System and methods for content-based financial database indexing, searching, analysis, and processing

• US 8,458,065 B1
• Filed: 01/31/2008
• Issued: 06/04/2013
• Est. Priority Date: 01/31/2007
• Status: Active Grant

• Current Assignee: Financialsharp, Inc.
• Original Assignee: Financialsharp, Inc.
• Inventors: Zhang, Xiaoping, Kedmey, David, Wang, Fang
• Primary Examiner(s): MILEF, ELDA G
• Application Number:
  US12/023,908
• Time in Patent Office: 1,951 Days
• Field of Search: 705/35
• US Class Current: 705/35
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06Q 10/0639 Performance analysis of employees; Performance analysis of enterprise or organisation operations
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange
• CPC Class Codes: G06Q 40/06 Asset management; Financial planning or analysis

Abstract

Robust content-based financial data processing is enabled by software with a customizable knowledge base. The software indexes a publicly or privately available financial database based on the content of the database records. Utilizing proprietary information contained within a knowledge base, the software enables users to search the indexed database by feature, example firm, or pattern. The information contained in the knowledge base enables results to be ranked by relevance. Moreover, users provide feedback about search results to enhance the knowledge base. Software analyzes data by using adaptive signal processing to remove market trends or noise and enable complicated market research.

Claims

1. A method for performing content-based indexing of a financial database, said method comprising:

causing at least one memory device to store data representative of a plurality of potential financial object features, each feature representing a complex set of characteristics of the financial object useful to a financial professional;

causing the at least one memory device to store a user preference including at least one of a past decision about a financial object and a preferred feature of a financial object;

causing at least one processor to retrieve a database record containing time series data from the financial database;

causing the at least one processor to apply a mathematical transform function to the time series data contained in the database record to generate transformed data;

causing the at least one processor to operate with the at least one memory device to create an indexed database wherein any of the plurality of potential financial object features is usable as a query to search for such feature in different time series data sets by, for each of the plurality of potential financial object features;

(i) causing the at least one processor to determine whether said potential financial object feature exists in at least one of the time series data and the transformed data, and

(ii) if the determination is that said potential financial object feature exists in at least one of the time series data and the transformed data, causing the at least one memory device to store, in said indexed database, an indication of an association between that potential financial object feature and said data in which the feature exists; and

causing the at least one processor to operate with at least one display device to display a subset of at least one of the time series data and the transformed data based on the user preference, the subset representing at least one stored feature that was found to exist in at least one of the time series data and the transformed data.

2. The method of claim 1, wherein the plurality of potential financial object features includes at least one selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an earnings-per-share indicator, a yield indicator, an amount of free cash flow, a coverage ratio, a foreign exchange rate, a plurality of accounting line items, at least one credit rating, at least one characteristic of at least one option, and at least one valuation metric.

3. The method of claim 1, wherein at least one of the plurality of potential financial object features is at least one selected from the group consisting of:

time series data that has undergone a Fourier transform, time series data that has undergone a singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

4. The method of claim 1, wherein at least one of the plurality of potential financial object features is at least one selected from the group consisting of:

common stock, preferred stock, exchange traded funds, mutual funds, closed-end funds, real estate investment trusts, master limited partnerships, royalty trusts, enhanced income securities, private equities in US markets, public equities in US markets, private equities in non-US markets, and public equities in non-US markets;

options on equities, futures, currencies, and other derivatives contracts, OTC, and exchange traded in US and non-US markets;

fixed income securities;

futures and forwards contracts such as commodities futures, interest rate futures, equity index futures, and single stock futures;

quantified credit ratings;

asset pricing model parameters;

GAAP accounting line items;

non-GAAP accounting measures;

valuation metrics;

projected and historical pro forma financial statement items;

items in accounting disclosures;

information disclosed in SEC Form-10K, Form 10-Q, Form S-1, proxy statements, and other public and private company disclosures;

profitability measures;

capital structure;

operational variables;

debt covenants;

technical analysis measures;

macro economic variables;

audit report items;

analyst estimates;

research report items; and

Commitment of Traders Report data.

5. The method of claim 1, wherein at least one of the plurality of potential financial object features is determined by comparing two or more selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an amount of earnings-per-share, a yield amount, a free cash flow amount, a coverage ratio, a foreign exchange rate, accounting line items, credit ratings, characteristics of options, valuation metrics, time series data that has undergone a Fourier transform, time series data that has undergone an singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

6. The method of claim 1, wherein at least one of the plurality of potential financial object features is derived from at least one selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an amount of earnings-per-share, a yield amount, a free cash flow amount, a coverage ratio, a foreign exchange rate, accounting line items, credit ratings, characteristics of options, valuation metrics, time series data that has undergone a Fourier transform, time series data that has undergone an singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

7. A method for performing content-based indexing of a financial database, said method comprising:

causing at least one memory device to store a plurality of keywords, each keyword indicating the presence of a financial object feature;

causing the at least one memory device to store a user preference, said user preference including at least one of;

a user past decision about a financial object and a user preferred feature of a financial object;

causing at least one processor to retrieve textual data about a financial object from the financial database;

causing the at least one processor to operate with the at least one memory device to create an indexed database wherein any of the plurality of keywords and quantifying modifiers are usable as a query to enable location of the feature represented thereby in different sets of textual data by, for each of the plurality of keywords;

(i) causing the at least one processor to parse the textual data for the keyword and a quantifying modifier based on the user preference, wherein the presence of the keyword and the quantifying modifier indicate that a feature associated with the keyword and the quantifying modifier is present in the textual data; and

(ii) if the keyword and the quantifying modifier are present in the textual data, causing the at least one memory device to store, in said indexed database, an indication of the association between the feature associated with that keyword in association with said textual data, the quantifying modifier representing a non-numerical modifier to the keyword indicating an assessment of the keyword in the context of the textual data.

