Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Ross Taylor
    @RJT1990
    Welcome to the PyFlux gitter chat! Have any questions? Any feature requests? Any bugs? Shout here.
    Ross Taylor
    @RJT1990
    There was an issue for version 0.4.5 for Python 2.7 because of some faulty imports. Looking to rectify this.
    Afterwards- not intending for big updates for a while (taking a break). But my main focus when returning will be (a) speeding up existing models, (b) finishing tests and implementing continuous integration, (c) implementing neural network AR models and (d) making a feature creation tool (which I am excited about)
    Ad34
    @ad34
    Hello, I just discovered pyflux , is it possible to use the algorithms on data which does not comes from a pandas dataframe , let s say I ve got a 1D numpy array
    Ross Taylor
    @RJT1990
    Hello - the models with exogenous variables currently require pandas input; everything else works with np.arrays
    Ad34
    @ad34
    Thanks @RJT1990
    Adam Orr
    @adamjorr
    Hi, how do you properly use a uniform prior? I'm trying model.adjust_prior(1,pf.Uniform()) and getting AttributeError: 'module' object has no attribute 'Uniform'. Is there an import I'm missing?
    Pranav Shekhar Bahl
    @pranavbahl2308
    Hi I recently came across pyflux and had few questions about its working
    1) Is there any way to automate determination of p,d,q values for ARIMA ?
    2) How to run SARIMA in pyflux
    Ross Taylor
    @RJT1990
    Hey - apologies for late responses. API is now pf.Flat()
    @pranavbahl2308 (1) No automatic determination yet - no - way to do this right now is to create multiple models (loop over them) and evaluate by the AIC which is stored in the results object. E.g. x = model.fit() -> x.aic
    (2) SARIMA - not yet implemented; will probably implement it sooner or later (along with seasonality for models in general)
    Stuart Gordon Reid
    @StuartGordonReid
    Hello,
    I just started using PyFlux last night and I was wondering if you could let me know how many of the models are designed with concurrency in mind?
    Stuart Gordon Reid
    @StuartGordonReid
    I'm asking because I ran a whole bunch of ARIMA models yesterday (50 models * ~1500 time steps) and the CPU load was ~2.7 out of 8. I do have a compiled BLAS on my machine and get close to 100% core utilization when working with most Numpy functions. Thanks in advance :-)
    Aalto
    @aalto1
    Hi I am trying to follow the ARIMAX tutorial but it seems the expression definition in the ARIMAX call is malformed
    in particular i get this assertionError
    ~/anaconda3/envs/denv/lib/python3.6/site-packages/patsy/tokens.py in python_tokenize(code)
         33                 break
         34             origin = Origin(code, start, end)
    ---> 35             assert pytype not in (tokenize.NL, tokenize.NEWLINE)
         36             if pytype == tokenize.ERRORTOKEN:
         37                 raise PatsyError("error tokenizing input "
    
    AssertionError:
    Aalto
    @aalto1
    ok it was a pasty problem
    not pyflux
    idrigo
    @idrigo
    Hi, guys. Trying to import PyFlux after installation and receive following error: ImportError: cannot import name 'gas_recursion_normal_orderone'. Using MacOS python 3.6.8