JP2008129984A - Transaction support system, transaction support program and recording medium thereof - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a transaction support system for providing a user with the best solution by reproducing the latest market model with the latest data attached thereto, using the model and searching the best solution by evolutionary computation, a transaction support program and its recording medium. <P>SOLUTION: The transaction support system comprises: an information collecting part for collecting information composed of a plurality of items on an individual object; an information recording part for storing the collected information; a data providing part for performing missing value processing of the collected information; a market reproducing part for reproducing the past market by using the collected information; a strategy creating part for creating a strategy on the basis of a predetermined index for target information at a time designated by a user; a fitness calculating part for calculating fitness at the past reproduction market by using the created strategy; and a providing part for providing a user with the calculated fitness. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a transaction support method and a transaction support program when trading financial products such as securities, commodity futures, and foreign exchange.

  With the rapid spread of communication networks such as the Internet, home trade where customers (investors) use the Internet or the like to perform commodity futures transactions or securities transactions has been realized. Various attempts have been made to assist those who conduct such transactions.

  In particular, as a method of predicting future market conditions, a virtual market is built in a computer, and the economic phenomena are analyzed by performing the activities of virtual investors and actual investors in the virtual market. Analysis is attempted (see, for example, Non-Patent Documents 1 and 2).

  In the method (U-Mart) described in Non-Patent Document 1, a stock approach using a stock index future that trades a stock price index on a fixed date in the future at a price decided at the present time is used. Therefore, the method described in Non-Patent Document 1 is not effective for analyzing a detrade strategy. Further, in the method described in Non-Patent Document 2 (Kababo contest), board information is not taken into consideration at all. The board information is the most important judgment material for de-trading. Therefore, even the method described in Non-Patent Document 2 cannot acquire daytime data effective for analyzing a detrade strategy.

  The limit of the sell order and the limit of the buy order for which trading has not been established are referred to as “indication price”. The board indicates the order status (indication price) of selling and buying in real time. The bid price is the desired price of the transaction indicated when trading stocks and bonds mainly on exchanges. The interval between quotes is called the quote unit. In stock trading, a quote unit is set according to the stock price, and a price smaller than this unit cannot be specified.

  The stock price is represented by a single number. In actual trading of stocks, it is necessary to consider the price that can be bought cheapest (best bid price) and the price that can be sold highest (best bid price). The best bid price will never be smaller than the best bid price. Immediately after buying a stock at the best bid price, the price that will definitely sell is the best bid price. For this reason, the purchaser bears the risk of the difference between the best bid price and the best bid price from the moment the stock is purchased. FIG. 1 is a diagram illustrating the relationship between the best bid price and the best bid price. In the example shown in FIG. 1, if you want to buy a stock immediately, you can only buy it for 2030 yen. If you want to sell your stock immediately, you can only sell it for 2025 yen. This problem does not become apparent in the case of medium- to long-term investment, but it is an essential problem in the case of detrade.

Evolutionary computation is known as a method for solving optimization problems by modifying, synthesizing, and selecting data structures by mimicking the mechanism of biological evolution. The market changes over time. Such changes in time series are said to be dynamic environment problems. Evolutionary computation is more difficult to apply to dynamic problems than static problems. Therefore, it is desirable to search for an optimal solution by successfully applying evolutionary computation to a dynamic environment.
Kita Hajime, “Introduction to U-Mart”, [online], May 24, 2000, U-Mart Project, [Search on November 25, 2006], <URL; http: //u-mart.econ. kyoto-u.ac.jp/articles/u-mart-introduction.pdf> “Super Cabrobo Tournament Site”, August 25, 2006, [Search November 25, 2006], <URL;http://www.kaburobo.jp/docs/super-kaburobo_rule_rev1-04_0825_2006.pdf>

  That is, the present invention has been made in view of the above problems, and its purpose is to reproduce the latest market model with the latest data added and to search for the best solution by evolutionary calculation using this model. Another object is to provide a transaction support system, a transaction support program, and a recording medium provided to the user.

  The transaction support system of the present invention includes an information collecting unit that collects information including a plurality of items related to individual objects, an information recording unit that accumulates the collected information, and a data providing unit that performs missing value processing on the collected information A market reproduction unit that reproduces a past market using the collected information, a strategy creation unit that creates a strategy based on a predetermined index for target information at a point specified by a user, and the created strategy The fitness calculation unit that calculates the fitness in the past reproduction market, and the provision unit that provides the calculated fitness to the user.