8. The method of claim 7, wherein the plurality of keywords includes at least one selected from the group consisting of:

a stock ticker symbol, a company name, an analyst opinion, a potential merger candidate, a board member name, institutional ownership, an insider name, a major holder name, a term indicating merger activity, a term indicating acquisition activity, a competitor firm, a firm identified by a user, and a keyword to describe a feature of a time series data set.

9. The method of claim 7, wherein at least one of the plurality of keywords is at least one selected from the group consisting of:

gross domestic product, GDP, earnings, Chief Executive Officer, and CEO.

10. The method of claim 7, wherein the quantifying modifier is at least one selected from the group consisting of:

good, neutral, bad, increase, steady, and decrease.

11. The method of claim 7, which includes causing the at least one memory device to store the keyword, the quantifying modifier, and a numeral included in the textual data as a feature of the textual data.

12. A method of searching a financial database, said method comprising:

causing at least one memory device to store a user preference, said user preference including at least one of;

a user past decision about a financial object and a user preferred feature of a financial object;

causing the at least one memory device to store a similarity measure, the similarity measure indicating a characteristic by which to determine similarity of a first financial object and a second financial object;

enabling a user to specify a query item using at least one input device, wherein the query item is a representation of an example financial object;

causing at least one processor to determine a feature of the query item, the feature representing a complex set of characteristics of the example financial object useful to a financial professional;

causing the at least one processor to compare the feature of the query item with a feature of an indexed financial object stored in an indexed database, the indexed database having been created by searching for each of a plurality of potential features in each a plurality of data sets, and for each feature found in each data set, said found feature having been stored in said indexed database in association with said data set to be usable as a query to search for said data set in said indexed database;

causing the at least one processor to generate a first similarity score for the indexed financial object based on at least one of the similarity measure and the user preference;

causing at least one display device to display the indexed financial object and the first similarity score;

enabling the user, using the at least one input device, to provide feedback about the first similarity score;

causing the at least one processor to compare the feature of the query item with the feature of the indexed financial object; and

causing the at least one processor to generate a second similarity score for the indexed financial object based on;

(1) the user feedback and (2) at least one of;

the similarity measure and the user preference.

13. The method of claim 12, wherein the user past decision about the financial object includes an investment decision based on a similarity score.

14. The method of claim 12, wherein the query item includes at least one item selected from a group consisting of:

a feature, a computer-generated graph, a hand-drawn graph, an image, a description of a feature, an example firm, an example financial instrument, an example financial index, and an example financial object.

15. The method of claim 12, wherein enabling the user to provide feedback about the first similarity score using the at least one interface device includes at least one selected from the group consisting of:

enabling the user to input a check in a check box indicating the financial object stored in the indexed database is similar, enabling the user to input a feedback score indicating the extent to which the financial object stored in the indexed database is similar, and enabling the user to input information indicating an important feature about the financial object stored in the indexed database.

16. The method of claim 15, wherein enabling the user to provide feedback about the first similarity score using the at least one interface device includes applying a learning algorithm to combine a user input provided using the at least one interface device with a user'"'"'s pattern of selecting results.

17. The method of claim 12, which includes generating a first similarity score for the financial object stored in the indexed database based on at least one of a pre-set similarity measure, the user decision, and a decision of a second user.

18. The method of claim 12, wherein enabling the user to provide feedback about the first similarity score using the at least one interface device includes at least one selected from the group consisting of:

implicitly monitoring a first user pattern of selecting results and implicitly monitoring a second user pattern of selecting results.

19. A method for analyzing data in an indexed financial database, said method comprising:

causing at least one memory device to store a user preference, said user preference including at least one of;

a user past decision about a financial object and a user preferred feature of a financial object;

enabling a user to define a feature which is of interest to the user using at least one input device, the feature representing a complex set of characteristics of the financial object useful to the user;

causing at least one processor to apply a mathematical transform to a feature of a financial object stored in said indexed financial database of financial objects, the indexed database having been created by searching for each of a plurality of potential features in each a plurality of data sets, and for each feature found in each data set, said found feature having been stored in said indexed database in association with said data set to be usable as a query to search for said data set in said indexed database, said application of the mathematical transform generating a transformed financial object feature;

causing the at least one processor to compare the feature of interest with the transformed financial object feature;

causing the at least one processor to generate a similarity score based on at least one of the user preference and a predetermined similarity score associated with the user, said similarity score indicating a similarity between the feature of interest and the transformed financial object feature; and

causing at least one display device to display the transformed financial object feature and the similarity score.

