Why the predicted values show an unrealistic
The predicted values show an unrealistic, overly smooth trend with almost linear progression starting from 2024-07-27. This pattern fails to reflect the significant variability and periodic trends observed in real outpatient visit data.
and in another issue, why torch model only could load on num_layers=50 even can not run on default layers
sorry for that i know for sure that's is pretty dumb question but i really facing problems on how can i use this model
your help is really appreciated
Can you provide the time-series as an array? I can try forecasting and see if we get the same result
okay, i just using pandas convert to json show that to you.
asct1.to_dict(orient='records')
[{'ZB1': 18, 'ds': Timestamp('2024-07-01 00:00:00'), 'unique_id': 11}, {'ZB1': 35, 'ds': Timestamp('2024-07-02 00:00:00'), 'unique_id': 24}, {'ZB1': 66, 'ds': Timestamp('2024-07-03 00:00:00'), 'unique_id': 21}, {'ZB1': 58, 'ds': Timestamp('2024-07-04 00:00:00'), 'unique_id': 22}, {'ZB1': 50, 'ds': Timestamp('2024-07-05 00:00:00'), 'unique_id': 0}, {'ZB1': 63, 'ds': Timestamp('2024-07-06 00:00:00'), 'unique_id': 9}, {'ZB1': 89, 'ds': Timestamp('2024-07-07 00:00:00'), 'unique_id': 12}, {'ZB1': 61, 'ds': Timestamp('2024-07-08 00:00:00'), 'unique_id': 15}, {'ZB1': 87, 'ds': Timestamp('2024-07-09 00:00:00'), 'unique_id': 16}, {'ZB1': 102, 'ds': Timestamp('2024-07-10 00:00:00'), 'unique_id': 1}, {'ZB1': 107, 'ds': Timestamp('2024-07-11 00:00:00'), 'unique_id': 13}, {'ZB1': 105, 'ds': Timestamp('2024-07-12 00:00:00'), 'unique_id': 25}, {'ZB1': 87, 'ds': Timestamp('2024-07-13 00:00:00'), 'unique_id': 2}, {'ZB1': 94, 'ds': Timestamp('2024-07-14 00:00:00'), 'unique_id': 23}, {'ZB1': 88, 'ds': Timestamp('2024-07-15 00:00:00'), 'unique_id': 3}, {'ZB1': 100, 'ds': Timestamp('2024-07-16 00:00:00'), 'unique_id': 6}, {'ZB1': 116, 'ds': Timestamp('2024-07-17 00:00:00'), 'unique_id': 10}, {'ZB1': 99, 'ds': Timestamp('2024-07-18 00:00:00'), 'unique_id': 4}, {'ZB1': 107, 'ds': Timestamp('2024-07-19 00:00:00'), 'unique_id': 26}, {'ZB1': 98, 'ds': Timestamp('2024-07-20 00:00:00'), 'unique_id': 27}, {'ZB1': 95, 'ds': Timestamp('2024-07-21 00:00:00'), 'unique_id': 7}, {'ZB1': 93, 'ds': Timestamp('2024-07-22 00:00:00'), 'unique_id': 28}, {'ZB1': 99, 'ds': Timestamp('2024-07-23 00:00:00'), 'unique_id': 17}, {'ZB1': 94, 'ds': Timestamp('2024-07-24 00:00:00'), 'unique_id': 19}, {'ZB1': 100, 'ds': Timestamp('2024-07-25 00:00:00'), 'unique_id': 8}, {'ZB1': 111, 'ds': Timestamp('2024-07-26 00:00:00'), 'unique_id': 5}, {'ZB1': 87, 'ds': Timestamp('2024-07-27 00:00:00'), 'unique_id': 20}, {'ZB1': 90, 'ds': Timestamp('2024-07-28 00:00:00'), 'unique_id': 14}, {'ZB1': 92, 'ds': Timestamp('2024-07-29 00:00:00'), 'unique_id': 29}, {'ZB1': 76, 'ds': Timestamp('2024-07-30 00:00:00'), 'unique_id': 30}, {'ZB1': 55, 'ds': Timestamp('2024-07-31 00:00:00'), 'unique_id': 18}]
Can you provide the time-series as an array? I can try forecasting and see if we get the same result