KR102110419B1 - Apparatus and method for determining investment portfolio - Google Patents

Abstract

The portfolio building device for building an investment portfolio includes an asset selection unit that selects a plurality of assets necessary for the construction of an investment portfolio, and a dimensionality reduction performance unit that performs dimensionality reduction on the returns of multiple assets that are selected using a dimension reduction algorithm. Add-on assets based on the asset classifier that targets multiple assets that are selected based on errors caused by dimensionality reduction into first and second asset groups, and target and benchmark asset returns. It may include an add-on asset yield calculation unit for calculating the return of the investment and the investment weight determining unit for determining the investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

Description

Apparatus and method of building an investment portfolio {APPARATUS AND METHOD FOR DETERMINING INVESTMENT PORTFOLIO}

The present invention relates to an apparatus and method for building an investment portfolio.

Most individuals and investors are operating on assets through investment products such as various financial institutions because they lack knowledge of financial products and it is difficult to understand the details of various financial products that have changed or newly created investment conditions. .

In a conventional financial institution, since a financial company employee has recommended a direct investment product and created a portfolio in a method of providing an investment portfolio to an individual investor, a portfolio of different quality may be created according to employee competency, and inconsistent investment Since the portfolio can be created according to the direction, the reliability of the portfolio was inevitably low.

In addition, despite the different asset conditions and investment conditions of customer investors, the uniformly provided financial product portfolio is at risk of causing significant losses to customer investors in the rapidly changing financial product market.

Korean Patent Publication No. 2014-0022707 (released on February 25, 2014)

The present invention performs a dimensional reduction on the yield of a plurality of assets selected by using a dimensionality reduction algorithm for a plurality of assets selected for the construction of an investment portfolio, and a plurality of assets selected based on errors caused by the dimensionality reduction We want to classify the assets of the first asset group and the second asset group. In addition, the present invention calculates the yield of an add-on asset based on the target yield and the yield of the benchmark asset, and each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset Wants to determine the share of the investment. However, the technical problems to be achieved by the present embodiment are not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above-described technical problem, a portfolio building apparatus for constructing an investment portfolio according to the first aspect of the present invention includes an asset selection unit for selecting a plurality of assets necessary for the construction of the investment portfolio; A dimensionality reduction unit performing dimensionality reduction on the returns of the plurality of assets selected using a dimensionality reduction algorithm; An asset classifying unit classifying the plurality of assets selected as first asset groups and second asset groups based on errors caused by the dimensionality reduction; An add-on asset yield calculator configured to calculate a yield of an add-on asset based on the target yield and the benchmark asset's return; And an investment weight determining unit determining an investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

A method of building an investment portfolio in the portfolio building apparatus according to the second aspect of the present invention comprises the steps of selecting a plurality of assets necessary for the construction of the investment portfolio; Performing a dimensionality reduction on the returns of the plurality of assets selected using a dimensionality reduction algorithm; Classifying the selected plurality of assets into a first asset group and a second asset group based on an error caused by the dimensionality reduction; Calculating a yield of an add-on asset based on the target yield and the yield of the benchmark asset; And determining an investment proportion of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

The above-described problem solving means are merely examples, and should not be construed as limiting the present invention. In addition to the exemplary embodiments described above, there may be additional embodiments described in the drawings and detailed description of the invention.

According to any one of the above-described problem solving means of the present invention, the present invention establishes a first investment portfolio corresponding to a benchmark asset and a second investment portfolio corresponding to an add-on asset, resulting in the risk of the rise or fall of a financial market or financial product. Can provide an investment portfolio that can minimize

In addition, since the present invention determines the share of investment in each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset, the investment is managed while the investor tracks the benchmark asset and invests at the same time Opportunity costs can be identified to respond flexibly, and as a result, more effective investment can be made.

