System for extracting historical market information with condition and attributed windows

Description Submit all comments and votes

1. Field of the Invention

The present invention relates to automated information search and retrieval systems, and more particularly to systems designed to facilitate research into the fluctuations of commodity markets and the like over time, such as fluctuations in the stock market.

2. Background of the Invention

Many different database systems exist that can aid one in searching through the history of stock market fluctuations. These systems permit one, for example, to request the value of a particular stock on a particular date. They permit one to ask to have plotted the fluctuations of one or more stocks over a particular range of dates.

It is also well known to provide a general purpose database system that includes a query language in which searches can be expressed in the form of a logical expression that utilizes AND and OR logical operators to conjoin various search conditions, several of which may relate to greater than or less than relations between algebraic expressions concerning the data. Examples of such systems are DBase III, Inmagic, SOL and PARADOX. Once a complex query is defined using the query formulation language of one these systems, the system typically provides for the construction of windows that prompt the user for the values needed to facilitate the evaluation of query. But systems designed for stock market research typically do not use windows to prompt the users through the construction of an entire query, particularly any general type of sophisticated query.

Using conventional systems and assuming that the necessary data is present within the system's database, it is always possible for a sophisticated programmer to come in and to design a database structure, a set of queries formulated in a formal language, and a set of reports similarly formulated in a formal manner, that can extract almost any kind of desired report from the data. However, the process of designing a particular query can be time consuming and expensive, since it typically requires programmer assistance. It is not something that can be done quickly and easily by a market trader or an institutional investment manager who is not skilled in programming.

Accordingly, a primary object of the present invention is to design a market information machine that permits traders and portfolio managers to pose sophisticated ad hoc queries against an extensive database containing historical stock, commodity, and economic data at a level that was previously not economical nor timely, so that high-level research can be performed quickly in a way that can substantially increase trading profits.

SUMMARY OF THE INVENTION

The present invention, briefly described, is a computer system that uses templates displayed to a user in a window to prompt a user who is a trader or portfolio manager through the entire process of creating a sophisticated search request. The same windowing system may be used to review and modify the query. The query is also echoed back to the user in the form of a search request formulated in near plain English to aid the user in understanding the query. The user need not understand computer programming. The user may also escape from the windowing point-and-click-mode of a query formulation and revise the near plain English expression of the search query directly with a few keystrokes.

In addition, the market information machine enables the user to combine a request for the retrieval of fundamental data relating to the value of a commodity on given dates with search limitations relating to technical trading rules, holidays and historical events, business events, government reports, and even particular days of the week, month, or year, all in the same easy and simple manner. For example, one can easily request a bar chart of U.S. bond prices on all days when the consumer price index report was released by the U.S. Government, when bonds were up on the preceding day, and limited to those occurrences that fall within ten days before or after a prime rate increase. Such a search could be further limited to stipulate that only those occurrences between Memorial Day and Labor Day when the prime rate was over 10% should be displayed.

Invention resides in combining stock price information with economic indicator information, market average information, and other forms of market domain knowledge to form a single database (or several databases accessible by the same program) and then providing a English-like query language powerful enough to formulate time and event limited search requests such as the above example that lends itself to search formulation entirely through the use of overlaid window prompts or, alternatively, through the use of a more natural language format within the confines of a single system that can be used by traders and portfolio managers who .are not skilled in programming.

Further objects and advantages of the invention are apparent from the detailed description which follows, and the invention is defined with particularity in the claims annexed to and forming a part of this specification.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a overview block diagram of the system 100 which embodies the invention illustrating the way in which it interacts with the user when installed on a professional workstation;

FIG. 2 is a block diagram of a database 114 element of the system 100;

FIGS. 3A and 3B are two different embodiments illustrating, in block diagram form, the way in which user queries are represented as a data structure defining overlapping windows, as a near-English command string, and as a data structure suitable for controlling the operation of a search request execution engine;

FIG. 4A is a pre-construction window appearing on the initial power-up screen of the market information machine (MIM);

