KR20190130535A - Pair Trading System based on the Stock bigdata and Controlling Method for the Same - Google Patents

Abstract

The present invention provides a pair trading system based on stock big data and a control method therefor. The pair trading system comprises: a collection engine module crawling securities-related big data provided from internet sites of various information collection media, classifying the crawled data, and storing the same in a DB; a control module including the collection engine module, controlling the entire process of big data-based pair trading, and performing control such that the pair trading information is processed to calculate a sell/buy time point of a set item and produce an expected scenario; an analysis engine module filtering a big data-based securities item pair trading list by processing and analyzing the big data crawled under the control of functions by the control module; and a result calculation module calculating an actual buy/sell entry timepoint for the securities item pair trading list filtered by the analysis engine module under the control of functions by the control module, calculating the profits, producing a final expected scenario, and transmitting the same to a set customer terminal or electronic address. The present invention crawls the securities-related big data, processes and analyzes the same in multi-stages, produces the securities item pair trading list, calculates the actual buy/sell entry timepoint and produces the best expected scenario, and transmits the same to the set customer, thereby extracting pair trading item groups from all the items regardless of item association and providing a service. As a result, a high-quality service can be provided to securities investors.

Description

Fair Trading System based on the Stock bigdata and Controlling Method for the Same}

The present invention relates to a securities big data-based pair trading system and a control method thereof, in particular, securities-related big data is processed in multiple stages by crawling and analyzing a securities item pair list, and then buying / selling entry points and best expected scenarios. The present invention relates to a securities big data-based pair trading system and a method of controlling the same, by providing a high yield service for a securities investor by calculating and transmitting the same to a set customer.

Generally, with the development of telecommunications and industry, the pattern of stock index is also changing rapidly. In other words, due to the segmentation of the index by industry, the increase in the weight of specific stocks, the composite index declines, but the correlation index increases rather than the specific index soaring. Indices, etc.), the rapid rise and fall of cyber traders, the deepening of groupings in which all stocks belonging to a particular index rise (upper) or all down (lower). The next day, due to the sharp rise in domestic stock prices, not only the fluctuation of the stock index has widened but also the change time has been speeding up. If the share price fluctuates sharply or the upper and lower limits are abolished, the index change will intensify further. Therefore, as the index fluctuates as mentioned above, it is very difficult for a stock investor to make a trading (buy / sell) decision, and at the same time, the decision has a significant impact on the overall profit and loss. Therefore, in recent years, a pair trading method has been used to track correlated stocks in order to maximize such risks and returns.

Then, referring to FIG. 1 for the conventional stock pair trading method as described above, a first step (S101) of crawling related stock groups in the dairy business;

A second step (S102) of analyzing items in the related item group crawled after the first step (S101);

Comprising a fourth step (S103) of calculating the items available for pair trading among the items analyzed after the second step (S102) to provide to the customer.

On the other hand, when looking at the conventional stock trading method more specifically, for example, the relevant stock group in the KOSPI or KOSDAQ listed on the stock exchange of Korea first crawls. In addition, we will analyze the stocks in the related stock groups, such as POSCO, which produces steel, and Hyundai Steel, or Samsung and LG, which produce electronic products. In addition, if there are stocks among the analyzed stocks among the related stocks, such as Samsung Electronics and LG Electronics, which produce electronic products during the analysis process, and they are available for pay trading, they may be collected and then sold or bought. The information is calculated and provided to the customer's terminal.

However, the conventional pair trading method as described above, since the trading for the pair trading items can be calculated only in the dairy business, the service is very limited, and the high quality service provided to the investor is very limited. As the fair trading item is calculated by simple processing at the time, the accuracy of the odds for the buy / sell entry point is low, resulting in a significant decrease in the utilization of information.

Therefore, the present invention has been invented to solve the problems of the prior art as described above, after crawling and analyzing securities related big data in multiple stages and calculating the stock list of stock items, the actual buying / selling entry point and the best expected scenario. By calculating and sending to the set customer, The purpose of this study is to provide a pair trading system based on securities big data that can provide the best high-profit service by discovering the pair trading stocks calculated for all stocks regardless of the relatedness of stocks.