  In the present invention, the past market is reproduced from the information that is collected and accumulated and includes a plurality of items relating to individual objects. A strategy is created based on a predetermined market for target information at a user's desired time. Next, the fitness of this strategy is calculated in the past reproduction market. Provide users with a highly adaptable strategy. The user can perform a more advantageous transaction by performing a transaction in the actual market of the next day using the obtained strategy.

  The predetermined index may include board information. Thereby, a strategy suitable for detrade can be obtained.

  The transaction support method of the present invention includes an information collecting step for collecting information consisting of a plurality of items related to individual objects, an information recording step for accumulating the collected information, and data for performing missing value processing on the collected information. A providing step, a market reproducing step for reproducing a past market using the collected information, a strategy creating step for creating a strategy based on a predetermined index for target information at a point specified by a user, and the creation The fitness level calculation step for calculating the fitness level in the past reproduction market using the strategy and the providing step for providing the calculated fitness level to the user. Further, when creating a strategy, a predetermined index including board information may be used.

  Further, the transaction support program of the present invention includes an information collecting step for collecting information consisting of a plurality of items related to individual objects, an information recording step for storing the collected information, and data for performing missing value processing on the collected information. A providing step, a market reproducing step for reproducing a past market using the collected information, a strategy creating step for creating a strategy based on a predetermined index for target information at a point specified by a user, and the creation The fitness level calculation step for calculating the fitness level in the past reproduction market using the strategy and the providing step for providing the calculated fitness level to the user. Further, when creating a strategy, a predetermined index including board information may be used.

  The transaction support program may be a computer-readable recording medium.

  The present invention reproduces the latest market model to which the latest data is added, searches for the best solution by evolutionary calculation using this model, and provides it to the user. Therefore, the user can provide a transaction support system, a transaction support program, and a recording medium thereof that can perform more advantageous transactions by performing transactions in the actual market the next day using the obtained strategy. it can.

  Hereinafter, the present invention will be described in detail with reference to the drawings.

[Constitution]
FIG. 2 is a diagram showing an outline of the transaction support system according to the present invention. The system includes an information collecting unit 1, an information recording unit 2 for accumulating the collected information, a data providing unit 3 for performing missing value processing on the collected information, and a past market using the collected information. The market reproduction unit 4 to be reproduced, the strategy creation unit 5 that creates a strategy based on a predetermined index for the target information specified by the user, and the fitness in the past reproduction market using the created strategy Is provided with at least an adaptability calculation unit 6 for calculating the value and a providing unit 7 for providing the calculated fitness value to the user. Although not shown, an automatic transaction apparatus may be included.

  The system shown in FIG. 1 is realized by a computer system such as a workstation including a central processing unit, an input / output interface device, and an external storage device. Also, one or a plurality of programs for realizing the present system is stored in a storage medium, the program is read by a storage medium reading device or the like, and the read program is executed by a computer system including a CPU and a memory. Thus, it may be configured to realize each device and information recording unit of the system shown in FIG.

  In this system, information such as stock prices, interest rates, interest rates and prices of bonds, exchange rates, futures products, etc. as information on the actual market at regular intervals (for example, every few seconds to several tens of seconds) using the information collecting unit 1 For example, the data is automatically extracted and collected using a computer network such as the Internet, LAN, and WAN. For example, stock price, stock name, current price, current price, current price time, volume, volume-weighted average, best bid price, best bid price, best buy, etc. Collect quotes, best bids, step information, etc. Information collected in this way is sent to and stored in the information recording unit 2. The information storage unit 2 stores stock price information obtained from the network in a CSV (Comma Separated Value) file, a multi-user multi-thread SQL database, or the like. Information is collected during the day when stocks are traded, for example.

  The data providing unit 3 extracts the collected information from the information recording unit 2 and performs missing value processing when the collected information has a missing value such as deletion of error data. The extracted information is used to reproduce the past market on the computer by the market reproduction unit 4. Reproduction of the market can reproduce the market trend of the day finely by the board information every few seconds.