FIG. 4B is a pre-compilation window appearing on the initial power-up screen of the MIM;

FIG. 4C is a machine command window appearing on the initial power-up screen of the MIM;

FIG. 5A is a pre-construction window showing the selection of the January expiration date;

FIG. 5B is a pre-compilation window shown with a prompt for the user to complete the attribute wheel;

FIG. 5C is a machine command window (the "show attribute) conditions". A pre-format has been collected from the pre-construction window;

FIG. 6A is a pre-construction window showing the selection of the March triple-witching day;

FIG. 6B is a pre-compilation window showing a prompt to select from the attribute field;

FIG. 6C is a machine command window showing the selection of the attribute field from the pre-format statement;

FIG. 7A is a pre-construction window showing the selection of the closing price of the IBM one day ago plus one point as a selection for the attribute field;

FIG. 7B is a pre-compilation window showing a prompt for the attribute field;

FIG. 7C is a machine command window showing the selection of the attribute field from the pre-format;

FIG. 8A is a pre-construction window showing the selection of one point over the previous day for the attribute field;

FIG. 8B is a pre-compilation window showing a prompt for an attribute field entry;

FIG. 8C is a machine command window showing the selection of an attribute field;

FIG. 9A is the first of five bar charts of IBM generated for the query shown in FIGS. 8A, 8B and 8C; FIG. 9B is the second of five bar charts of IBM generated for the query shown in FIGS. 8A, 8B and 8C; FIG. 9C is the third of five bar charts of IBM generated for the query of FIGS. 8A, 8B and 8C; FIG. 9D is the fourth of five bar charts of IBM generated for the query of FIGS. 8A, 8B and 8C; FIG. 9B is the fifth of the five bar charts generated for the query shown in FIGS. 8A, 8B and 8C;

FIGS. 10 through 35 represent, in flowchart form, a formal syntax definition of a near-natural retrieval language for use in the preferred embodiment of the present invention;

FIG. 36 shows Menu Bars which represent "pull-down" windows;

FIG. 37 shows the XMIM mini-keypad;

FIG. 38 shows window scroll bars for use when all information in a window is not visible at the same time;

FIG. 39 shows sliders bars used in many places throughout the XMIM system for numerical input;

FIG. 40 illustrates two types of XMIM buttons: (1) single command Select Buttons and (2) Toggle Buttons;

FIG. 41 shows the interval function window in the XMIM system;

FIG. 42 is an interval function query example showing the graph of execution corresponding to the query for: "What happens when the US Financial Future hits 95 and then falls off to 94 sometime during the day?";

FIG. 43A is a conditional operator menu illustrating that the XMIM system utilizes concise connectors such as "crosses above" and "crosses below";

FIG. 43B is an example of a query utilizing concise connectors and moving averages showing the query: "the 18 day moving average of the SP's crosses above the 50 day moving average.";

FIG. 44 shows the Schema window;

FIG. 45 shows the Schema "SeriesSearch window" which allows the user to select an existing time series symbol for viewing the series' attributes and the associated data, modifying the series' attributes, or deleting a series;

FIG. 46 illustrates the "Series Add Window" for creating new security symbols;

FIG. 47 shows the Schema "Modify Facts window"; and

FIG. 48 shows a sample Macro window.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT

a. A Hardware and Software Systems

The preferred embodiment of the invention is shown in block diagram form in FIG. 1 of the drawings. The invention is implemented on a professional workstation 102 which, in the preferred embodiment, falls within the Sun Microsystem, SUN3 or SUNSPARC family of workstations, or it may be a Hewlett Packard HP9000 Series 400 or 700 workstation. The operating system used in the preferred embodiment is SUN-OS 4.0.3 or higher, or HPUX 7.0 or higher. In addition, the workstation is equipped with x/MOTIF, the industry standard windowing software designed to work with Unix under X Windows.

