search

AutoViML / Auto_TS

Automatically build ARIMA, SARIMAX, VAR, FB Prophet and XGBoost Models on Time Series data sets with a Single Line of Code. Created by Ram Seshadri. Collaborators welcome.
Apache License 2.0
729 stars 114 forks source link

Error while training auto_ts: Line search cannot locate an adequate point after 20 function and gradient evaluations #6

Closed vreddy-cs closed 4 years ago

vreddy-cs commented 4 years ago

I'm getting this error, while running auto_ts model training on Sales_and_Marketing.csv

Tit   = total number of iterations
Tnf   = total number of function evaluations
Tnint = total number of segments explored during Cauchy searches
Skip  = number of BFGS updates skipped
Nact  = number of active bounds at final generalized Cauchy point
Projg = norm of the final projected gradient
F     = final function value

           * * *

   N    Tit     Tnf  Tnint  Skip  Nact     Projg        F
    9     29     32      1     0     0   2.301D-04   1.150D+01
  F =   11.498879635408754

CONVERGENCE: REL_REDUCTION_OF_F_<=_FACTR*EPSMCH
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.
 This problem is unconstrained.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Bad direction in the line search;
   refresh the lbfgs memory and restart the iteration.

 Warning:  more than 10 function and gradient
   evaluations in the last line search.  Termination
   may possibly be caused by a bad search direction.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.

 Line search cannot locate an adequate point after 20 function
  and gradient evaluations.  Previous x, f and g restored.
 Possible causes: 1 error in function or gradient evaluation;
                  2 rounding error dominate computation.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Warning:  more than 10 function and gradient
   evaluations in the last line search.  Termination
   may possibly be caused by a bad search direction.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Warning:  more than 10 function and gradient
   evaluations in the last line search.  Termination
   may possibly be caused by a bad search direction.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

 Warning:  more than 10 function and gradient
   evaluations in the last line search.  Termination
   may possibly be caused by a bad search direction.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.
 This problem is unconstrained.

any idea, how to solve this issue?

AutoViML commented 4 years ago

Can you please cut and paste what was the auto-ts.Auto_TimeSeries() function call parameters that you used in order to hit this error? For example what was the time interval used, model type, etc. this will help me trouble shoot it. Please cut and paste as much as possible of your notebook. Thanks a lot Ram

On Thu, Jun 11, 2020 at 6:52 AM Venkat notifications@github.com wrote:

I'm getting this error, while running auto_ts model training on Sales_and_Marketing.csv https://github.com/AutoViML/Auto_TS/blob/master/example_datasets/Sales_and_Marketing.csv

Tit = total number of iterations Tnf = total number of function evaluations Tnint = total number of segments explored during Cauchy searches Skip = number of BFGS updates skipped Nact = number of active bounds at final generalized Cauchy point Projg = norm of the final projected gradient F = final function value

       * * *

N Tit Tnf Tnint Skip Nact Projg F 9 29 32 1 0 0 2.301D-04 1.150D+01 F = 11.498879635408754

CONVERGENCE: REL_REDUCTION_OFF<=_FACTR*EPSMCH This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration. This problem is unconstrained.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Bad direction in the line search; refresh the lbfgs memory and restart the iteration.

Warning: more than 10 function and gradient evaluations in the last line search. Termination may possibly be caused by a bad search direction. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained.

Line search cannot locate an adequate point after 20 function and gradient evaluations. Previous x, f and g restored. Possible causes: 1 error in function or gradient evaluation; 2 rounding error dominate computation. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Warning: more than 10 function and gradient evaluations in the last line search. Termination may possibly be caused by a bad search direction. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Warning: more than 10 function and gradient evaluations in the last line search. Termination may possibly be caused by a bad search direction. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

Warning: more than 10 function and gradient evaluations in the last line search. Termination may possibly be caused by a bad search direction. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained. This problem is unconstrained.

any idea, how to solve this issue?

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/AutoViML/Auto_TS/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AMKBH6BL4LAPIKYMLUCRQIDRWCZQBANCNFSM4N3JPWUA .

vreddy-cs commented 4 years ago

@AutoViML , these are call params inside the Auto_Timeseries

ml_dict = AT.Auto_Timeseries(train_data, date_column,
    target_name, ",", score_type = 'rmse', forecast_period = 8,
    time_interval = 'Month', non_seasonal_pdq = None, seasonality = False,
    seasonal_period = 12, seasonal_PDQ = None, model_type = 'best',
    verbose = 1)
AutoViML commented 4 years ago

Sorry couldn't reproduce it - this runs fine on our desktop/laptop. You might want to run it on Colab and share a link to that colab so we can reproduce it. At the moment, suspicion is on PyFlux - may be it is not installed properly on your machine or you might have an older version of statsmodels. Can't tell for sure.

ngupta23 commented 4 years ago

I was able to reproduce this error. It is coming from the optimizer that is being used by the ARIMA (and other functions) used in the statsmodels package. Changing the solver from 'lbfgs' to 'bfgs' resolves the warning messages.

However the result is not very different with the different solver. My unit test are giving an RMSE of 169.09 with lbfgs vs an RMSE of 169.00 for bfgs. So I am not sure that it is a real issue. I can add an option to customize this in the next release, but I feel the ROI will be minimal.

ngupta23 commented 4 years ago

Closing this issue since it is coming from optimizer. Will reopen is this turns out to be a major issue.