1 is a block diagram of a portfolio building system according to an embodiment of the present invention.
2 is a block diagram of the portfolio building apparatus shown in FIG. 1 according to an embodiment of the present invention.
3A to 3C are diagrams for explaining a method of performing a dimensional reduction on a yield of a plurality of assets according to an embodiment of the present invention.
4 is a view showing a target rate of return, a benchmark asset, and a return rate of an add-on asset according to an embodiment of the present invention.
5 is a flowchart illustrating a method of building an investment portfolio according to an embodiment of the present invention.
6 is a flowchart illustrating a method of building an investment portfolio according to another embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art to which the present invention pertains may easily practice. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. In addition, in order to clearly describe the present invention in the drawings, parts irrelevant to the description are omitted, and like reference numerals are assigned to similar parts throughout the specification.

Throughout the specification, when a part is "connected" to another part, this includes not only "directly connected" but also "electrically connected" with another element in between. . Also, when a part “includes” a certain component, this means that other components may be further included, rather than excluding other components, unless otherwise specified.

In the present specification, the term “unit” includes a unit realized by hardware, a unit realized by software, and a unit realized by using both. Further, one unit may be realized by using two or more hardware, and two or more units may be realized by one hardware.

Some of the operations or functions described in this specification as being performed by a terminal or device may be performed instead on a server connected to the corresponding terminal or device. Similarly, some of the operations or functions described as being performed by the server may be performed in a terminal or device connected to the corresponding server.

Hereinafter, specific contents for carrying out the present invention will be described with reference to the accompanying drawings or process flow charts.

1 is a block diagram of a portfolio building system according to an embodiment of the present invention.

Referring to FIG. 1, the portfolio building system may include a portfolio building device 100 and a plurality of investor terminals 110. However, since the portfolio building system of FIG. 1 is only an embodiment of the present invention, the present invention is not limitedly interpreted through FIG. 1, and may be configured differently from FIG. 1 according to various embodiments of the present invention.

Generally, each component of the portfolio building system of FIG. 1 is connected through a network (not shown). Network refers to a connection structure capable of exchanging information between nodes such as terminals and servers, and a local area network (LAN), a wide area network (WAN), and the Internet (WWW: World) Wide Web), wired and wireless data communication networks, telephone networks, and wired and wireless television communication networks. Examples of wireless data communication networks include 3G, 4G, 5G, 3rd Generation Partnership Project (3GPP), Long Term Evolution (LTE), World Interoperability for Microwave Access (WIMAX), Wi-Fi, Bluetooth communication, infrared communication, ultrasound Communication, Visible Light Communication (VLC), LiFi, and the like are included, but are not limited thereto.

The portfolio building apparatus 100 may select a plurality of assets required for building an investment portfolio.

The portfolio building apparatus 100 may perform dimensionality reduction on the returns of a plurality of assets selected using a dimensionality reduction algorithm. Here, the dimensionality reduction algorithm may be a principal component analysis algorithm or an auto encoder algorithm.

The portfolio building apparatus 100 may classify the selected plurality of assets into a first asset group and a second asset group based on an error caused by a dimensional decrease in the yield of the plurality of selected assets. Specifically, the portfolio building apparatus 100 groups assets belonging to a first criterion (assets belonging to or smaller than a first size) in which the size of an error generated due to a dimensional decrease in the yield of the asset is grouped into a first asset group, It is possible to group assets belonging to a second criterion (assets belonging to a second size or more) in which the size of the error caused by the dimensionality reduction is preset into a second asset group.

The portfolio building apparatus 100 may calculate the return of the add-on asset based on the target return and the return of the benchmark asset. Here, the target rate of return means the desired rate of return of the investor received from the investor terminal, and the rate of return of the benchmark asset refers to the rate of return to be compared when measuring portfolio investment performance. For example, the yield of the benchmark asset may include the rate of return shown in KOSPI200, S & P500, and the like. In addition, the return of an add-on asset is an excess return on a benchmark, and may mean a return on an asset that does not follow the flow of the general financial market.

The portfolio building apparatus 100 extracts the asset that follows the yield of the benchmark asset from the first asset group, builds the first investment portfolio, and extracts the asset that follows the yield of the add-on asset from the second asset group, and extracts the second Build a portfolio.

The portfolio building apparatus 100 may determine an investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset. For example, the portfolio building apparatus 100 is based on the investment propensity information of each investor of the plurality of investor terminals 110, the proportion of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset And may provide a customized portfolio to the plurality of investor terminals 110 based on the determined specific gravity.