Another object of the present invention is to provide stock investors with the best buy and sell entry point and best forecast scenario for the pair trading items calculated on the basis of securities big data to make an optimized trading decision accordingly To provide a securities big data-based pair trading system and a control method thereof that can minimize losses.

The present invention for achieving the above object is a collection engine module for crawling the securities-related big data provided from the Internet site of the various information collection media and classify the crawled data in a DB;

A control module for controlling the entire process of big data-based pair trading including the collection engine module and processing the pair trading information to calculate a sell / buy time point and an expected scenario of a set item;

An analysis engine module for filtering and analyzing big data crawled under a function control of the control module in multiple stages to filter a big data-based stock trading pair trading list;

Under the control of the control module, the actual buying / selling entry point is calculated on the stock trading pair trading list filtered by the analysis engine module, and then the profit is calculated to calculate the final expected scenario to the set customer terminal or electronic address. It provides a securities big data-based pair trading system including a result calculation module for transmitting.

Still another aspect of the present invention provides a method comprising: a first process of a collecting engine module crawling securities related big data provided from internet sites of various information collecting media under function control of a control module, and classifying the crawled data in a DB;

A second step of analyzing, by the analysis engine module, the crawled big data in a multi-step process and analyzing the big data-based securities pair list after the first process by controlling a function of a control module;

After the second process, the result calculation module calculates the actual buy / sell entry point for the securities item pair list filtered by the analysis engine module under the control of the control module, calculates the profit and calculates the final expected scenario. It provides a control method of a securities big data-based pair trading system including a third process of transmitting to a set customer terminal or an electronic address.

According to the present invention as described above, by crawling securities-related big data processing and analysis in a multi-step, calculate the securities item pair list, calculate the actual buy / sell entry point and the best expected scenario and send to the set customer, the item relevance Irrespective of whether or not the fair trading group is discovered and provided to all stocks, high-quality services can be provided to securities investors.

The present invention as described above also provides the stock investor with the buy / sell entry point and the best forecast scenario for the pair trading items calculated on the basis of the securities big data to make an optimized trading decision accordingly In addition to minimizing losses, it also maximizes returns.

1 is a flowchart illustrating an example of a conventional item pair trading method.
2 is an explanatory diagram illustrating a securities big data based pair trading system according to the present invention.
3 is a flowchart of the present invention.
4 is an explanatory diagram for explaining an example of an item pairing process executed by the analysis engine module of the present invention system.
5 is an explanatory diagram for explaining an example of a result calculation process executed by the result calculation module of the system of the present invention;

As the invention allows for various changes and numerous embodiments, particular embodiments will be illustrated in the drawings and described in detail in the written description. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all transformations, equivalents, and substitutes included in the spirit and scope of the present invention. In the following description of the present invention, if it is determined that the detailed description of the related known technology may obscure the gist of the present invention, the detailed description thereof will be omitted.

Terms such as first and second may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. As used herein, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, action, component, part, or combination thereof described on the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

In addition, in describing the present invention, when it is determined that the detailed description of the related known technology may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted.

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

2 is an explanatory diagram illustrating a securities big data based pair trading system according to the present invention, FIG. 3 is a flowchart of the present invention, and FIG. 4 is an example of an item pairing process executed by the analysis engine module of the present invention system. 5 is an explanatory diagram for explaining an example of a result calculation process executed by the result calculation module of the system of the present invention.

Referring to FIG. 2, various information collecting media 1 crawls securities related big data provided from various portal sites, stock exchange sites, and stock information providing sites, and classifies the crawled data into DB 2. A collection engine module 3 for storing;

A control module (4) including the collection engine module (3) to control the entire process of big data-based pair trading and to process the pair trading information to calculate a sell / buy time point and an expected scenario of a set item;

An analysis engine module (5) for filtering and analyzing big data crawled under a function control of the control module (4) in multiple stages to filter the big data-based securities stock trading list;

Under the control of the control module 4, the actual buying / selling entry point is calculated on the stock trading pair trading list filtered by the analysis engine module 5, and then the profit is calculated to calculate the final expected scenario. And a result calculation module 6 for transmitting to a customer terminal or an electronic address.

