Intermarket Analysis

• US 20110137781A1
• Filed: 12/07/2009
• Published: 06/09/2011
• Est. Priority Date: 12/07/2009
• Status: Active Grant

• Current Assignee: Predictive Technologies
• Original Assignee: Predictive Technologies
• Inventors: Mendelsohn, Louis B.
• Granted Patent: US 8,442,891 B2
• Time in Patent Office: Days
• US Class Current: 705/35
• CPC Class Codes: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes

Abstract

A method and a system for performing intermarket analysis. The method can include, from a pool of available markets in which at least one key intermarket has been selected and removed, selecting at least one market as a general intermarket, and removing the market selected as the general intermarket from the pool of available markets. From the pool of available markets from which the general intermarket has been removed, at least one market can be selected as a predictive intermarket and removed from the pool of available markets. Market data for each of the key intermarket, the general intermarket and the predictive intermarket can be processed to train a neural network. After training the neural network, market data for the primary market can be processed with the neural network to predict future market data for the primary market. The predicted future market data can be output.

Claims

1. A method of performing intermarket analysis, comprising:

from a pool of available markets in which at least one key intermarket has been selected and removed, selecting at least one market as a general intermarket, and removing the market selected as the general intermarket from the pool of available markets;

from the pool of available markets from which the general intermarket has been removed, selecting at least one market as a predictive intermarket and removing the market selected as the predictive intermarket from the pool of available markets;

processing market data for each of the key intermarket, the general intermarket and the predictive intermarket to train a neural network of a processing system so that when the neural network processes input data, predictive output data generated by the neural network is close to desired output data;

after training the neural network, processing market data for the primary market with the neural network to predict future market data for the primary market; and

outputting the predicted future market data.

2. The method of claim 1, wherein selecting the at least one market as a predictive intermarket comprises:

for each of the markets remaining in the pool of available markets after removal of the general intermarket;

processing historical market data for the market to generate predictive data for the primary market; and

comparing the predictive data to actual market data of a primary market and determining an improvement metric between the predictive data and the actual market data; and

selecting as the predictive intermarket the market that provides a best improvement metric.

3. The method of claim 2, wherein determining the improvement metric between the predictive data and the actual market data comprises:

measuring at least one statistic for each of the markets remaining in the pool of markets, the statistic selected from a group of performance metrics consisting of an accuracy, an average error and a mean squared error.

4. The method of claim 2, wherein determining the improvement metric between the predictive data and the actual market data comprises:

assigning a grade to each of the markets remaining in the pool of markets, each grade indicating at least one impact by the market in a trading simulation, the impact correlating to at least one measure of criteria selected from a group of performance metrics consisting of a measure of profitability, a measure of maximum drawdown, and number of trades.

5. The method of claim 1, wherein selecting at least one market as the general intermarket comprises:

for each of the markets remaining in the pool of available markets after removal of the key intermarket, determining a level of correlation between the market data and data of the primary market for each of a plurality of conditions; and

for at least one condition, selecting the market having a highest level of correlation with the primary market, and associating that market with the condition as a general intermarket.

6. The method of claim 1, further comprising:

from the pool of available markets, before the general intermarket has been selected and removed, selecting at least one of the markets as a key intermarket; and

removing the market selected as the key intermarket from the pool of available markets.

7. The method of claim 6, wherein selecting at least one market as the key intermarket comprises:

selecting a plurality of hierarchy levels for grouping the key intermarkets; and

for each hierarchy level;

selecting as a key intermarket for that hierarchy level a market from the pool of available markets that captures the most significant intermarket relationship for that hierarchy level; and

removing the market selected as the key intermarket for that hierarchy level from the pool of available markets.

8. A system that performs intermarket analysis, comprising:

a processor configured to;

from a pool of available markets in which at least one key intermarket has been selected and removed, select at least one market as a general intermarket, and remove the market selected as the general intermarket from the pool of available markets;

from the pool of available markets from which the general intermarket has been removed, select at least one market as a predictive intermarket and remove the market selected as the predictive intermarket from the pool of available markets;

process market data for each of the key intermarket, the general intermarket and the predictive intermarket to train a neural network so that when the neural network processes input data, predictive output data generated by the neural network is close to desired output data;

after training the neural network, process market data for the primary market with the neural network to predict future market data for the primary market; and

output the predicted future market data.

