Methods, systems and computer readable code for forecasting time series and for forecasting commodity consumption

• US 8,108,243 B2
• Filed: 06/19/2005
• Issued: 01/31/2012
• Est. Priority Date: 06/18/2004
• Status: Active Grant

• Current Assignee: Amdocs Development Limited (Amdocs Ltd.)
• Original Assignee: Cvidya Networks
• Inventors: Solotorevsky, Gad, Pedan, Sveta
• Primary Examiner(s): Jeanty, Romain
• Application Number:
  US11/570,775
• Publication Number:
  US 20080059284A1
• Time in Patent Office: 2,417 Days
• Field of Search: 705/7.29
• US Class Current: 705/7.29
• CPC Class Codes: G06Q 10/04 Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem" market predictions or forecasting for commercial activities G06Q30/0202
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data

Abstract

Methods, systems and computer readable code for forecasting commodity consumption and for forecasting time series are provided. According to some embodiments, the forecasting includes deriving at least one population commodity consumption forecasting model from population historical consumption data, deriving an individual commodity consumption forecasting model for at least one individual of the population from at least one population commodity consumption forecasting model and from individual historical consumption data, and forecasting future individual commodity consumption for the individual using individual commodity consumption forecasting model. According to some embodiments, the presently disclosed forecasting includes forecasting future values of an individual time series within a population of time series, where each time series on the same domain. Thus, according to some embodiments, the forecasting includes deriving at least one population forecasting model from past values of the population of time series, deriving an individual time series for at least one individual time series forecasting model from past individual time series values and from at least one population forecasting model and forecasting future values of said individual time series using the individual time series forecasting model.

Claims

1. A computer program product encoding a computer program stored on a non-transitory computer readable storage medium for executing a process on a digital computer processor, the process comprising a set of instructions for forecasting commodity consumption by a target individual of a large population of a size of at least on the order of magnitude of one million individuals, the large population including individuals exhibiting irregular historical consumption patterns of a commodity, the target individual not necessarily having a commodity consumption pattern representative of the large population, wherein the set of instructions comprises instructions for:

a) selecting, from the large population of individuals having a size on the order of magnitude of one million, a representative sub-set of the population, the representative sub-set of the population, when combined with the target individual, not required to coincide with the large population;

b) providing historical consumption data describing actual historical consumptions of the commodity by the representative sub-set of the large population during one or more less-recent historical time period(s) and during one or more more-recent historical time periods(s);

c) for a plurality of commodity consumption forecast models, evaluating, for the specific case of the representative sub-set selected in step (b), performance of each forecast model of the plurality of forecast models by determining the ability of each forecast model of the plurality of commodity consumption forecast models to predict;

i) consumption of the commodity by the representative population sub-set during the more-recent historical time period(s) from

ii) data describing consumption of the commodity by the representative population sub-set during the less-recent historical time period(s)

d) according to the results of the forecast performance evaluating of step (c), selecting a sub-plurality of commodity consumption forecast models from the plurality of consumption forecast models;

e) for the target individual of the large population, providing historical consumption data describing actual historical consumptions of the commodity by the target individual during one or more less-recent historical time period(s) and during one or more more-recent historical time periods(s), the historical consumption data of the target individual not necessarily representative of the large population or of the representative sub-set of the large population;

f) for the selected sub-plurality of commodity consumption forecast models, evaluating, for the specific case of the target individual, performance of each forecast model of the sub-plurality of forecast models by determining the ability of each forecast model of the sub-plurality to predict;

i) consumption of the commodity by the target individual during the more-recent historical time period(s) from

ii) data describing consumption of the commodity by the target individual during the less recent historical time period(s);

g) formulating a combined forecasting model adapted for the target individual that includes at least some forecast models of the sub-plurality weighted for the target individual, in accordance with the results of the model performance-evaluating for the specific case of the target individual; and

h) forecasting, future consumption of the commodity by the target individual using the combined forecast model.

2. The computer program product of claim 1 wherein the consumed commodity is network bandwidth in a telecommunications network.

3. The computer program product of claim 1 wherein the consumed commodity is water provided by a metered water-delivery system.

4. The computer program product claim 1 wherein the consumed commodity is electricity provided by a metered electricity-delivery system.

5. The method computer program product claim 1 wherein at least steps (e)-(g) are repeated a plurality of times for a plurality of different target individuals, each being associated with a different respective historical consumption pattern of the commodity, thereby obtaining a plurality of different forecasts of future consumptions of the commodity.