The invention focuses upon the interaction between a user 105 and a system of programs 100 installed on the workstation 102 which results in the user being able to develop database queries 107 using a windows interface program 104 and to view the queries in a natural English form on the screen (query feedback 109) while preparing and revising the queries using the windows interface program 104 for guidance.

The windows interface program 104 provides a vehicle whereby the user may develop a query on the screen of the workstation 102. The windows interface program 104 places in front of the user 105 templates which list the alternative query 105 formulations that the user can select using a mouse pointing device and a typewriter to prepare specific queries. The query windows, illustrated below, correspond to the language options that are set forth in a formal description of the query language that is also presented below, and thus guide the user through all of the possible query formulations that the language is capable of achieving. As the query is developed, it is represented in a tree structure composed of the various overlapping windows that are used by the user to create the query. The query language is basically an English-like language especially tailored to be embedded within a window's prompting system. But the queries are equally capable of being represented in what appears to be normal English. Accordingly, as the query is developed, a natural English version of the query is developed by the windows interface 104 and is passed to the display manager 106 for presentation as query feedback 109 to the user 105, so that the user can see a natural English language expression of the query or partial query that has been developed. In one embodiment, the user may enter or edit the English language query expression directly, rather than enter or edit the query using the windowing prompts. This facility enables all users to save time, and advanced users may bypass the windowing system on occasion.

The query itself is represented internally as a query representation 108 which is, in reality, a database structure that defines the overlapping set of windows which the user uses to define the query. When the user is finished defining the query and calls for the generation of reports based upon the query, the query representation 108 is compressed into a more compact, parsable form from which redundancies have been eliminated and is passed to an execution engine 110 which parses the query and calls upon a database manager 112 to extract the necessary information from a database 114; to search through the data performing the called for logical, mathematical, time range, and numeric computations; and to present the resulting information to a report manager 116 and also to a graphics manager 118 which respectively prepare reports 120 and graphs 122 for display to the user 105.

With reference to FIG. 12.2, the database 114, in the preferred embodiment of the invention, contains individual stock price information 202 concerning the price of stock on a daily (and optionally on an hourly) basis over the years. In addition, the database contains market averages information 204, such as the Dow Jones averages each day over extended periods. The database 114 also contains economic indicators information 206, such as the consumer price index, the GNP, automobile production figures, IBM's sales and profits (for example), and other such indicators, together with the dates upon which this information is released if appropriate. In addition, the database 114 contains market domain knowledge 208 relating to dates that can have an influence upon the market; such dates as holidays, special holidays, triple-witching days, and option contract expiration days, for example. Dividend distribution dates are also included. The user may also add special databases, such as bear or bull market dates, for example. Even the dates of the lunar cycles may be added to the system for use in the market studies.

The databases shown in FIG. 2 used in combination with the system 100 shown at FIG. 1 creates the potential for the user to generate reports that are of particular use to market traders and analysts in trying to examine past actions of the market at certain times, such as when dividends are paid or when special market events occur. In addition, the user may add the user's own market domain knowledge to the market domain knowledge 208 within the database 114 and thereby create a separate body of information for special reports. But to realize the potential of the database, a carefully human-engineered interface between the user 105 and the remaining parts of the system is provided in the form of the queries 107 which are implemented using overlapping windows and which are based upon an English-like search query language.

FIGS. 3A and 3B illustrate a way that user queries are processed in two different embodiments of the invention. In FIG. 3A, the user 105 interacts directly with the windows interface 104 to create and later to modify the query. During this interaction, information from the windows interface 104 enables an echo generator 302 to construct a natural language formulation of the query (or partial query) which is displayed on a screen 306 to the user but which is not directly editable by the user. When the query formulation process is completed, the windows data structure 304, created by the windows interface 104, is passed in a simplified form, stripped of redundancies, to an execution engine 110 which parses the data structure and then executes the database retrieval actions for and also causes the necessary reports to be generated, and formatted for printing and display. Note that report definition is inherent in the search query formulation process.