Hereinafter, the operation of each component of the portfolio building system of FIG. 1 will be described in more detail.

2 is a block diagram of the portfolio building apparatus 100 shown in FIG. 1 according to an embodiment of the present invention.

Referring to FIG. 2, the portfolio building apparatus 100 includes an asset selection unit 200, a dimension reduction performing unit 210, an asset classification unit 220, an add-on asset yield calculation unit 230, and an investment portfolio construction unit 240 ) And the investment weight determining unit 250. Here, the dimension reduction execution unit 210 may include an error calculation unit 212. However, the portfolio building apparatus 100 illustrated in FIG. 2 is only one implementation example of the present invention, and various modifications are possible based on the components illustrated in FIG. 2.

The asset selection unit 200 may select a plurality of assets necessary for building an investment portfolio. Here, as a plurality of types of assets, for example, stock assets, stock mixed assets, bond assets, bond mixed assets, commodities, stock funds (eg, repurchase condition bond-linked funds (repo), short-term financial funds (MMF), etc.) ), And the style of the asset may include, for example, large-cap, small-cap, dividend-share, value-added, index, leverage, unlisted, commodity spot, commodity futures.

For example, the asset selection unit 200 may select assets based on the items included in the ranking (eg, Copis 200) in which the KOSPI market capitalization ranking is preset. Alternatively, the asset selection unit 200 selects an asset mainly based on an item set according to an investment manager's interest item (for example, a stock item mainly based on large stocks, a stock item mainly based on commodity futures, and an item intensively recommended by stock analysts). can do.

The dimensionality reduction performing unit 210 may perform dimensionality reduction on the returns of a plurality of assets selected using a dimensionality reduction algorithm. Here, the dimensionality reduction algorithm may be, for example, a principal component analysis algorithm or an auto encoder algorithm.

Specifically, the dimension reduction performing unit 210 extracts a plurality of duplicate assets having characteristics similar to each of the plurality of selected assets by inputting a yield of each of the plurality of selected assets into the dimensionality reduction algorithm, and extracts the plurality of extracted duplicate assets The rate of return for a predetermined period for each may be output.

Referring to FIG. 3A for a moment, a method of performing dimensional reduction on the yields of a plurality of assets will be described.

Referring to FIG. 3A, the dimensionality reduction performing unit 210 inputs a rate of return 300 for each of the N plurality of assets (assets selected by the asset selection unit 200) into the principal component analysis algorithm, and the N plurality of The dimensionality reduction may be performed by returning the assets of the assets as the yields 302 of K multiple assets.

Here, the principal component analysis algorithm may be an algorithm that performs dimensional reduction of the inputted N plurality of assets into K plurality of assets.

Here, K multiple assets having the same or similar asset characteristic condition among N multiple assets may be extracted based on the asset characteristic conditions preset in the principal component analysis algorithm, and each of the N multiple assets in the principal component analysis algorithm itself It is also possible to extract K assets having the same or similar asset characteristics by grouping multiple assets having the same or similar characteristics by analyzing the characteristics of.

For example, the dimensionality reduction performing unit 210 has a plurality of K assets having a rate of return directly proportional to the stock index (eg, composite index, KOSDAQ index, KOSPI index, etc.) among the N plurality of assets through the principal component analysis algorithm. Can be extracted. Thereafter, the dimensionality reduction performing unit 210 may recover the yields of the K plurality of assets to the yields of the N plurality of assets based on the characteristics of each of the K plurality of assets extracted by the principal component analysis algorithm. Here, the restored N plurality of assets may be duplicate assets in which the N plurality of assets selected by the asset selection unit 200 are duplicated.

The dimensionality reduction performing unit 210 may derive the yield 304 for each of the plurality of extracted duplicate assets.

In another embodiment, the dimensionality reduction performing unit 210 extracts features for N multiple assets input using an auto-encoder algorithm, and K multiples having the same or similar asset characteristics based on the extracted features. After extracting the duplicate assets, and recovering (i.e., replicating) from the extracted K multiple duplicate assets to N multiple duplicate assets that are the same or similar to the input multiple N multiple assets, that is, the duplicate assets are stored for each of the multiple duplicate assets. Yields for a set period can be derived.