  The strategy creation unit 5 creates a strategy based on a predetermined index for target information at the time point designated by the user. This index is recorded as a module in the information recording unit 2. Next, the fitness calculation unit 6 calculates the fitness in the past market using this strategy. The obtained fitness is provided to the user from the providing unit 7, and the user can perform a transaction in an actual market based on the fitness. Further, since the strategy is verified in a market in which past information is accurately reproduced in a state where information is not collected (offline), the strategy can be evolved or learned.

[Transaction support system and transaction support method]
In the following explanation, explanation will be made based on the case where stock trading is performed. In the present invention, in order to use the evolutionary calculation more effectively, an actual dynamic problem is modeled on a computer, and the evolutionary calculation is applied to the modeled problem. Specifically, during the day when the stock price fluctuates, stock price information is collected in real time, and the latest market model with new data added is constructed after the market ends, and the best solution is searched.

[Create strategy]
In the present invention, an evolutionary calculation method is used as a trading strategy. In the following example, description will be made using an example of a genetic algorithm (GA) and an example of a genetic program (GP). The evolutionary calculation method used in the present invention is not limited to these, and a known method can be used.

(Genetic algorithm (GA))
In the solid expression in GA, an index is given to each gene, and {0, 1} is assigned to the given index. Specifically, as shown in FIG. 3, the gene code with a gene length of 14 is used as one trading strategy, and the seven genes in the first half of the gene code are purchased and the seven genes in the second half are sold. Allocate and trade according to the conditions. As an index to be assigned to a gene, an index that can be assigned {0, 1} can be used. In addition, the number of genes can be arbitrarily determined.

  An example of an index assigned to a gene is shown below. Indexes relating to past stock prices, turnovers, and plate thicknesses are given, and {0, 1} is determined based on the data before t, 2t, 3t, 4t seconds and the current data using the reference interval t. The reference interval t is a parameter given from the outside.

1. The stock at the time that the index search about stock and Pc, to the time of searching t, 2t, 3t, 4t seconds before the stock each _{P -t, P -2t, P -3t} , and _{P of -4t} . The stock price indicators are as follows.
(1) The magnitude of Pc and P- _t ((1) in FIG. 3): 1 if Pc> P- _t , 0 otherwise
(2) _Magnitude of Pc and P- _2t ((2) in FIG. 3): 1 if Pc> P- _2t , 0 otherwise
(3) The magnitude of Pc and P _-3t ((3) in FIG. 3): 1 if Pc> P _-3t , 0 otherwise
(4) _Magnitude of Pc and P- _4t ((4) in FIG. 3): 1 if Pc> P- _4t , 0 otherwise

2. The volume _{V 0} which from the index (5) 2t relating volume to t seconds before, volume _{V 1} of the magnitude of up to the time of searching the t seconds before ((5 in FIG. _3)): if V 0 _{<V 1} 1; otherwise 0
(6) Volume V ₂ from 4 t to 2 t seconds ago and volume V ₃ from 2 t seconds to the time of searching ((6) in FIG. 3): 1 if V ₂ <V ₃ Other than 0

3. As shown in FIG. 4, the number of orders corresponding to five buying plates from the best bid and five selling plates from the best bid are obtained. The total value of the number of each of the five orders ((7) in FIG. 3): 1 if the total value of the best bid is greater than the total value of the best bid, 0 otherwise

  The values obtained as described above are assigned to (1) to (7) in FIG. Similarly, regarding the buying conditions, the obtained values are assigned to (1) to (7) in FIG.

  Using the strategy obtained in this way, a transaction is performed with a market reproduction model. The profit of the virtual trading agent that conducts trading is regarded as fitness. For example, one day profit is obtained by subtracting the best bid price from the best bid price. The reproduction period of the market reproduction model is not particularly limited, and a period and a date for which trading is desired can be arbitrarily selected. The fitness is the sum of the profits of the period or the profits on a predetermined date.

  When conducting the simulation, certain trading conditions are set. For example, a one-day simulation is performed under the following conditions. The agent starts from the buying state and buys at the best bid price if the buying conditions are met. After buying, if the selling conditions are met, sell at the best bid price. Each agent buys and sells only once a day. If the buying condition is met but the selling condition is not met and the stock is not bought or sold, it is sold at the closing price. Furthermore, the profit is set to 0 when the buying conditions are not matched.

