Robust forecasting techniques with reduced sensitivity to anomalous data
First Claim
1. A computer-implemented method of computing a forecast F(t+x) for a second time t+x from at least an observation y(t) at a first time t earlier in time than the second time t+x, where x is a forecast time horizon, the method comprising:
- obtaining from a computer memory a first plurality of seasonality factors from prior seasons for the first time t;
determining a first seasonal adjustment factor Id(t) from the first plurality of seasonality factors;
computing a level component L(t) using at least the first seasonal adjustment factor Id(t) and the observation y(t);
obtaining a second plurality of seasonality factors from prior seasons for the second time t+x;
determining a second seasonal adjustment factor Id(t+x) from the second plurality of seasonality factors; and
computing, by execution of software by one or more processors, at least a first forecast F1(t+x) at least partially based on the level component L(t) and the second seasonal adjustment factor Id(t+x), wherein the first forecast F1(t+x) is considered in determining the computed forecast F(t+x).
1 Assignment
0 Petitions
Accused Products
Abstract
Robust forecasting techniques are relatively immune from anomalies or outliers in observed data, such as a stream of data values reflective of the operation or use of a computer system. One robust technique provides a relatively accurate forecast of seasonal behavior even in the presence of an anomaly in corresponding historical data. Another robust forecasting technique provides a relatively accurate forecast even in the presence of an anomaly that spans multiple recent observations. In one embodiment, both techniques are used in combination to automatically detect anomalies in the operation and/or use of a multi-user computer system.
104 Citations
47 Claims
-
1. A computer-implemented method of computing a forecast F(t+x) for a second time t+x from at least an observation y(t) at a first time t earlier in time than the second time t+x, where x is a forecast time horizon, the method comprising:
-
obtaining from a computer memory a first plurality of seasonality factors from prior seasons for the first time t; determining a first seasonal adjustment factor Id(t) from the first plurality of seasonality factors; computing a level component L(t) using at least the first seasonal adjustment factor Id(t) and the observation y(t); obtaining a second plurality of seasonality factors from prior seasons for the second time t+x; determining a second seasonal adjustment factor Id(t+x) from the second plurality of seasonality factors; and computing, by execution of software by one or more processors, at least a first forecast F1(t+x) at least partially based on the level component L(t) and the second seasonal adjustment factor Id(t+x), wherein the first forecast F1(t+x) is considered in determining the computed forecast F(t+x). - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34)
-
-
35. A computer-readable medium having stored thereon a set of program modules that, when executed by a computer, cause the computer to perform a method of computing a forecast F(t+x) for a second time t+x from at least an observation y(t) at a first time t earlier in time than the second time t+x, where x is a forecast time horizon, the method comprising:
-
obtaining from a computer memory a first plurality of seasonality factors from prior seasons for the first time t; determining a first seasonal adjustment factor Id(t) from the first plurality of seasonality factors; computing a level component L(t) using at least the first seasonal adjustment factor Id(t) and the observation y(t); obtaining a second plurality of seasonality factors from prior seasons for the second time t+x determining a second seasonal adjustment factor Id(t+x) from the second plurality of seasonality factors; and computing at least a first forecast F1(t+x) at least partially based on the level component L(t) and the second seasonal adjustment factor Id(t+x), wherein the first forecast F1(t+x) is considered in determining the computed forecast F(t+x).
-
-
36. A computer-implemented method of computing a forecast F(t+x) for a second time t+x from at least an observation y(t) at a first time t earlier in time than the second time t+x, where x is a forecast time horizon, the method comprising:
-
collecting a predetermined number of seasonality factors from prior seasons from a computer memory for the first time t, where a seasonality factor corresponds to a time that is an integer multiple of seasonal periods from the first time t; deseasonalizing the observation y(t) at the first time t using fewer than all of the collected seasonality factors; computing a level component using the observation y(t) as deseasonalized; and computing, by execution of software by one or more processors, at least a first forecast F1(t+x) at least partially based on the level component and a second seasonal adjustment factor for the second time t+x. - View Dependent Claims (37, 38, 39, 40, 41, 42, 43)
-
-
44. A computer-readable medium having stored thereon a set of program modules that, when executed by a computer, cause the computer to perform a method of computing a forecast F(t+x) for a second time t+x from at least an observation y(t) at a first time t earlier in time than the second time t+x, where x is a forecast time horizon, the method comprising:
-
collecting a predetermined number of seasonality factors from prior seasons from a computer memory for the first time t, where a seasonality factor corresponds to a time that is an integer multiple of seasonal periods from the first time t; deseasonalizing the observation y(t) at the first time t using fewer than all of the collected seasonality factors; computing a level component using the observation y(t) as deseasonalized; and computing at least a first forecast F1(t+x) at least partially based on the level component and a second seasonal adjustment factor for the second time t+x.
-
-
45. A computer-implemented method of computing a forecast for a time t+x from a time t for a monitored system, the method comprising:
-
computing a seasonal adjustment factor for the time t+x and a seasonal adjustment factor for the time t, wherein at least one of the computed seasonal adjustment factors is computed from a plurality of seasonality factors such that the at least one computed seasonal adjustment factor is representative of a normal operating condition of the monitored system; and using the computed seasonal adjustment factors to compute the forecast by execution of software by one or more processors. - View Dependent Claims (46, 47)
-
Specification