9. The system of claim 8, wherein to select the at least one market as a predictive intermarket, the processor is configured to:

for each of the markets remaining in the pool of available markets after removal of the general intermarket;

process historical market data for the market to generate predictive data for the primary market; and

compare the predictive data to actual market data of a primary market and determine an improvement metric between the predictive data and the actual market data; and

select as the predictive intermarket the market that provides a best improvement metric.

10. The system of claim 9, wherein to determine the improvement metric between the predictive data and the actual market data the processor is configured to:

measure at least one statistic for each of the markets remaining in the pool of markets, the statistic selected from a group of performance metrics consisting of an accuracy, an average error and a mean squared error.

11. The system of claim 9, wherein to determine the improvement metric between the predictive data and the actual market data the processor is configured to:

assign a grade to each of the markets remaining in the pool of markets, each grade indicating at least one impact by the market in a trading simulation, the impact correlating to at least one measure of criteria selected from a group of performance metrics consisting of a measure of profitability, a measure of maximum drawdown, and a number of trades.

12. The system of claim 8, wherein to select at least one market as the general intermarket the processor is configured to:

for each of the markets remaining in the pool of available markets after removal of the key intermarket, determine a level of correlation between the market data and data of the primary market for each of a plurality of conditions; and

for at least one condition, select the market having a highest level of correlation with the primary market, and associate that market with the condition as a general intermarket.

13. The system of claim 8, wherein the processor is further configured to:

from the pool of available markets, before the general intermarket has been selected and removed, select at least one of the markets as a key intermarket; and

remove the market selected as the key intermarket from the pool of available markets.

14. The system of claim 13, wherein to select at least one market as the key intermarket the processor is configured to:

select a plurality of hierarchy levels for grouping the key intermarkets; and

for each hierarchy level;

select as a key intermarket for that hierarchy level a market from the pool of available markets that captures the most significant intermarket relationship for that hierarchy level; and

remove the market selected as the key intermarket for that hierarchy level from the pool of available markets.

15. A computer program product comprising:

a computer-usable medium comprising computer-usable program code that performs intermarket analysis, the computer-usable medium comprising;

computer-usable program code that, from a pool of available markets in which at least one key intermarket has been selected and removed, selects at least one market as a general intermarket, and removes the market selected as the general intermarket from the pool of available markets;

computer-usable program code that, from the pool of available markets from which the general intermarket has been removed, selects at least one market as a predictive intermarket and removes the market selected as the predictive intermarket from the pool of available markets;

computer-usable program code that processes market data for each of the key intermarket, the general intermarket and the predictive intermarket to train a neural network of a processing system so that when the neural network processes input data, predictive output data generated by the neural network is close to desired output data;

computer-usable program code that, after training the neural network, processes market data for the primary market with the neural network to predict future market data for the primary market; and

computer-usable program code that outputs the predicted future market data.

16. The computer program product of claim 15, wherein the computer-usable program code further comprises:

computer-usable program code that for each of the markets remaining in the pool of available markets after removal of the general intermarket;

processes historical market data for the market to generates predictive data for the primary market; and

compares the predictive data to actual market data of a primary market and determine an improvement metric between the predictive data and the actual market data; and

selects as the predictive intermarket the market that provides a best improvement metric.

17. The computer program product of claim 16, wherein the computer-usable program code that determines the improvement metric between the predictive data and the actual market data comprises:

computer-usable program code that measures at least one statistic for each of the markets remaining in the pool of markets, the statistic selected from a group of performance metrics consisting of accuracy, an average error and a mean squared error.

18. The computer program product of claim 16, wherein the computer-usable program code that determines the improvement metric between the predictive data and the actual market data comprises:

computer-usable program code that assigns a grade to each of the markets remaining in the pool of markets, each grade indicating at least one impact by the market in a trading simulation, the impact correlating to at least one measure of criteria selected from a group of performance metrics consisting of a measure of profitability, a measure of maximum drawdown, and a number of trades.

19. The computer program product of claim 15, wherein the computer-usable program code that selects at least one market as the general intermarket the processor comprises:

computer-usable program code that, for each of the markets remaining in the pool of available markets after removal of the key intermarket, determines a level of correlation between the market data and data of the primary market for each of a plurality of conditions; and

computer-usable program code that, for at least one condition, selects the market having a highest level of correlation with the primary market, and associates that market with the condition as a general intermarket.

20. The computer program product of claim 15, wherein the computer-usable program code further comprises:

computer-usable program code that, from the pool of available markets, before the general intermarket has been selected and removed, selects at least one of the markets as a key intermarket; and

computer-usable program code that removes the market selected as the key intermarket from the pool of available markets.