20. The method of claim 19, wherein the plurality of potential features includes at least one selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an earnings-per-share indicator, a yield indicator, an amount of free cash flow, a coverage ratio, a foreign exchange rate, a plurality of accounting line items, at least one credit rating, at least one characteristic of at least one option, and at least one valuation metric.

21. The method of claim 19, wherein at least one of the plurality of potential features is at least one selected from the group consisting of:

time series data that has undergone a Fourier transform, time series data that has undergone a singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

22. The method of claim 19, wherein at least one of the plurality of potential features is at least one selected from the group consisting of:

common stock, preferred stock, exchange traded funds, mutual funds, closed-end funds, real estate investment trusts, master limited partnerships, royalty trusts, enhanced income securities, private equities in US markets, public equities in US markets, private equities in non-US markets, and public equities in non-US markets;

options on equities, futures, currencies, and other derivatives contracts, OTC, and exchange traded in US and non-US markets;

fixed income securities;

futures and forwards contracts such as commodities futures, interest rate futures, equity index futures, and single stock futures;

quantified credit ratings;

asset pricing model parameters;

GAAP accounting line items;

non-GAAP accounting measures;

valuation metrics;

projected and historical pro form a financial statement items;

items in accounting disclosures;

information disclosed in SEC Form-10K, Form 10-Q, Form S-1, proxy statements, and other public and private company disclosures;

profitability measures;

capital structure;

operational variables;

debt covenants;

technical analysis measures;

macro economic variables;

audit report items;

analyst estimates;

research report items; and

Commitment of Traders Report data.

23. The method of claim 19, wherein at least one of the plurality of potential features is determined by comparing two or more selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an amount of earnings-per-share, a yield amount, a free cash flow amount, a coverage ratio, a foreign exchange rate, accounting line items, credit ratings, characteristics of options, valuation metrics, time series data that has undergone a Fourier transform, time series data that has undergone a singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

24. The method of claim 19, wherein at least one of the plurality of potential object features is derived from at least one selected from the group consisting of:

a price-to-earnings ratio, a stock price, a stock volume, a bid-ask spread, an amount of earnings-per-share, a yield amount, a free cash flow amount, a coverage ratio, a foreign exchange rate, accounting line items, credit ratings, characteristics of options, valuation metrics, time series data that has undergone a Fourier transform, time series data that has undergone an singular value decomposition transform, time series data that has undergone a wavelet transform, and time series data that has undergone a time-frequency transform.

25. The method of claim 19, which includes enabling a user to provide feedback about the similarity score and the transformed financial object feature using the at least one input device.

26. The method of claim 25 which includes causing the at least one processor to compare the feature of interest with the transformed financial object feature and to generate an updated similarity score based on the feedback about the similarity score and the transformed financial object feature.

27. The method of claim 25, wherein the user provides feedback about the similarity score and the transformed financial object feature using the at least one input device by a mechanism selected from the group consisting of:

an interface to enable a user to change the similarity score, an interface to enable a user to indicate the importance of a feature, and an algorithm to infer user feedback by monitoring user decisions based on the displayed financial object feature.

28. The method of claim 19, which includes causing the at least one processor to generate a similarity score based on at least one selected from the group consisting of:

the user preference, a user preference associated with a second user, and a predetermined similarity score associated with the user, the predetermined similarity score indicating a similarity between the feature of interest and the transformed financial object feature.

29. The method of claim 25, which includes causing the at least one processor to generate a second similarity score indicating a similarity between the feature of interest and the transformed financial object feature based on user feedback about the similarity score.

30. A method of processing an indexed financial database, said method comprising:

causing at least one memory device to store a user preference including at least one of a past decision about a financial object and a preferred feature of a financial object;

causing at least one processor to determine a processing algorithm based on at least one of the user preference and a user input;

causing the at least one processor to apply the processing algorithm to a feature contained in the indexed financial database to result in a processed feature, the indexed financial database having been created by searching for each of a plurality of potential features in each a plurality of data sets, and for each feature found in each data set, said found feature having been stored in said indexed financial database in association with said data set to be usable as a query to search for said data set in said indexed financial database;

causing at least one display device to display the processed feature to a user;

enabling the user to provide feedback, via at least one input device, about the processed feature;

causing the at least one memory device to store the user provided feedback;

causing the at least one processor to modify the processing algorithm based on the user provided feedback; and

causing the at least one processor to apply the modified processing algorithm to the feature contained in the indexed financial database.

31. The method of claim 30, wherein the processing algorithm is at least one selected from the group consisting of:

a noise removal algorithm, a customized trend removal algorithm, a customized signal filter, and a customized modeling algorithm.

32. The method of claim 30, wherein the user provided feedback about the processed feature is provided by a mechanism selected from the group consisting of:

an interface to enable a user to describe user preferences using a feature, an interface to enable a user to describe user preferences using at least one predefined processing parameter with the processing algorithm, an algorithm to infer user feedback by monitoring user decisions based on the displayed processed feature, and an algorithm to combine user provided feedback with inferred user feedback.