Here, the auto-encoder algorithm is a process in which input variables (N multiple assets) are extracted through the encoder as K multiple duplicate assets having the same or similar asset characteristics, and K multiple duplicate assets based on the extracted asset characteristics. The process of recovering from the output variable (N plurality of duplicate assets) through the decoder may be an algorithm implemented as a decal-combined shape.

The error calculator 212 may calculate an error for the replication yield of each of the plurality of extracted duplicate assets. Referring to FIGS. 3A to 3B, based on the difference between the yield 300 of each of the N plurality of assets and the yield 304 of the replicated assets corresponding to each of the N plurality of assets, for the replication yield for each N replication assets Error 306 can be calculated.

The asset classifying unit 220 may classify a plurality of assets selected based on an error caused by the dimensionality reduction into a first asset group and a second asset group.

For example, the asset classifying unit 220 may classify the plurality of assets selected based on the size of the error for the replication yield of each of the plurality of duplicate assets into a first asset group and a second asset group. Here, the first asset group is a group including an asset belonging to a first criterion (an asset belonging to a first size or less) in which the size of the error for the replication yield is preset, and the second asset group is the size of the error for the replication yield It may be a group that includes assets belonging to a predetermined second criterion (assets belonging to a second size or higher). If classified as such, the first asset group reflects the overall market situation that commonly affects assets. Groups can be formed of assets. In the second asset group, a group may be formed of assets with unique characteristics that exhibit returns that are different from the common flow of the market. Here, the plurality of assets belonging to the first asset group are assets having a small error size for the replication yield (ie, assets having similar origins to the replication yield), and the plurality of assets belonging to the second asset group are errors for the replication yield. It may be a large asset (i.e., an asset with a large difference between the replication yield and the original yield).

The asset classifying unit 220 first and second assets of the plurality of assets selected based on the cosine similarity or Euclidean distance for the error size of each of the plurality of duplicate assets. Can be classified into groups. Here, the cosine similarity refers to the degree of similarity to the error size of each replication yield for each replication asset, and the Euclidean distance may mean a range to which the error size of each replication yield for each replication asset belongs.

For example, the asset classifying unit 220 classifies assets belonging to the first range (eg, lower 30%) of the Euclidean distance for the size of the error of the duplication rate among the plurality of selected assets as the first asset group, Assets in which the Euclidean distance for the magnitude of the error of the replication yield rate falls within the second range (eg, the top 30%) may be classified into a second asset group. The classification into the first asset group and the second asset group can be configured in various ways based on the relationship between errors in addition to the relative magnitude of errors.

The add-on asset yield calculation unit 230 may calculate the return of the add-on asset based on the target return and the benchmark asset return. For example, the add-on asset yield calculation unit 230 may calculate the return of the add-on asset during the predetermined period based on the target return rate desired by the investor (customer) and the benchmark asset during the predetermined period.

Referring to FIG. 4, since the yield 402 of the benchmark asset follows the flow of the general financial market, it will show a movement as shown in FIG. 4.

On the other hand, in general, since the target rate of return 400 is set higher than the rate of return in the general financial market, as shown in FIG. 4, the target rate of return 400 is set higher than the rate of return of the benchmark asset 402.

That is, if only the benchmark asset that follows the flow of the general financial market is invested, it is difficult to achieve the target rate of return 400.

In the present invention, the target return rate 400 may be achieved by extracting an investment strategy that duplicates not only the benchmark asset but also the add-on asset return rate that does not follow the general financial market flow and reflects it in the investment portfolio.

Here, the target rate of return 400 refers to an investor's desired rate of return received from the investor terminal, and the benchmark asset's rate of return 402 is a general financial market, such as a portfolio of stocks representing stocks such as large or superior stocks. It can mean the rate of return for assets with the flow of. The benchmark asset is likely to have a close relationship with the K plurality of duplicate assets on which the auto-encoder algorithm has been performed among the input N plurality of assets, for example, as shown in FIG. 3A. Because the K plurality of duplicate assets on which the auto-encoder algorithm is performed may include information describing the common movement between the N multiple assets, and the benchmark asset also reflects the common movement of assets in the market. Because it is designed.

