Guest User

Untitled

a guest
May 25th, 2018
78
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 0.96 KB | None | 0 0
  1. sar_m = sarimax.SARIMAX(df_train,
  2. trend='n',
  3. order=(2,1,1),
  4. seasonal_order=(2, 1, 1, 24),
  5. enforce_stationarity=False,
  6. enforce_invertibility=False,
  7. simple_differencing=False).fit()
  8. # predict for every hour of the next month
  9. predict_steps = 24*30
  10. forecast = sar_m.forecast(steps=predict_steps)
  11.  
  12. # plot against real data
  13. plot_forecast(df_test[:predict_steps], forecast,
  14. title='SARIMAX - Predicted vs Actual (September 2014)',
  15. xlabel='Day in September 2014',
  16. ylabel='Number of Pizza Orders')
  17.  
  18. # calculate RMSE error
  19. rmse(df_test[:predict_steps].numOfOrders, forecast)
  20.  
  21. date,numOfPizzaOrders
  22. 2014-04-01 00:00:00,12
  23. 2014-04-01 01:00:00,5
  24. 2014-04-01 02:00:00,2
  25. 2014-04-01 03:00:00,4
  26. 2014-04-01 04:00:00,3
  27. 2014-04-01 05:00:00,3
  28. 2014-04-01 06:00:00,7
  29. 2014-04-01 07:00:00,5
  30. 2014-04-01 08:00:00,17
Add Comment
Please, Sign In to add comment