KR102222928B1 - Apparatus, method and computer program for driving financial estimates - Google Patents

Abstract

Provided is a financial estimate deriving device, which comprises: a communication unit communicating with a database or a terminal; an input unit; and a processor which receives user identification information from the terminal through the communication unit or through the input unit; determines a financial estimate target fund based on the user identification information or the user input through the input unit, receives base data corresponding to the financial estimation target fund, extracts at least one principal component from relevant variables included in the base data using a macroscopic financial estimation model, predicts the relevant variables in consideration of population characteristics using a micro-financial estimation model, and derives a financial estimate using the extracted principal component and the predicted relevant variables.

Description

A device for deriving a financial estimate, a method for deriving a financial estimate, and a computer program {APPARATUS, METHOD AND COMPUTER PROGRAM FOR DRIVING FINANCIAL ESTIMATES}

The present invention relates to an apparatus for deriving a financial estimate for a fund, a method for deriving a financial estimate, and a computer program.

Currently, there are 67 funds from the central government of the Republic of Korea, with a total amount of 639 trillion won, and the fund is expected to exceed 1,000 trillion won in 2025. Funds are classified into social insurance fund, financial fund, and business fund depending on their characteristics, and the fund's assets and liabilities may arise due to the purpose of establishing the fund and the original business of the fund. Funds need to be managed in a management area with a stable asset-to-liability ratio.

The main clients in need of financial estimates for the fund are funds, mutual aid associations and asset managers. However, the fund or mutual aid association often lacks the expertise necessary to manage the fund due to rotational positions. In the case of asset managers, asset managers that perform the Outsourced Chief Investment Officer (OCIO) program are the main clients. The OCIO program is an external consignment of fund management, in which an asset manager or securities company with expertise in management is entrusted to manage the fund for a certain period of time.

In the case of asset management companies and securities companies, they have expertise in fund management, but they lack the ability to manage funds in consideration of the characteristics of the funds and the identified characteristics. Therefore, there is a need for an advisory service to compensate for the lack of expertise or management know-how for the party operating the fund.

The present invention provides an apparatus for deriving a fiscal estimate for a fund subject to fiscal estimates, a method for deriving a fiscal estimate, and a computer program.

In addition, the present invention provides an apparatus, a method, and a computer program for grasping the relationship between funds based on the characteristics of each fund.

As a technical means for achieving the above-described technical problem, an apparatus for deriving a financial estimate according to an embodiment of the present specification includes: a communication unit for communicating with a database or a terminal; Input unit; And receiving user identification information input through the terminal or the input unit through the communication unit, determining a fund for financial estimation based on the user identification information or a user input through the input unit, and Receive the corresponding base data, extract at least one principal component from the related variables included in the base data using a macroscopic financial estimation model based on principal component analysis, and determine the population characteristics using a microscopic financial estimation model based on Markov chain. It may include a processor that predicts the considered related variable and derives a financial estimate using the extracted principal component and the predicted related variable.

According to another embodiment of the present specification, a method of deriving a financial estimate includes receiving user identification information, determining a fund to be subject to financial estimate based on the user identification information or input of the user, and responding to the fund to be subject to financial estimate. Receiving base data, extracting at least one principal component from related variables included in the base data using a macroscopic financial estimation model, predicting the related variable considering population characteristics using a microscopic financial estimation model And deriving a financial estimate using the extracted principal component and the predicted related variable.

According to another embodiment of the present specification, a computer program stored in a medium including a sequence of instructions for deriving a financial estimate, when the computer program is executed by a computing device, receives user identification information, and the user identification information or Determine a fund to be subject to financial estimation based on a user's input, receive base data corresponding to the fund to be subject to financial estimate, and extract at least one principal component from related variables included in the base data using a macroscopic financial estimate model And, it may include a sequence of instructions for predicting the related variable in consideration of population characteristics using a microscopic fiscal estimation model, and deriving a fiscal estimate using the extracted principal component and the predicted related variable.

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

According to various embodiments of the present specification, even if there is a lack of expertise or management know-how for the parties operating the fund, it is possible to secure expertise and continuity in fund management by providing an advisory service using a financial estimate, A stable consignment operation of the fund is possible.

In addition, according to various embodiments of the present specification, a more accurate fiscal estimate may be derived using a macro fiscal estimate model and a micro fiscal estimate model.

In addition, according to various embodiments of the present specification, by extracting the principal component using a macroscopic financial estimation model, it is possible to derive a financial estimation plan having high accuracy while excluding the subjectivity of an analyst.

In addition, according to various embodiments of the present specification, it is possible to effectively predict a variable sensitive to a change in population distribution by using a microscopic fiscal estimate model, and accordingly, a fiscal estimate with higher accuracy can be derived.