The return rate of the add-on asset (404) corresponds to the return on assets that do not follow the flow of the general financial market (e.g., small caps, futures options, etc.), and the add-on asset deviates from the general flow of financial markets exposed through benchmark assets. Sexual assets can contain information that describes the unique movement of each individual stock.

The add-on asset yield calculation unit 230 calculates the return 404 of the add-on asset for a predetermined period based on the difference between the target return 400 and the return 402 of the benchmark asset, as shown in [Equation 1]. can do.

[Equation 1]

Return on add-on assets = Target return-Return on benchmark assets

The investment portfolio construction unit 240 extracts the first asset group from the asset having high similarity with the yield of the benchmark asset among the first asset group (the group including the assets having the error size for the replicated yield that is less than or equal to the preset first size). Build an investment portfolio, extract assets with high returns and similarities to add-on assets from the second asset group (groups with assets that fall within the second size or more with an error size for the replicated yield) to extract the second investment portfolio. Can build.

Specifically, the investment portfolio building unit 240 inputs the yields of the plurality of assets belonging to the first asset group and the yields of the benchmark assets into a machine learning algorithm (eg, an artificial neural network analysis algorithm, etc.), and You can build a first investment portfolio by extracting highly similar assets.

In addition, the investment portfolio construction unit 240 extracts assets with high returns and similarities to those of add-on assets by inputting the returns of multiple assets belonging to the second asset group and the returns of add-on assets into the machine learning algorithm to extract the second investment portfolio. Can build.

The investment weight determining unit 250 may determine an investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

The investment weight determining unit 250 may determine the weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset, based on the investment propensity information of the investor or the investor's investment and asset management situation. have. For example, if the investor has a conservative tendency to invest, the investment proportion determining unit 250 corresponds to the benchmark asset rather than the investment proportion of the second investment portfolio corresponding to the add-on asset (for example, 20%). The investment share (eg 80%) of the first investment portfolio can be set high. Conversely, when an investor has an aggressive tendency to invest, the investment proportion determining unit 250 is a second investment corresponding to an add-on asset rather than an investment proportion (eg, 20%) of the first investment portfolio corresponding to the benchmark asset. You can set the portfolio's investment share (eg 80%) high.

In addition, the investment proportion determining unit 250 may determine an investment ratio of each of a plurality of assets included in each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

For example, the investment weight determining unit 250 simulates the returns on each of the plurality of assets constituting the first investment portfolio corresponding to the benchmark asset, and the respective assets constituting the simulated first investment portfolio. You can simulate the final rate of return of your first investment portfolio by entering the rate of return for your machine learning algorithm.

Subsequently, the investment weight determining unit 250 performs a linear regression analysis in which the final returns of the simulated first investment portfolio are dependent variables, and the returns for each of the plurality of assets constituting the simulated first investment portfolio are independent variables. It can be done.

The investment weight determining unit 250 may derive a yield coefficient for each of the plurality of assets constituting the first investment portfolio from the linear regression analysis.

The investment weight determining unit 250 may determine an investment ratio of each of the plurality of assets included in the first investment portfolio based on the yield coefficients for each of the plurality of assets constituting the derived first investment portfolio.

Meanwhile, the investment proportion determining unit 250 simulates the returns on each of the plurality of assets constituting the second investment portfolio corresponding to the add-on assets, and returns the returns on each of the plurality of assets constituting the simulated second investment portfolio. You can simulate the final return of your second investment portfolio by entering it into a machine learning algorithm.

Thereafter, the investment weight determining unit 250 may perform a linear regression analysis in which the final rate of return of the simulated second investment portfolio is a dependent variable and the rate of return for each of the plurality of simulated assets is an independent variable.

The investment weight determining unit 250 may derive a return coefficient for each of the plurality of assets constituting the second investment portfolio from the linear regression analysis. The investment weight determining unit 250 may determine an investment ratio of each of the plurality of assets included in the second investment portfolio based on the yield coefficients for each of the plurality of assets constituting the derived second investment portfolio.

