Improving accuracy of predictions using seasonal relationships of time series data

• US 10,474,968 B2
• Filed: 12/04/2018
• Issued: 11/12/2019
• Est. Priority Date: 04/17/2014
• Status: Active Grant

• Current Assignee: SAS Institute Incorporated
• Original Assignee: SAS Institute Incorporated
• Inventors: Chien, Yung-Hsin, Wang, Pu, Li, Yue
• Primary Examiner(s): Waldron, Scott A.
• Assistant Examiner(s): Lamardo, Viker A
• Application Number:
  US16/209,716
• Publication Number:
  US 20190108460A1
• Time in Patent Office: 343 Days
• Field of Search: 706 12
• CPC Class Codes: G06F 16/285 Clustering or classification
• CPC Class Codes: G06F 2216/03 Data mining
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models
• CPC Class Codes: G06Q 10/063 Operations research, analysis or management
• CPC Class Codes: G06Q 30/0202 Market predictions or forecasting for commercial activities

Abstract

Systems and methods are provided for performing data mining and statistical learning techniques on a big data set. More specifically, systems and methods are provided for linear regression using safe screening techniques. Techniques may include receiving a plurality of time series included in a prediction hierarchy for performing statistical learning to develop an improved prediction hierarchy. It may include pre-processing data associated with each of the plurality of time series, wherein the pre-processing includes tasks performed in parallel using a grid-enabled computing environment. For each time series, the system may determine a classification for the individual time series, a pattern group for the individual time series, and a level of the prediction hierarchy at which the each individual time series comprises an need output amount greater than a threshold amount. The computing system may generate an additional prediction hierarchy using the first prediction hierarchy, the classification, the pattern group, and the level.

Claims

1. A system for performing data mining and statistical learning techniques on a data set, the system comprising:

a processor; and

a non-transitory computer-readable storage medium including instructions stored thereon, which when executed by the processor, cause the system to perform operations including;

receiving a time series for performing statistical learning to develop improved object prediction intervals, wherein the time series includes one or more demand characteristics and a demand pattern for an object;

pre-processing data associated with the time series, wherein the pre-processing includes tasks performed in parallel using a grid-enabled computing environment;

determining a number of low demand periods within the time series, a low demand period being a time interval for which demand for the object is less than a threshold value;

determining a series type for the time series using the number of low demand periods;

determining an in-season interval of the time series using the number of low demand periods and the series type, the in-season interval indicating a demand period for which demand for the object has historically been greater than the threshold value;

deriving a future in-season interval using the determined in-season interval, the future in-season interval being a predicted time interval during which demand for the object is predicted to be greater than the threshold value; and

transmitting, to one or more nodes in the grid-enabled computing environment, prediction data related to the time series based on the derived future in-season interval, wherein the derived future in-season interval provides user control of the data set when the derived future in-season interval is applied to the data set and characteristics of the data set.

2. The system of claim 1, wherein determining the number of low demand periods within the time series further includes determining an approximate time series utilizing a segmentation algorithm and the time series, wherein the number of low demand periods are determined using the approximate time series.

3. The system of claim 1, wherein determining the number of low demand periods further includes determining whether the demand period exceeds a threshold amount of time.

4. The system of claim 1, wherein the time series includes time series aggregated data from a plurality of time series, the plurality of time series individually comprising at least one demand characteristic and at least one demand pattern for an object.

5. The system of claim 1, wherein the operations further include identifying an incomplete seasonal period of the time series, the incomplete seasonal period beginning with data that is greater than the threshold value.

6. The system of claim 5, wherein the operations further include excluding the incomplete seasonal period from the time series.

7. The system of claim 1, wherein the operations further include:

determining a mean period for the in-season interval of the time series;

associating the time series with a seasonal series type based on a determination that the mean period is greater than a seasonal threshold amount; and

associating the time series with an event series type based on a determination that the mean period is less than the seasonal threshold amount.

8. The system of claim 7, wherein the operations further include determining, when the time series is associated with the seasonal series type, a season start index corresponding to a day of a year on which a season begins, wherein determining the in-season interval is based on the season start index.

9. The system of claim 7, wherein the operations further include, when the time series is associated with the event series type:

determining an event index corresponding to a day of a year on which an event occurs in the time series;

determining a season start index for the in-season interval based on the event index; and

determining a season end index for the in-season interval based on th the event index, wherein the determined in-season interval is further based on the season start index and the season end index.

10. The system of claim 1, wherein determining the in-season interval using the number of low demand periods and the series type further include receiving user input selecting a calculation of at least one of a minimum period, a maximum period, a mean period, and a mode period.

11. A computer-program product tangibly embodied in a non-transitory machine-readable storage medium, the computer-program product including instructions for performing data mining and statistical learning techniques on a data set, the instructions configured to be executed to cause a data processing apparatus to perform operations including:

receiving a time series for performing statistical learning to develop improved object prediction intervals, wherein the time series includes one or more demand characteristics and a demand pattern for an object;

pre-processing data associated with the time series, wherein the pre-processing includes tasks performed in parallel using a grid-enabled computing environment, the tasks comprising;

determining a number of low demand periods within the time series, a low demand period being a time interval for which demand for the object is less than a threshold value;

determining a series type for the time series using the number of low demand periods;

determining an in-season interval of the time series using the number of low demand periods and the series type, the in-season interval indicating a demand period for which demand for the object has historically been greater than the threshold value;

deriving a future in-season interval using the determined in-season interval, the future in-season interval being a predicted time interval during which demand for the object is predicted to be greater than the threshold value; and

transmitting, to one or more nodes in the grid-enabled computing environment, prediction data related to the time series based on the derived future in-season interval, wherein the derived future in-season interval provides user control of the data set when the derived future in-season interval is applied to the data set and characteristics of the data set.