In addition, the collection engine module 3 is a stock exchange site and securities providing various information collecting media 1, for example, various portal sites, KOSPI, KOSDAQ, and various futures information through the Internet under the control of the function of the control module 4. It crawls various securities-related big data provided from the information providing site and stores the crawled data in the DB 2 according to the set classification procedure. For example, the collection engine module 3 may be configured by setting quote period data including all KOSPI stock data and all KOSDAQ stock data, index data including KOSPI 200 and KOSDAQ 50, and industry data including KOSPI and KOSDAQ industries. Crawl

In addition, the analysis engine module 5 first extracts only indexes and items that are correlated to all items among the big data crawled by the collection engine module 3 under the function control of the control module 4, and then extracts the items. Calculated indexes and spreads among stocks (differences between stocks). Furthermore, the analysis engine module 5 has a result obtained by processing as described above, and firstly analyzes whether there is a correlation using the correlation formula of [Equation 1] and then cannot predict a random item. Secondly, this normality (stability) (where the position of a certain range and the price volatility of a specific item within a set period) is analyzed secondly, and the result extracted after the second analysis is the third normality ( Analyze whether there is a normal distribution-an item that can predict price fluctuations easily. For example, the analysis engine module 5 selects two random items and calculates correlations between the corresponding items in Equation 1. At this time, the price change is calculated through logarithmic difference, and then Equation 2 is used. The correlation between the two stocks is calculated by obtaining the price change.

[Equation 1]

[Equation 2]

In addition, the analysis engine module 5 verifies normality (stability) by using correlation coefficients of two items obtained by Equations 1 and 2 under the control of the control module 4. First, variables are calculated through linear regression of Equation 3.

[Equation 3]

In addition, the analysis engine module 5 calculates the variables through the linear regression of Equation 3 as described above, and then verifies normality through the regression variables calculated by Equation 3 through Equation 4 ( The verification condition is p-value <0.05 of the regression model based on equation (4).

[Equation 4]

Therefore, when the analysis engine module 5 executes through the above steps, it reduces the data size and calculates the optimal item 1 and item 2 for pair trading.

In addition, the result calculation module 6 calculates selling / buying entry points for the pair trading items calculated by the analysis engine module 5 and creates a scenario for the expected profit and provides it to the customer. In this case, the result calculation module 6 will be described in more detail. The result calculation module 6 obtains a jet score (standard score) from a spread calculated by the analysis engine module 5, and sets a parameter threshold value, for example, When the spread is the biggest and the spread is the largest or smaller than 2 (parameter threshold), the sell / buy entry point is determined. Provide to customers. That is, the result calculation module 6 verifies normality (stability) using Equation 5, wherein the result calculation module 6 first calculates an entry point strategic model and uses the result of calculation A pair trading item pair is calculated.

[Equation 5]

Next, the control method of the present invention having the above configuration will be described.

According to the method of the present invention, as shown in FIG. 3, the collection engine module crawls the securities related big data provided from the Internet site of various information collecting media under the function control of the control module, and classifies the crawled data and stores them in the DB A first process (S1);

A second step (S2) of, after the first step (S1), an analysis engine module filtering and analyzing the big data-based securities pair list by processing and analyzing the crawled big data in multiple stages under the function control of a control module;

After the second process (S2), the result calculation module calculates the actual buying / selling entry point for the securities item pair list filtered by the analysis engine module under the function control of the control module, and then calculates the profit to be the final expected scenario. It is configured to include a third process (S3) to calculate and transmit to the set customer terminal or electronic address.

In the first step (S1), the collection engine module provides various information collecting media such as various portal sites, KOSPI, KOSDAQ, and various futures information through the Internet under the control of the control module. The method may further include crawling various securities related big data provided from the site and classifying the crawled data according to a set classification procedure and storing the crawled data in a DB.