6. The method computer program product of claim 1 wherein the performance evaluating of step (c) includes:

i) applying each commodity consumption model of the plurality of models to historical consumption data for the representative sub-set of the population during less-recent historical time period(s) in order to generate a test forecast of the consumption of the commodity by the representative sub-set during the more- recent historical time period(s); and

ii) comparing the test forecasts to the actual historical consumptions of the commodity by the representative sub-set of the large population during the more-recent historical time period(s).

7. The computer program product claim 1 wherein the selecting the sub-plurality of commodity consumption forecast models from the plurality of consumption forecast models is done using model redundancy between models of the plurality of forecast models.

8. A system for forecasting commodity consumption by a target individual of a large population of a size of at least on the order of magnitude of one million individuals, the large population including individuals exhibiting irregular historical consumption patterns of a commodity, the target individual not necessarily having a commodity consumption pattern representative of the large population, the system comprising:

a) a data storage including at least one of volatile and non-volatile memory configured to store;

i) a description of a representative sub-set of the population selected from the large population of individuals having a size at least on the order of magnitude of one million, the representative sub-set of the population, when combined with the target individual, not required to coincide with the large population;

ii) historical consumption data describing actual historical consumptions of the commodity by the representative sub-set of the large population during one or more less-recent historical time period(s) and during one or more more-recent historical time periods(s);

iii) a description of a plurality of commodity consumption forecast models; and

iv) historical consumption data describing actual historical consumptions of the commodity by the target individual during one or more less-recent historical time period(s) and during one or more more-recent historical time periods(s), the historical consumption data of the target individual not necessarily representative of the large population or of the representative sub-set of the large population;

b) a digital computer processor; and

c) computer code stored on a computer-readable medium, the digital computer processor and the computer code configured such that execution of the computer code by the digital computer processor effects the following steps;

i) selecting the representative sub-set of the population from the large population of individuals having a size at least on the order of magnitude of one million;

ii) for the plurality of commodity consumption forecast models, evaluating, for the specific case of the representative sub-set selected in step (i), performance of each forecast model of the plurality of forecast models by determining the ability of each forecast model of the plurality of commodity consumption forecast models to predict;

A) consumption of the commodity by the representative population sub-set during the more-recent historical time period(s) from

B) data describing consumption of the commodity by the representative population sub-set during the less-recent historical time period(s)

iii) according to the results of the forecast performance evaluating of step (ii), selecting a sub-plurality of commodity consumption forecast models from the plurality of consumption forecast models;

iv) for the selected sub-plurality of commodity consumption forecast models, evaluating, for the specific ease of the target individual, performance of each forecast model of the sub-plurality of forecast models by determining the ability of each forecast model of the sub-plurality to predict;

A) consumption of the commodity by the target individual during the more-recent historical time period(s) from

B) data describing consumption of the commodity by the target individual during the more recent historical time period(s); and

v) formulating a combined forecast model adapted for the target individual from that includes at least some forecast models of the sub-plurality weighted, for the target individual, in accordance with the results of the model performance-evaluating for the specific case of the target individual; and

vi) forecasting future consumption of the commodity by the target individual using the combined forecast model.

9. The system of claim 8 wherein the consumed commodity is network bandwidth in a telecommunications network.

10. The system of claim 8 wherein the consumed commodity is water provided by a metered water-delivery system.

11. The system of claim 8 wherein the consumed commodity is electricity provided by a metered electricity-delivery system.

12. The system of claim 8 wherein either the digital computer processor or the computer code stored on the computer-readable medium or the digital computer processor and the computer code are configured such that at least steps (iv)-(vi) are repeated a plurality of times for a plurality of different target individuals, each being associated with a different respective historical consumption pattern of the commodity, thereby obtaining a plurality of different forecasts of future consumptions of the commodity.

13. The system of claim 8 wherein either the digital computer processor or the computer code stored on the computer-readable medium or the digital computer processor and the computer code are configured such that the performance evaluating of step (ii) includes:

i) applying each commodity consumption model of the plurality of models to historical consumption data for the representative sub-set of the population during less-recent historical time period(s) in order to generate a test forecast of the consumption of the commodity by the representative sub-set during the more-recent historical time period(s); and

ii) comparing the test forecasts to the actual historical consumptions of the commodity by the representative sub-set of the large population during the more-recent historical time period(s).

14. The system of claim 8 wherein either the digital computer processor or the computer code stored on the computer-readable medium or the digital computer processor and the computer code are configured such that the selecting the sub-plurality of commodity consumption forecast models from the plurality of consumption forecast models is done using model redundancy between models of the plurality of forecast models.