12. The computer-program product of claim 11, wherein determining the number of low demand periods within the time series further includes determining an approximate time series utilizing a segmentation algorithm and the time series, wherein the number of low demand periods are determined using the approximate time series.

13. The computer-program product of claim 11, wherein determining the number of low demand periods further includes determining whether the demand period exceeds a threshold amount of time.

14. The computer-program product of claim 11, wherein the time series includes time series aggregated data from a plurality of time series, the plurality of time series individually comprising at least one demand characteristic and at least one demand pattern for an object.

15. The computer-program product of claim 11, wherein the operations further include identifying an incomplete seasonal period of the time series, the incomplete seasonal period beginning with data that is greater than the threshold value.

16. The computer-program product of claim 15, wherein the operations further include excluding the incomplete seasonal period from the time series.

17. The computer-program product of claim 11, wherein the operations further include:

determining a mean period for the in-season interval of the time series;

associating the time series with a seasonal series type based on a determination that the mean period is greater than a seasonal threshold amount; and

associating the time series with an event series type based on a determination that the mean period is less than the seasonal threshold amount.

18. The computer-program product of claim 17, wherein the operations further include determining, when the time series is associated with the seasonal series type, a season start index corresponding to a day of a year on which a season begins, wherein determining the in-season interval is based on the season start index.

19. The computer-program product of claim 17, wherein the operations further include, when the time series is associated with the event series type:

determining an event index corresponding to a day of a year on which an event occurs in the time series;

determining a season start index for the in-season interval based on the event index; and

determining a season end index for the in-season interval based on the event index, wherein the determined in-season interval is further based on the season start index and the season end index.

20. The computer-program product of claim 11, wherein determining the in-season interval using the number of low demand periods and the series type further include receiving user input selecting a calculation of at least one of a minimum period, a maximum period, a mean period, and a mode period.

21. A computer-implemented method for performing data mining and statistical learning techniques on a data set, comprising:

receiving a time series for performing statistical learning to develop improved object prediction intervals, wherein the time series includes one or more demand characteristics and a demand pattern for an object;

pre-processing data associated with the time series, wherein the pre-processing includes tasks performed in parallel using a grid-enabled computing environment, the tasks comprising;

determining a number of low demand periods within the time series, a low demand period being a time interval for which demand for the object is less than a threshold value;

determining a series type for the time series using the number of low demand periods;

determining an in-season interval of the time series using the number of low demand periods and the series type, the in-season interval indicating a demand period for which demand for the object has historically been greater than the threshold value;

deriving a future in-season interval using the determined in-season interval, the future in-season interval being a predicted time interval during which demand for the object is predicted to be greater than the threshold value; and

transmitting, to one or more nodes in the grid-enabled computing environment, prediction data related to the time series based on the derived future in-season interval, wherein the derived future in-season interval provides user control of the data set when the derived future in-season interval is applied to the data set and characteristics of the data set.

22. The computer-implemented method of claim 21, wherein determining the number of low demand periods within the time series further includes determining an approximate time series utilizing a segmentation algorithm and the time series, wherein the number of low demand periods are determined using the approximate time series.

23. The computer-implemented method of claim 21, wherein determining the number of low demand periods further includes determining whether the demand period exceeds a threshold amount of time.

24. The computer-implemented method of claim 21, wherein the time series includes time series aggregated data from a plurality of time series, the plurality of time series individually comprising at least one demand characteristic and at least one demand pattern for an object.

25. The computer-implemented method of claim 21, further comprising identifying an incomplete seasonal period of the time series, the incomplete seasonal period beginning with data that is greater than the threshold value.

26. The computer-implemented method of claim 25, further comprising excluding the incomplete seasonal period from the time series.

27. The computer-implemented method of claim 21, further comprising:

determining a mean period for the in-season interval of the time series;

associating the time series with a seasonal series type based on a determination that the mean period is greater than a seasonal threshold amount; and

associating the time series with an event series type based on a determination that the mean period is less than the seasonal threshold amount.

28. The computer-implemented method of claim 27, further comprising determining, when the time series is associated with the seasonal series type, a season start index corresponding to a day of a year on which a season begins, wherein determining the in-season interval is based on the season start index.

29. The computer-implemented method of claim 27, further comprising, when the time series is associated with the event series type:

determining an event index corresponding to a day of a year on which an event occurs in the time series;

determining a season start index for the in-season interval based on the event index; and

determining a season end index for the in-season interval based on the event index, wherein the determined in-season interval is further based on the season start index and the season end index.

30. The computer-implemented method of claim 21, wherein determining the in-season interval using the number of low demand periods and the series type further include receiving user input selecting a calculation of at least one of a minimum period, a maximum period, a mean period, and a mode period.