A second embodiment of the invention, illustrated in FIG. 3B, also permits the user 105 to interact with the user's windowface 104 to create a window data structure 306 that embodies the search request. The user 105 may then review and revise the query using the window interface just as in the embodiment of FIG. 3A. In addition, the echo generator 308 is designed to be bi-directional so that the window data structure 306 may be translated into an ASCII command string which is then presented to the user by a command string interface 310. The user may not only review the ASCII command string but may also edit the command string and revise it. Then the command string interface 310 reverses the process and essentially parses the command string back into a windows data structure 306 which may then be presented to the user 105 through the windows interface 104. The user is thus given the choice of entering and revising the query either by writing and revising an ASCII command string or by working through the series of overlaying window queries and prompts. Typically, a beginning user will stick with the windows queries and will only use the ASCII command string to make minor changes, as in the spelling of a particular stock name. Advanced users may then wish to save time by entering the query initially using the command string editing facility and possibly only use the windowing system to check out additional options that the user may have been forgotten. In either case, the final query data structure 306 is passed through a compiler 312 and is compiled into a compact tabular or tree form suitable for guiding the performance of an execution engine 110 through the tasks of retrieving the information, carrying out the relational steps called for by the search query, and generating and formatting the necessary reports.

b. The window System That Interacts With The User

FIGS. 4A, 4B and 4C represent the windows which appear on the power-up screen of the market information machine (MIM). FIG. 4A is the query construction window. The query construction window is always located in the lower half of the MIM screen. On power-up, this window presents the user with seven query format options 402-414 to be used in the construction of queries to the market information database. Each of the seven query formats provide a unique framework to be used in the construction of a query for searching the database. The query format also defines the sequencing of the later MIM screens and the order in which the query is constructed. This provides autosequencing which walks the user through the query construction, thereby preventing the construction of bad queries. FIG. 4A shows option 402 as being selected because it is illustrated in darker print than the latter six options 404-414. Option format 402 defines a "show attributes when conditions" query format. Option 402 is selected by placing the mouse pointer 418, shown in FIG. 4B on option format 402 with the mouse and clicking-on option 402.

Other formats are defined in option 404-414. Option format 404 is the "show count of days when conditions" query format. Option format 406 is the "show percent of days in universe when conditions" query format. Option format 408 is the "first values of attributes when conditions" query format. Option format 410 is the "show last values of attributes when conditions" query format. Option format 412 is the "show maximum values of single attributes when conditions" query format. And, option format 414 is the "show minimum values of single attributes when conditions" query format.

Once the user has chosen one of the query formats 402-414, then the user completes the selection by moving the mouse to the COMMIT 416 and clicking. When the user is committed to a particular option 402-414, then the query format is defined. The query format will then appear in the machine command window (FIG. 4C) and its construction will be compiled according to a predetermined autosequencing of window menu selections on the MIM screen display. This construction takes place in the query construction window (FIG. 4A) and the compilation is shown in the query compilation window (FIG. 4B) as the query construction progresses.

FIG. 4B is the query compilation window on power-up. The query compilation window is located in the upper left-hand corner of the MIM screen. FIG. 4B is shown with a blank window because no commands have been compiled on power-up. The MIM is idle waiting for commands on power-up. The mouse pointer 418 is used to click on selections with the mouse.

FIG. 4C illustrates the machine command window on power-up. The machine command window illustrated in FIG. 4C is located in the upper right-hand corner of the MIM screen display. The window is shown empty on power-up because a query format has not been chosen. The machine command window will show the query format 402-414 which is being operated on by the user. Eight soft key commands are located along the top of the machine command window. QUIT 420 is used to terminate the query the MIM program. RESTART 422 is used to re-initiate a query command process. EXEC 424 is used when a query is completely constructed. Clicking-on EXEC 424 will begin execution of the database search with the query shown in the query compilation window. PLOT 426 will plot the results of a database search. Examples of plots are shown in FIGS. 9A-9E. CONT 428 is used to continue the execution of a command on the MIM when it has been paused. SAVE 430 is used to save the results of a query on the MIM. PRINT 432 is used to print the results of query. FRONT 434 is used to position windows at the front of the MIM screen display.