On the other hand, those skilled in the art, the asset selection unit 200, the dimension reduction performing unit 210, the error calculation unit 212, the asset classification unit 220, the add-on asset yield calculation unit 230, the investment portfolio construction unit 240 And it will be fully understood that each of the investment weight determining unit 250 may be implemented separately, or one or more of them may be integrated and implemented.

5 is a flowchart illustrating a method of building an investment portfolio according to an embodiment of the present invention.

Referring to FIG. 5, in step S501, the portfolio building apparatus 100 may select a plurality of assets necessary for building an investment portfolio.

In step S503, the portfolio building apparatus 100 may perform dimensional reduction on the returns of the plurality of assets selected using the dimensionality reduction algorithm. Here, the dimensionality reduction algorithm may be, for example, a principal component analysis algorithm or an automatic encoder algorithm.

In step S505, the portfolio building apparatus 100 may classify the plurality of assets selected based on the error caused by the dimensionality reduction into the first asset group and the second asset group.

In step S507, the portfolio building apparatus 100 may calculate the rate of return of the add-on asset based on the target rate of return and the rate of return of the benchmark asset.

In step S509, the portfolio building apparatus 100 may build a first investment portfolio corresponding to the benchmark asset and a second investment portfolio corresponding to the add-on asset.

In step S511, the portfolio building apparatus 100 may determine an investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.

In the above description, steps S501 to S511 may be further divided into additional steps or combined into fewer steps, according to an embodiment of the present invention. Also, some steps may be omitted if necessary, and the order between the steps may be changed.

6 is a flowchart illustrating a method of building an investment portfolio according to another embodiment of the present invention.

Referring to FIG. 6, in step S601, the portfolio building apparatus 100 may select a plurality of assets necessary for building an investment portfolio.

In step S603, the portfolio building apparatus 100 may perform dimensional reduction on the yields of the plurality of assets selected using the dimensionality reduction algorithm. Here, the dimensionality reduction algorithm may be, for example, a principal component analysis algorithm or an automatic encoder algorithm.

In steps S605 and S607, the portfolio building apparatus 100 classifies assets belonging to a first size or less of a predetermined size of an error that occurs due to a dimensional decrease in yield among a plurality of selected assets, and returns the Assets belonging to a second size or more having a predetermined size of error due to dimensionality reduction may be classified into a second asset group.

In step S609, the portfolio building apparatus 100 extracts an asset having a similarity to the yield of the benchmark asset among the plurality of assets included in the first asset group, and an asset having a similarity to the yield of the benchmark asset extracted in the step S611. Can be used to build a first investment portfolio corresponding to a benchmark asset.

In step S613, the portfolio building apparatus 100 extracts an asset having a similarity to the yield of the add-on asset among the plurality of assets included in the second asset group, and uses an asset having a similarity to the yield of the add-on asset extracted in the step S615. By doing so, it is possible to build a second investment portfolio corresponding to add-on assets.

In step S617, the portfolio building apparatus 100 may determine an investment weight of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset based on the investment propensity information of the investor.

In the above description, steps S601 to S617 may be further divided into additional steps or combined into fewer steps, according to an embodiment of the present invention. Also, some steps may be omitted if necessary, and the order between the steps may be changed.

One embodiment of the present invention can also be implemented in the form of a recording medium including instructions executable by a computer, such as program modules, being executed by a computer. Computer readable media can be any available media that can be accessed by a computer and includes both volatile and nonvolatile media, removable and non-removable media. Also, the computer-readable medium may include any computer storage medium. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data.

The above description of the present invention is for illustration only, and those skilled in the art to which the present invention pertains can understand that the present invention can be easily modified into other specific forms without changing the technical spirit or essential features of the present invention. will be. Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as distributed may be implemented in a combined form.

The scope of the present invention is indicated by the following claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalent concepts should be interpreted to be included in the scope of the present invention. .