For example, in the first process (S1), the collection engine module includes quote data including KOSPI stock data and KOSDAQ stock data, index data including KOSPI 200 and KOSDAQ 50, and KOSPI and KOSDAQ stocks. The method further includes a precise crawling step for crawling industry data.

In addition, in the second process (S2), the analysis engine module first extracts only indexes and items that are correlated to all items among the big data crawled by the collection engine module under the control of the control module;

A data processing step of calculating a spread (difference between items) between the extracted indexes and items after the extraction step;

A first analysis step of analyzing whether or not there is a correlation by using a correlation equation first with the result extracted after the data processing step;

A second analysis step of analyzing whether random items that cannot be predicted after the first analysis step have normality (stability) (whether there is less price volatility within a range range or within a predetermined period of time); ,

After the second analysis step, the third step further includes a third analysis step of analyzing whether the extracted result is third-order normality (regular distribution-whether the item is easy to predict price fluctuations).

Therefore, if the data is executed through the detailed steps of the second process S2 as described above, the data size can be reduced by the analysis engine module, and the optimal item 1 and item 2 for pair trading can be calculated. .

Furthermore, in the third process S3, the result calculation module calculates selling / buying entry points for the pair trading items calculated by the analysis engine module, respectively, and provides a scenario for the expected profit to the customer. It includes more. At this time, in the third process (S3), the result calculation module obtains a jet score (standard score) from the spread calculated by the analysis engine module, and obtains a parameter threshold, for example, larger or smaller than 2, so that the spread will be greatest. In this case, an entry point judgment step of judging the point of entry of a sell / buy and an anticipation of providing prices to the customer by creating scenarios for expected profits by date on a graph or report after the entry point judgment step are made. Further comprising the scenario calculation step.

In other words, in order to use the securities big data-based pair trading system 7 according to the present invention, first, the control module 4 uses the collection engine module 3 to collect various information collecting media 1, for example, various portal sites, It crawls the stock-related big data provided from the stock exchange site and the stock information providing site and stores the crawled data in the DB (2). When the crawling process is completed as described above, the control module 4 processes and analyzes the crawled big data in multiple stages through the analysis engine module 5 to filter the big data-based stock item pair list. In addition, after the analysis process as described above, the control module 4 calculates the actual buy / sell entry point for the securities item pair list filtered by the analysis engine module 5 through the result calculation module 6, and then generates profits. The final expected scenario is calculated and then sent to the set customer terminal or electronic address, so that pair trading based on securities big data is executed.

At this time, looking at the data crawling process in more detail, the collection engine module (3) is a variety of information gathering medium (1), for example, various portal sites, KOSPI, KOSDAQ and the Internet through the function control of the control module (4) It crawls various stock-related big data provided from the stock exchange site and the stock information providing site that provide various futures information, and stores the crawled data in the DB 2 according to the set classification procedure. In the data crawling process, the collection engine module 3 may include big data collection targets such as quote data including all KOSPI stock data and all KOSDAQ stock data, index data including KOSPI 200 and KOSDAQ 50, and KOSPI. Industry data, including industry and KOSDAQ, will be classified and crawled precisely.

Further, in the analysis process, the analysis engine module 5 first extracts only indexes and items that are correlated to all items among the big data crawled by the collection engine module 3 under the control of the control module 4. Then, the extracted index and the spread between the stocks (differences between stocks) are calculated. At this time, the analysis engine module 5 analyzes whether there is a correlation using the correlation expression with the extracted result calculated through the process of calculating the spread and the first, and then predicts the extracted result with the same result. Secondly, we analyze whether the random item which cannot be done is normal (stability), and then again, if the extracted result is normal (regular distribution – item that is easy to predict price fluctuation). Therefore, if the analysis is carried out through the detailed analysis process as described above, the data size of the big data, which was largely crawled by the analysis engine module 5, is considerably reduced, and the best stocks 1 and 2 can be traded. Will yield.

