-
Can you please clarify the following situation? I am trying to build a model with SARIMAX including exog variables and having difficulties to determine the orders as I do not have access to regression…
-
#2836 works around the missing gfortran and gets full test results again
only some failures in statespace,
- 2 small precision problems
- test_sarimax.TestFriedmanStateRegression.test_mle large fai…
-
Recent versions of numpy / scipy / pandas have issues deprecation or future warnings that affect the state space files, including:
- Various places: FutureWarning: TimeSeries is deprecated. Please use…
-
**soon**
- [x] `specification` dict should be a Bunch, IMO (I don't have tab completion for dict keys)
- [ ] related to `specification` dict: `_init_kwds` for recreating model (generic, not used in ts…
-
Hello all!
I just upgraded to pandas 0.17 and I have issues with the `filter` method from the `SARIMAX` class. This notebook is throwing errors (and my code too): https://github.com/statsmodels/stats…
-
The code below crashes on the last line but changing the SARIMAX object to an ARIMA object (see commented line) makes the last line execute successfully. Possibly a bug?
``` python
import matplotlib…
-
In the diagnostics branch that is awaiting merge, sarimax.predict breaks when simple_differencing=True. You get the size matching error like
ValueError: Invalid dimensions for time-varying state inte…
-
Steps:
I am a beginner in using Python for TSA. Had Anaconda packaged 0.6 statsmodels. Wanted SARIMAX functionality and tried building and installing it from source. Build is fine but cannot see any …
-
In multivariate state space models with variance parameters, complex step derivatives can fail because the Cholesky factorization (zpotrf) silently ignores the complex part. This is because zpotrf ass…
-
I'm not able to replicate dynamic prediction with exog from Stata.
My impression is that the exog slice is wrong in dynamic prediction, or I still doen't understand this.
ARMA with dynamic=False agr…