In addition, according to various embodiments of the present specification, by providing relationship information between the funds, a pool between similar funds is formed, so that high profitability can be expected against the same risk.

1 is a block diagram showing the configuration of a system for deriving a financial estimate according to an embodiment of the present specification.
2 is a block diagram showing the configuration of an apparatus for deriving a financial estimate according to an embodiment of the present specification.
3 is a flowchart illustrating a method of deriving a financial estimate according to an embodiment of the present specification.
4 is a diagram illustrating an example of a method of extracting a principal component using a macroscopic financial estimation model according to an embodiment of the present specification.
5A and 5B are diagrams illustrating an example of extracting a main component using a macroscopic financial estimation model for a housing and urban fund according to an embodiment of the present specification.
6 is a diagram illustrating an example of predicting the number of housing subscription subscribers using a microscopic financial estimation model for a housing and city fund according to an embodiment of the present specification.
7A and 7B are diagrams illustrating an example of deriving a financial estimate for a retirement pension fund according to an embodiment of the present specification.
8 is a view showing an example of clustering several funds in consideration of the characteristics of the funds according to an embodiment of the present specification.

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

Throughout the specification, when a part is said to be "connected" with another part, this includes not only "directly connected" but also "electrically connected" with another element interposed therebetween. . In addition, when a part "includes" a certain component, it means that other components may be further included rather than excluding other components unless specifically stated to the contrary.

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

In this specification, some of the operations or functions described as being performed by the terminal or device may be performed instead in a server connected to the terminal or device. Likewise, some of the operations or functions described as being performed by the server may also be performed by a terminal or device connected to the server.

Hereinafter, with reference to the accompanying configuration diagram or processing flow chart, it will be described in detail for the implementation of the present invention.

Hereinafter, the term "fund" and the term "pension fund" may be used interchangeably.

1 is a block diagram showing the configuration of a system 1 for deriving a financial estimate according to an embodiment of the present specification. Referring to FIG. 1, a system 1 for deriving a fiscal estimate according to an embodiment may include an apparatus 100 for deriving a fiscal estimate, a database 200 and a terminal 300. The apparatus 100 for deriving a financial estimate may communicate with the database 200 and the terminal 300 through a wired or wireless network. In FIG. 1, the database 200 is shown separately from the apparatus 100 for deriving a financial estimate, but may be included in the apparatus 100 for deriving a financial estimate.

The device 100 for deriving a financial estimate receives base data used to derive a financial estimate for each fund from the database 200 or the terminal 300, and estimates a financial estimate for the fund subject to financial estimate based on the received base data. It is possible to draw a proposal and provide a financial estimate for the fund subject to fiscal estimates.

The financial estimation plan derivation device 100 receives base data from the database 200 or the terminal 300 at a predetermined period or in real time, updates the base data with the latest data, and finances using the updated base data. It is possible to derive a financial estimate of the fund to be estimated.

The underlying data used to derive the fiscal estimate includes the income and expenditure elements of the fund for which the financial projection is to be estimated, and may further include statistical records to be used to analyze each income and expenditure element. For example, the income elements of the Housing and Urban Fund include national housing bonds and housing subscription savings, and expenditure elements include public housing projects, urban regeneration projects, and loan support related to housing purchases and rentals for vulnerable social groups. The issuance amount of national housing bonds changes according to the status of real estate transactions, and the number of subscribers to the housing subscription savings may change depending on the housing economy. Therefore, the base data on the Housing and Urban Fund includes the income and expenditure elements of the Housing and Urban Fund, and furthermore, as a statistical record that can analyze the income and expenditure elements, various statistical records related to the status of real estate transactions, housing It can include a variety of statistics about the game.

The financial estimation plan derivation apparatus 100 receives the first base data from the database 200, receives the second base data from the terminal 300, and uses the received first base data and the second base data. A financial estimate for the fund to be estimated can be drawn.

When receiving user account information including an ID and a password, the device 100 for deriving a financial estimate, identifies a user corresponding to the received user account information, determines a fund for financial estimation based on the identified user, and determines Financial estimates derived for funds subject to fiscal estimates can be provided. In this case, a fund for which an advisory contract has been concluded may be registered for each user, and the financial estimate derivation apparatus 100 may determine a fund registered for the identified user as a fund for financial estimation.

