Time series analysis and forecasting program
View Patent ↗A method and apparatus for performing forecasting using time series data are described. In one embodiment, the method includes designating a last portion of the time series data as a base pattern, searching the remaining portion of the time series data for candidate patterns using the base pattern, and identifying candidate forecasts corresponding to the candidate patterns. The method further includes calculating a set of final forecast values based on the candidate forecasts.
1. A computer program product comprising a computer useable medium having computer program logic recorded thereon for enabling a processor in a computer system to perform forecasting, said computer program logic comprising: means for enabling the processor to designate a portion of time series data as a base pattern; means for enabling the processor to identify one or more candidate patterns from a remaining portion of the time series data based on a comparison with the base pattern; means for enabling the processor to identify one or more candidate patterns from a remaining portion of the time series data based on a comparison with the base pattern; means for enabling the processor to construct a forecast based on one or more candidate forecasts, wherein each of the one or more candidate forecasts is associated with a corresponding candidate pattern; and means for storing the forecast in memory for further processing or display.
2. The computer program product of claim 1 , further comprising:
means for enabling the processor to receive a number of time intervals for which a forecast is requested.
3. The computer program product of claim 2 , wherein the means for enabling the processor to designate a portion of time series data as a base pattern comprises:
means for enabling the processor to designate a portion of time series data having a length that is the same as the number of time intervals for which the forecast is requested as a base pattern.
4. The computer program product of claim 1 , further comprising:
means for enabling the processor to designate a pattern following each of the one or more candidate patterns as a candidate forecast.
5. The computer program product of claim 1 , further comprising:
means for enabling the processor to assign a weight to each of the one or more candidate patterns.
6. The computer program product of claim 5 , further comprising:
means for enabling the processor to select the one or more candidate forecasts based on the weight assigned to each candidate pattern associated with the one or more candidate forecasts.
7. The computer program product of claim 1 , further comprising:
means for enabling the processor to remove a seasonal effect from the time series data.
8. The computer program product of claim 1 , further comprising:
means for enabling the processor to remove noise from the time series data.
9. A computerized forecasting method comprising designating a portion of time series data as a base pattern; identifying one or more candidate patterns from a remaining portion of the time series data based on a comparison with the base pattern; constructing a forecast based on one or more candidate forecasts, wherein each of the one or more candidate forecasts is associated with a corresponding candidate pattern; and storing the forcast in memory for further processing or display.
10. The method of claim 9 , further comprising:
receiving a number of time intervals for which a forecast is requested.
11. The method of claim 10 , wherein designating a portion of time series data as a base pattern comprises:
designating a portion of time series data having a length that is the same as the number of time intervals for which the forecast is requested as a base pattern.
12. The method of claim 9 , further comprising:
designating a pattern following each of the one or more candidate patterns as a candidate forecast.
13. The method of claim 9 , further comprising:
assigning a weight to each of the one or more candidate patterns.
14. The method of claim 13 , further comprising: selecting the one or more candidate forecasts based on the weight assigned to each candidate pattern associated with the one or more candidate forecasts.
15. The method of claim 9 , further comprising:
removing a seasonal effect from the time series data.
16. The method of claim 9 , further comprising:
removing noise from the time series data.