(Genetic Program (GP))
In the solid representation in GP, a strategy is represented by a tree structure. The condition based on the technical index is used for the non-terminal symbol, and the agent action (trading) is used for the terminal symbol. FIG. 5 is a diagram illustrating an example of a solid in GP. In the example of this figure, buying and selling is determined by comparing the current stock price with the index of the non-terminal node. In this way, in the GP, first, actions are determined with reference to stock information from the root to the present. A strategy with complex conditional branches can be generated by evolving a tree representing the strategy. Use the board information to reflect the agent's behavior, such as buying, selling, or doing nothing, using the selling board, buying board thickness, bid price, bid price, bid price, bid price, bid price, etc. By doing so, a strategy using GP can be established.

  Trade in a reproducible market using the strategies described above. A thing with a lot of profit of a virtual trading agent who performs trading is assumed to have high fitness. The fitness is provided to the user. In the case of a strategy with high fitness, the user performs a transaction in the actual market the next day using the strategy. In the case of a strategy with low fitness, the user can trade again in the reproduction market using another strategy.

  The transaction support system of the present invention can be used for transactions of stock prices, interest rates, bond interest rates and prices, foreign exchange, futures products, and the like. In particular, it is preferable if board information is used for the transaction.

The table monitoring process may be performed by recording a program for realizing the above steps on a computer-readable recording medium, causing the computer system to read and execute the program recorded on the recording medium. . Here, the “computer system” may include an OS and hardware such as peripheral devices.
Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” means a flexible disk, a magneto-optical disk, a ROM, a writable nonvolatile memory such as a flash memory, a portable medium such as a CD-ROM, a hard disk built in a computer system, etc. This is a storage device.

Further, the “computer-readable recording medium” refers to a volatile memory (for example, DRAM (Dynamic) in a computer system serving as a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. Random Access Memory)) that holds a program for a certain period of time is also included.
The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, and what is called a difference file (difference program) may be sufficient.

FIG. 1 is a diagram illustrating the relationship between the best bid price and the best bid price. FIG. 2 is a diagram showing an outline of the transaction support system according to the present invention. FIG. 3 is a diagram showing an example of a gene code that is solid-stated by a gene algorithm. FIG. 4 is a diagram showing the number of orders for the best bid price and the best bid price. FIG. 5 is a diagram illustrating an example of a solid in GP.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Information collection part 2 Information recording part 3 Data provision part 4 Market reproduction part 5 Strategy creation part 6 Fitness calculation part 7 Provision part

Claims (7)

An information collection unit that collects multiple items of information about individual subjects;
An information recording unit for storing the collected information;
A data providing unit that performs missing value processing on the collected information;
A market reproduction unit that reproduces a past market using the collected information;
A strategy creation unit that creates a strategy based on a predetermined index for target information at the time specified by the user,
Using the created strategy, fitness calculation unit for calculating fitness in the past reproduction market,
A transaction support system comprising: a providing unit that provides the calculated fitness to a user.   The transaction support system according to claim 1, wherein the predetermined index includes board information. An information gathering step for gathering multiple items of information about individual subjects;
An information recording step for accumulating the collected information;
A data providing step for performing missing value processing on the collected information;
A market reproduction step of reproducing a past market using the collected information;
A strategy creation step for creating a strategy based on a predetermined index for target information at the time specified by the user,
Using the created strategy, fitness calculation step for calculating fitness in the past reproduction market,
And a providing step of providing the calculated fitness to a user.   The transaction support method according to claim 3, wherein a predetermined index including board information is used when creating a strategy. An information gathering step for gathering multiple items of information about individual subjects;
An information recording step for accumulating the collected information;
A data providing step for performing missing value processing on the collected information;
A market reproduction step of reproducing a past market using the collected information;
A strategy creation step for creating a strategy based on a predetermined index for target information at the time specified by the user,
Using the created strategy, fitness calculation step for calculating fitness in the past reproduction market,
And a providing step of providing the calculated fitness to a user.   The transaction support program according to claim 5, wherein a predetermined index including board information is used when creating a strategy.   A recording medium capable of reading the transaction support program according to claim 5 or 6 into a computer.