The apparatus 100 for deriving a financial estimate may provide data obtained by analyzing a relationship between business details and asset management details and various characteristics (eg, GDP growth rate, exchange rate, real estate price index, etc.) for each fund. In addition, the apparatus 100 for deriving a financial estimate may provide relationship information between the funds based on the characteristics of each fund. For example, the apparatus 100 for deriving a financial estimate may cluster funds based on characteristics, and may provide information on funds with high similarity or information on funds in conflicting relationships based on the clustering result. . In the past, each fund was classified on the basis of business, financial, and social insurance, but there was a limitation in that the characteristics of each fund were not clearly reflected in this classification.

On the contrary, in the present invention, funds having similar characteristics are pooled and operated, so that a high rate of return can be expected against the same risk.

The database 200 may store base data used to derive a financial estimate. That is, the database 200 can store information on each fund, portfolio information of each fund, information on income and expenditure elements of each fund, and statistical records to be used for analysis of income and expenditure elements of each fund. have.

The terminal 300 may refer to a device used by a remote user to remotely access the device 100 for deriving a financial estimate. The terminal 300 may access the device 100 for deriving a financial estimate based on wired/wireless communication technology to control the operation of the device 100 for deriving a financial estimate and receive information on the derived financial estimate. .

When the base data necessary for the process of deriving the financial estimate is input, the terminal 300 may transmit the input base data to the device 100 for deriving the financial estimate. In this case, the device 100 for deriving a financial estimate may function as a server, and the terminal 300 may function as a client, and the device 100 for deriving a financial estimate may output the derived financial estimate to the terminal 300. I can.

2 is a block diagram showing the configuration of an apparatus 100 for deriving a financial estimate according to an embodiment of the present specification. Referring to FIG. 2, the apparatus 100 for deriving a financial estimate according to an embodiment may include a processor 110, a communication unit 120, a memory 130, an input unit 140, and an output unit 150. .

The processor 110 controls the communication unit 120, the memory 130, the input unit 140, and the output unit 150 to derive a financial estimate for each fund or perform an operation required to derive a relationship between each fund. You can do it.

The processor 110 derives a fiscal estimate for the fund subject to fiscal estimate using a macro fiscal estimate model and a micro fiscal estimate model, and outputs the derived fiscal estimate through the output unit 150 or through the communication unit 120. It may be transmitted to the terminal 300 and provided to the user.

The communication unit 120 may transmit and receive data to and from the database 200 or the terminal 300 using wired or wireless communication technology. For example, the communication unit 120 may receive base data necessary for deriving a financial estimate of a fund to be financially estimated from the database 200 or the terminal 300.

Communication technologies used by the communication unit 120 include Global System for Mobile communication (GSM), Code Division Multi Access (CDMA), Long Term Evolution (LTE), 5G, Wireless LAN (WLAN), Wireless-Fidelity (Wi-Fi), There are Bluetooth, Radio Frequency Identification (RFID), Infrared Data Association (IrDA), ZigBee, and Near Field Communication (NFC).

The memory 130 may store information on each fund, base data corresponding to each fund, user identification information, a fund analysis program, and the like.

The memory 130 may include not only an auxiliary storage device such as a hard disk drive or a solid state drive (SSD), but also a main storage device such as a random access memory (RAM).

The input unit 140 receives user identification information or fund data to be used to derive a financial estimate from a fund manager or a user.

The input unit 140 may include a mechanical input means (eg, a physical button) or a touch input means. The touch input means may include a virtual button displayed on a display implemented as an input device and an output device, a touch screen, or a touch button that is a physical button and interacts with a touch.

The output unit 150 may include at least one of the speaker 151 and the display 152. The output unit 150 may provide sound feedback to a user through the speaker 151, and the provided sound feedback may include synthetic voice, beep sound, music, and the like. Alternatively, the output unit 150 may output a financial estimate derived for the fund to be subject to financial estimate or relationship information between the funds through the display 152.

The display 152 may display information processed by the apparatus 100 for deriving a financial estimate. For example, the display 152 may display execution screen information of an application program driven by the device 100 for deriving a financial estimate, a user interface (UI), a graphical user interface (GUI), and the like according to the execution screen information.

3 is a flowchart illustrating a method of deriving a financial estimate according to an embodiment of the present specification.

Referring to FIG. 3, the processor 110 of the apparatus 100 for deriving a financial estimate receives user identification information through the communication unit 120 or the input unit 140 (S301).

The processor 110 may receive user identification information from the terminal 300 through the communication unit 120 or may receive user identification information through the input unit 140.

User identification information may include a user ID and a password corresponding to the user ID. A fund manager may be included in a user who uses the device 100 for deriving a financial estimate. Since the fund to be subject to financial estimation may be different for each fund manager, the apparatus 100 for deriving a financial estimation plan may provide a financial estimation service classified for each user by creating and managing accounts classified for each user.