In more detail, the analysis engine module 5 analyzes whether or not there is a correlation using the correlation equation [Equation 1] with the result extracted by processing as described above. Secondly, we analyze whether random stocks that cannot be predicted have normality (stability) (that is, when the price range of certain stocks are small within a certain period of time) and after the second analysis. Analyze whether the extracted results are third-order normality (regular distribution-stocks with easy prediction of price changes). For example, the analysis engine module 5 selects two random items and calculates correlations between the corresponding items in Equation 1. At this time, the price change is calculated through logarithmic difference, and then Equation 2 is used. The correlation between the two stocks is calculated by obtaining the price change.

[Equation 1]

[Equation 2]

In addition, the analysis engine module 5 verifies normality (stability) by using correlation coefficients of two items obtained by Equations 1 and 2 under the control of the control module 4. First, variables are calculated through linear regression of Equation 3.

[Equation 3]

In addition, the analysis engine module 5 checks the normality through the regression variable calculated by Equation 3 through Equation 4 after calculating the variables through linear regression of Equation 3 as described above ( The verification condition is p-value <0.05 of the regression model based on equation (4).

[Equation 4]

Therefore, when the analysis engine module 5 executes through the above steps, it reduces the data size and calculates the optimal item 1 and item 2 for pair trading.

For example, referring to the graph of FIG. 4, the targets of the stocks were acquired by Daewoo Securities and Gwangju Shinsegae as non-identical sectors, and the period was set to about one year (June 2015-June 2016). The correlation coefficient calculated in the equations 1 to 4 is 0.95, and the Daewoo Securities (Item 1) and Gwangju Shinsegae (Item 2) showed a correlation of 0.95 for 180 business days.

Furthermore, in the result calculation process, the result calculation module 6 calculates selling / buying entry points for the pair trading items calculated by the analysis engine module 5 and creates a scenario for the expected profit. In this case, the result calculation module 6 obtains a jet score (standard score) from the spread calculated by the analysis engine module 5 to obtain a parameter threshold value, for example, greater than or less than 2, and the spread is calculated. At the highest level, we determine when to enter a sell / buy entry, and then provide prices to the customer by creating forecast scenarios for expected earnings on a date or graph in a report.

Herein, the result calculation process will be described in more detail. Furthermore, the result calculation module 6 calculates selling / buying entry points for the pair trading items calculated by the analysis engine module 5, respectively. Create a scenario for the expected revenue and provide it to the customer. In this case, the result calculation module 6 will be described in more detail. The result calculation module 6 obtains a jet score (standard score) from a spread calculated by the analysis engine module 5, and sets a parameter threshold value, for example, When the spread is the biggest and the spread is determined to be greater than or less than 2 (parameter threshold), the sell / buy entry point is determined, and the price of the determined sell / sell entry point is created in a graph or report for scenarios of estimated earnings by date. Provide to customers.

That is, the result calculation module 6 verifies normality (stability) using Equation 5, wherein the result calculation module 6 first calculates an entry point strategic model and uses the result of calculation A pair trading item pair is calculated.

[Equation 5]

For example, when the result calculation module 6 assumes Z-Scores of the Daewoo Securities (item 1) and Gwangju Shinsegae (item 2) as ± 2, the number of changes of the spread sign is 180 based on the business day. 18 times during the day, ensure that there are enough days available for trading during that period. After that, investors can run a profit simulation targeting the two stock combinations. As a result, as shown in FIG. 5, the two stocks can be regarded as a stable stock price fluctuation during the period. In addition, it can be seen that the quote graph drawn by the corresponding pair of products becomes a significant regression model having P-value <= 0.5 in the safety test.

