.
  • p

    Phi Nguyen

    09/08/2022, 4:02 PM
    Copy code
    ValueError                                Traceback (most recent call last)
    /home/ray/workspace-project-nixtla-m5/nixtla-m5-v2.ipynb Cell 10' in <cell line: 2>()
          1 init = time()
    ----> 2 Y_hat = fcst.forecast(28)
          3 end = time()
          4 print(f'Minutes taken by StatsForecast using: {(end - init) / 60}')
    
    File ~/anaconda3/lib/python3.8/site-packages/statsforecast/core.py:405, in StatsForecast.forecast(self, h, df, X_df, level, fitted, sort_df)
        403 X, level = self._parse_X_level(h=h, X=X_df, level=level)
        404 if self.n_jobs == 1:
    --> 405     res_fcsts = self.ga.forecast(models=self.models, h=h, fitted=fitted, X=X, level=level)
        406 else:
        407     res_fcsts = self._forecast_parallel(h=h, fitted=fitted, X=X, level=level)
    
    File ~/anaconda3/lib/python3.8/site-packages/statsforecast/core.py:148, in GroupedArray.forecast(self, models, h, fitted, X, level)
        146 if has_level:
        147     kwargs['level'] = level
    --> 148 res_i = model.forecast(h=h, y=y_train, X=X_train, X_future=X_f, fitted=fitted, **kwargs)
        149 cols_m = [key for key in res_i.keys() if any(key.startswith(m) for m in matches)]
        150 fcsts_i = np.vstack([res_i[key] for key in cols_m]).T
    
    File ~/anaconda3/lib/python3.8/site-packages/statsforecast/models.py:245, in ETS.forecast(self, y, h, X, X_future, fitted)
        237 def forecast(
        238         self, 
    ...
    --> 950 data_positive = min(y) > 0
        951 if (not data_positive) and errortype == 'M':
        952     raise ValueError('Inappropriate model for data with negative or zero values')
    
    ValueError: min() arg is an empty sequence
  • p

    Phi Nguyen

    09/08/2022, 4:03 PM
    I have filtered down my dataset to just one unique ID
  • p

    Phi Nguyen

    09/08/2022, 4:03 PM
    Copy code
    Y_df.describe()
  • p

    Phi Nguyen

    09/08/2022, 4:03 PM
    Copy code
    y
    count	1941.000000
    mean	10.786193
    std	1.259658
    min	10.000000
    25%	10.000000
    50%	10.000000
    75%	11.000000
    max	25.000000
  • p

    Phi Nguyen

    09/08/2022, 4:03 PM
    and running this code
  • p

    Phi Nguyen

    09/08/2022, 4:03 PM
    Copy code
    fcst = StatsForecast(
        df=Y_df, 
        models=[ETS(season_length=7, model='ZNA')], 
        freq='D'
    )
  • p

    Phi Nguyen

    09/08/2022, 4:04 PM
    Y_hat = fcst.forecast(28)
  • p

    Phi Nguyen

    09/08/2022, 4:04 PM
    I have drop the na rows in my dataset but it doesn’t look I have zeros of negative value
  • m

    Max (Nixtla)

    09/08/2022, 4:28 PM
    Hi @Phi Nguyen, could you share a head of your DF?
  • m

    Max (Nixtla)

    09/08/2022, 4:31 PM
    I have filtered down my dataset to just one unique ID
    Probably you droped the id column and that is causing the error. Just add a
    unique_id
    column with any value (e.g. “series_1”) so you have a DF with the following columns:`unique_id:string` ,
    ds:datestamp
    ,
    y:float
    .
12345...12Latest