FIG. 5A illustrates a query construction window showing attribute field options 502-520. DATE option 502 allows the user to select a data in field 522 in order to use a specific date as an attribute for determining when to display market information from a particular date. Field 522 is shown with Jun. 6, 1991. Day of the week option 504 uses the day field 524 to select a particular day of the week by clicking on field 524 with the mouse and then typing in the day of the week. Day field 524 is shown having Monday selected as the desired day of the week. Month option 506 is used with field 526 to select a particular month as an attribute. Here, January is the selected month. Year option 508 is used with field 528 to allow the user to type in a particular year. Illustrated here with 1991. HOLIDAYS option 510 is used with field 530 to select predetermined holidays by typing them into field 530. Major holidays such as Christmas and New Years are included. SPECIAL DAYS 512 is used with field 532 to allow the selection of attribute options corresponding to other important days of the year. Shown in field 532 is Martin Luther King's birthday. Other predetermined special days such as Lincoln's birthday, Washington's birthday or President's Day could also be used in option 512. Option 514 is used for menu selection. This is illustrated by a double-arrow icon 536. Icon 536 is shown having "Expiration in (month)" as being selected from the menu. Field 534 which is used to click on and type in the month of the option expiration date which the user is interested in. Menu selection double-arrow icon 536 is used to present other options to the user. When the user clicks on double-arrow icon 536 with the mouse, the user can select the option expiration date as shown in option 514. However, clicking-on the double-arrow icon 536 will change options to the triple-witching expiration date or the next dividend date for a particular stock. WITHIN OPTION 516 is used with unit field 538 and menu icon 540 to select an attribute of stock performance within a particular time period. Field 538 will select a number of time units of interest and icon 540 will change the units.

Icon 540 is used to select from a menu of time units for use in the WITHIN option 516. Other time units which will become available by a clicking on icon 540 are days, weeks, months, and years. FIG. 5A is shown with days selected by icon 540. FROM OPTION 518 is used with field option 544 and field option 542 to select an attribute for determining when to display market information by defining the attribute from a particular date 544 to a latter date 542. Option 518 is shown with the attribute from Dec. 28, 1959 to Jun. 6, 1991.

Options 502 through 518 are considered domain knowledge for the MIM, because they contemplate the use of predetermined dates for making database queries. Date, day, month and year information is preprogrammed into the query format structure of the MIM as are certain holidays, such as New Years and Christmas. Special days, such as Martin Luther King's birthday as illustrated in option 512, Lincoln's birthday, President's Day, and so on are preprogrammed. Special market dates are also preprogrammed into the MIM, these include ex-dividend dates for particular stocks, option expiration days in a particular month for a stock, and the triple-witching expiration dates. The domain knowledge offers flexibility in the standard construction of queries for the MIM database. This offers a great deal of flexibility for the user in making ad hoc queries about market conditions and testing the effectiveness of particular market rules for buying or selling stock or options. IN FILE option 520, however, allows the user added flexibility in that custom date attributes may be compiled by the user for the purpose of determining when to plot market information. These special dates, chosen by the user, are entered into a file and the file name is typed into field 546. FIG. 5A shows the date file typed into field 546. These dates may include dates corresponding to bull markets or bear markets, lunar cycles, sun spot cycles or any other dates which the user has an interest in using as an attribute for determining when to compile market information. So, IN FILE option 520 provides the user with added flexibility to make custom queries. Option 514 is shown darker than the other options because the user has clicked-on option 514. An interpretation of option 514, as illustrated in FIG. 5A, is constructed in 552 which reads: "January expiration date." Once the user has determined that the query construction shown in compilation 552 is as desired, the user can commit to the query construction by moving the mouse to COMMIT 548 and clicking on COMMIT 548. If the user wishes to cancel the query shown in 552, the user may do so by moving the mouse to CANCEL 550 and clicking on CANCEL 550. This enables the user to start over with attribute selection for the purpose of constructing a query.