1: Various information collecting media 2: DB
3: collection engine module 4; Control module
5: Analysis engine module 6: Result calculation module
7: Fair Trading System

Claims (2)

A collection engine module for crawling securities related big data provided from internet sites of various information collecting media and classifying the crawled data in a DB;
A control module for controlling the big data-based pair trading process including the collection engine module and processing the pair trading information to calculate a sell / buy time point and an expected scenario of a set item;
An analysis engine module for filtering the big data-based securities stock trading list by processing and analyzing the big data crawled in multiple stages under the control of the control module;
Under the control of the control module, the actual buying / selling entry point is calculated on the stock trading pair trading list filtered by the analysis engine module, and then the profit is calculated to calculate the final expected scenario to the set customer terminal or electronic address. It further comprises a result calculation module for transmitting;
The collection engine module further performs a function of crawling quote data including all KOSPI stock data and KOSDAQ stock data, index data including KOSPI 200 and KOSDAQ 50, and industry data including KOSPI and KOSDAQ stocks. ;
The analysis engine module extracts only indexes and stocks that are correlated to all stocks among the big data crawled by the collection engine module under the control of the control module, and calculates a spread between the extracted indexes and stocks. Including more;
The analysis engine module analyzes whether the correlation is a first order using a correlation expression with the processed extracted result, and then analyzes whether the random items that cannot be predicted are normal (stability) in the second order. Further including a function of analyzing whether the extracted result is third-order normal after the second analysis;
The result calculation module obtains a jet score from the spread calculated by the analysis engine module, obtains a value larger or smaller than the set parameter threshold, and judges the time when the spread is the largest, when the sell / buy entry point is made. Further comprising providing prices to entry points to the customer by creating scenarios for expected revenue per day in graphs or reports;
The analysis engine module selects two random stocks in order to calculate the pair trading stocks on the stock trading pair trading list, calculates a change in price through a log difference, and calculates a correlation between the stocks in Equation 1 below. After calculating the correlation coefficient of the two items, the variables are calculated through the linear regression of Equation 3 below using the correlation coefficient of the two items obtained under the control function of the control module to verify the normality (stability) Securities big data-based pair trading system.
[Equation 1]

[Equation 3]

A first process of the collecting engine module crawling securities related big data provided from internet sites of various information collecting media under function control of the control module, and classifying the crawled data in a DB;
A second process of analyzing and analyzing the crawled big data in a multi-step process by the analysis engine module under the function control of a control module after the first process to filter the big data-based stock item pair list;
After the second process, the result calculation module calculates the actual buy / sell entry point for the securities item pair list filtered by the analysis engine module under the control of the control module, calculates the profit and calculates the final expected scenario. Including a third process of transmitting to a set customer terminal or an electronic address;
In the first process, the collection engine module crawls quote data including all KOSPI and KOSDAQ stock data, index data including KOSPI 200 and KOSDAQ 50, and industry data including KOSPI and KOSDAQ industries. Further comprising steps;
In the second process, an analysis engine module extracts an index and an item that are correlated to all items among the big data crawled by the collection engine module under the control of the control module; A data processing step of calculating the indexes and spreads between the items (differences between the items);
In the third process, the result calculation module obtains a jet score (standard score) from the spread calculated by the analysis engine module, and obtains a value larger or smaller than a set parameter threshold to determine when to enter a sell / buy point when the spread is most widespread. An entry scenario determination step, and an expected scenario calculation step of providing prices to the customer by creating scenarios for expected revenue by date on a graph or report after the entry point determination step;
In the data processing step, a first analysis step of analyzing whether the analysis engine module has a result extracted by processing through the data processing step and correlation using a correlation equation firstly, and after the first analysis step A second analysis step of analyzing whether the random item that the engine module cannot predict is normal (stability) in a second step, and whether the extracted result is third-order normality after the second analysis step; The module further includes a third analysis step;
In the second process, the analysis engine module selects two random stocks to calculate the pair trading stocks on the stock trading list and calculates the correlation between the stocks in Equation 1 below to calculate the price change through the log difference. Calculating the correlation coefficients of the two items, and calculating the variables through the linear regression of Equation 3 below using the correlation coefficient of the two items obtained by the analysis engine module under the control of the control module. Stability) further comprising the step of controlling the securities big data-based pair trading system control method.
[Equation 1]

[Equation 3]

Images