100: portfolio building device
110: multiple investor terminals
200: asset selection department
210: dimension reduction execution unit
212: error calculation unit
220: asset classification department
230: add-on asset yield calculator
240: Investment portfolio building department
250: investment proportion determining unit

Claims (15)

In the portfolio building device to build an investment portfolio,
An asset selection unit for selecting a plurality of assets necessary for the construction of the investment portfolio;
A dimensionality reduction unit performing dimensionality reduction on the returns of the plurality of assets selected using a dimensionality reduction algorithm;
An asset classifying unit classifying the plurality of assets selected as first asset groups and second asset groups based on errors caused by the dimensionality reduction;
An add-on asset yield calculator configured to calculate a yield of an add-on asset based on the target yield and the benchmark asset's return; And
And a first investment portfolio corresponding to the benchmark asset and a second investment portfolio corresponding to the add-on asset.
According to claim 1,
The dimension reduction performing unit
Entering the yield of each of the selected plurality of assets into the dimensionality reduction algorithm to extract a plurality of duplicate assets having the same characteristics as each of the selected plurality of assets,
And outputting a yield for a predetermined period of time for each of the extracted plurality of duplicate assets.
According to claim 2,
The dimensionality reduction algorithm is a principal component analysis algorithm or an auto encoder (auto encoder) algorithm, portfolio building apparatus.
According to claim 2,
The dimensionality reduction performing unit further comprises an error calculation unit for calculating an error for the replication yield of each of the extracted plurality of replication assets.
The method of claim 4,
The error calculation unit
The portfolio building apparatus calculates an error for the replication yield based on a difference between the yield for each asset and the yield of the corresponding duplicate asset.
The method of claim 5,
The asset classifying unit is to classify the plurality of assets selected as the first asset group and the second asset group based on the magnitude of the error in the replication yield.
The method of claim 6,
The first asset group is a group including assets belonging to a first criterion in which the size of the error for the duplication rate is preset,
The second asset group is a group that includes an asset that belongs to a second criterion in which the size of the error for the duplication rate is predetermined.
The method of claim 7,
The first investment portfolio is constructed by extracting an asset having a similarity with the yield of the benchmark asset from the first asset group,
And an investment portfolio construction unit configured to extract the asset having a high similarity to the yield of the add-on asset from the second asset group and build the second investment portfolio.
The method of claim 8,
The investment portfolio construction unit extracts assets having high similarities to the yields of the benchmark assets by inputting the yields of the plurality of assets belonging to the first asset group and the yields of the benchmark assets into a machine learning algorithm,
A portfolio building apparatus that extracts assets having high similarities to the yields of the add-on assets by inputting the yields of the plurality of assets belonging to the second asset group and the returns of the add-on assets into the machine learning algorithm.
In the method of building an investment portfolio in the portfolio building device,
Selecting a plurality of assets necessary for the construction of the investment portfolio;
Performing a dimensionality reduction on the returns of the plurality of assets selected using a dimensionality reduction algorithm;
Classifying the selected plurality of assets into a first asset group and a second asset group based on an error caused by the dimensionality reduction;
Calculating a yield of an add-on asset based on the target yield and the yield of the benchmark asset; And
And determining an investment share of each of the first investment portfolio corresponding to the benchmark asset and the second investment portfolio corresponding to the add-on asset.
The method of claim 10,
The step of performing a dimensional reduction on the yield of the plurality of assets selected above is
Extracting a plurality of duplicate assets having the same characteristics as each of the selected plurality of assets by inputting a return rate of each of the selected plurality of assets into the dimension reduction algorithm; and
And outputting a yield for a predetermined period of time for each of the extracted plurality of duplicate assets.
The method of claim 11,
The step of performing a dimensional reduction on the yield of the plurality of assets selected above is
And calculating an error for a replication yield rate of each of the extracted plurality of replication assets.
The method of claim 12,
The step of classifying the selected plurality of assets into a first asset group and a second asset group is
And classifying the selected plurality of assets into a first asset group and a second asset group based on the magnitude of the error for the duplication rate of return.
The method of claim 13,
The first asset group is a group including assets belonging to a first criterion in which the size of the error for the replication yield rate is predetermined,
The second asset group is a group of assets containing assets belonging to a second criterion in which the size of the error for the replication yield is predetermined.
The method of claim 14,
The first investment portfolio is constructed by extracting an asset having a similarity with the yield of the benchmark asset from the first asset group,
And constructing the second investment portfolio by extracting an asset having a similarity to the yield of the add-on asset among the second asset group.

Images