FIG. 5B shows the query compilation window on the MIM screen. The attribute selected in FIG. 5A is prompted in FIG. 5B with the <attr> symbol 554. When the user clicks on COMMIT 548, the attribute symbol 554 will be replaced by the query construction 552 which is "January expiration day". So the query illustrated in FIG. 5B will read "show the movement over the previous day of the closing price of IBM repeated for the current and next five days when the date is January expiration day." This then represents an appropriate query compilation. With the query compilation finished in FIG. 5B, the user can proceed to the machine command window, FIG. 5C, to instruct the MIM as to what to do with the query command. Clicking-on EXECUTE 424 will then cause the MIM to execute the query from the query compilation window and display the appropriate values. Clicking on PLOT 426 will command the MIM to plot appropriate marketing information as illustrated in FIGS. 9A-9E. The machine command window, shown in FIG. 5C, also shows the basic structural format of the query illustrated to show the attributes which condition the market information search. Fields 556 through. 568 are used to click on with the mouse in order to flesh out the query format and construct a query for compilation. Attribute field 556 is used as an attribute of the desired market information, for example, what stock is desired or what market index is desired. REPEAT 558 will command the MIM to repeat attributes 556. ATTRIBUTE 560 which is a field for determining when market information defines an attribute 560 should be displayed. OP 562 represents an operating field for operating on attributes under the WHEN condition. Attribute 564 is shown enlarged relative to the other fields because it has been clicked on by the mouse and the field is being operated on in the query construction window as illustrated in FIG. 5A where the January expiration date has been selected. Repeat 566 can be used to repeat WHEN conditions in the query format. CUM-ATTR 568 is used for cumulative attribute conditions for the purposes of determining what market information to be queried from the database.

FIG. 6A shows a query construction window which is similar to that described in FIG. 5A except that icon 536 has been clicked on with the mouse and has switched from the "expiration date" to the "triple-witching date." The March date has been selected in field 534. So in FIG. 6A, the 514 option now represents the "March triple-witching day" as shown in the construction interpretation 552. Clicking-on COMMIT 548 will replace attribute 554 in FIG. 6B with "March triple-witching date". This completes the query compilation in FIG. 6B.

FIG. 6C, illustrating the machine command window with a selected query format, again attribute 564 is shown enlarged because it was clicked on and selected by the mouse. This has been determined in the query construction window to be the March triple-witching date 522.

FIG. 7A illustrates a query construction window for determining attributes which determine when to display market information based on stock prices, market indices, such as composite averages for volumes, or other market performance criteria. Command line attribute 702 has been clicked-on by the mouse and is illustrated in darker print than options 704-710. Security field 712 has been filled in with IBM. So the query construction illustrated by 522 is the closing pricing of IBM one day ago plus one point. If this is what the user intended, the user may click-on COMMIT 548 and the construction 522 will be used as the market information attribute 700 as illustrated in the query compilation window of FIG. 7B.

Option 704 enables the user to use volume as an attribute. Field 714 is shown with the Dow Jones Industrial Average (DJIA) Market Index as being selected. Option 706 is used when selecting actual values of stock or commodity prices as an attribute for searching marketing information. The open market price has been selected in field 716 and commodity selected in field 718 is the Deutsch mark (DM). Field 720 illustrates a default option of the same day and Field 722 illustrates a default option of no offset in comparison of the actual value attribute 706.

Option 708 enables the user to switch from actual values to changing values. Icon 724 causes the user to switch between the previous or the next number of days being defined in field 726 which is shown with one time unit. The units can be switched with icon 728. Icon 728 allows selection of days, weeks, months or years as a time unit. The market change information can be made relative to an economic indicator with field 730. Field 730 is shown with the consumer price index (CPI). The number of days which the user is interested in retrieving is entered in field 732. The time period can be switched with icon 734 and icon 736. Icon 738 will provide offsets. Clicking-on icon 738 will provide either positive, negative or no offset. Icon 783 is shown as having a pos