Next, the processor 110 of the apparatus 100 for deriving a financial estimate determines a fund to be subject to financial estimate based on the received user identification information or a user input (S303).

Funds subject to fiscal estimates may mean funds subject to fiscal estimates. The user's account information may include setting information on a fund to be subject to financial estimation, and the processor 110 may determine a fund to be subject to financial estimation based on user account information corresponding to the received user identification information. Alternatively, the processor 110 may determine a fund other than the fund set in the existing user account information as a fund for financial estimation based on the user's input input through the communication unit 120 or the input unit 140. When the new fund is determined as the fund for financial estimation, the processor 110 may add setting information for the new fund for financial estimation to the user account information.

Next, the processor 110 of the apparatus 100 for deriving a financial estimate receives base data corresponding to the fund to be estimated from the database 200 or the terminal 300 through the communication unit 120 (S305).

Base data corresponding to the fund subject to fiscal estimates is data necessary to derive a fiscal estimate for the fund subject to fiscal estimates, and may include cumulative data on related variables or related factors that affect the income or expenditure of the fund subject to fiscal estimates. . For example, if the fund for financial estimation is the Employment Insurance Fund, the underlying data include the overall unemployment rate, the unemployment rate for each region, the total GDP growth rate, the GDP growth rate for each region, the total consumer price index, the consumer price index for each region, the stock price index, and the exchange rate as related variables. Accumulated data for the etc. may be included. Alternatively, if the target fund for financial estimation is the Housing City Fund, the base data includes the total real estate sales price index, real estate sales price index for each region, total real estate charter price index, real estate charter price index for each region, total real estate transaction volume, and each Real estate transaction volume by region, real estate transaction volume by type, total area unit real estate transaction volume, area unit real estate transaction volume by region, mortgage loan interest rate, average loan interest rate, real estate supply and demand trend index, subscription savings balance, subscription income, subscription principal repayment, number of subscription accounts, etc. Accumulated data for may be included.

In addition, the base data may include statistical data related to a fund subject to financial estimates used in a microscopic fiscal estimation model and statistics on population distribution. Specifically, the base data may include statistical data of population distribution on related variables of the fund subject to financial estimation. For example, when the fund subject to financial estimation is a retirement pension fund, the base data may include statistical data of subscribers of retirement pension insurance, statistical data of recipients of retirement pension insurance, statistical data of survivors benefit recipients, and the like. Likewise, when the fund to be subject to financial estimation is the Housing City Fund, the base data may include statistical data of housing subscription subscribers.

Next, the processor 110 of the fiscal estimate derivation device 100 extracts at least one principal component from related variables included in the base data using a macroscopic fiscal estimate model based on principal component analysis (PCA). (S307).

The macroscopic fiscal estimation model is a model that performs a macroscopic analysis of the fiscal estimating fund by extracting major components from the relevant variables of the fiscal estimating fund using PCA. PCA is a principal component analysis that can be used to reduce the dimension of variables in situations where various variables are mixed.

There are feature selection techniques and feature extraction techniques to reduce the dimensions by reducing variables, and the PCA technique belongs to the unsupervised feature extraction technique. In the variable selection technique, there is a great deal of room for human subjectivity to be involved in the process of selecting a variable, and there is a problem that the dimension of the curse occurs as the dimension is maintained as much as the number of selected variables. The curse of dimensionality can mean inefficiency of operation or learning that occurs when the dimension increases while using multivariate data to explain and estimate a phenomenon. When there are many selected variables, data visualization is difficult, redundant or unnecessary data may exist, and inefficiency may occur in modeling, making calculation difficult and increasing complexity, resulting in a large waste of computing power.

If only some of the variables are selected by using the variable selection technique, there is a correlation between the variables, and multicollinearity may occur, and it is not possible to know which variable is an important variable. Therefore, in this specification, a variable extraction technique is used. For example, there is a strong correlation between GDP growth rate and unemployment rate, or between real estate transaction volume and real estate price, so if the variable selection method is used, there is a high possibility that an appropriate analysis will not be derived.

The PCA technique is a technique that preserves the variance of data as much as possible, finds a new basis (axis) orthogonal to each other, and transforms samples from a high-dimensional space into a low-dimensional space with no linear association. For example, as shown in Equation 1 below, when the PCA technique is used, k variables (or components) Z that are not correlated with each other can be extracted from data consisting of p variables X (here, p>k). In Equation 1 below, X is the original variable, a is the loading, and Z is the extracted main component. The k extracted variables may be expressed as a linear combination of p existing variables. In the case of using the PCA technique, there is an advantage in that the dimension can be effectively reduced while preserving information (or variability) between components.

The processor 110 normalizes the related variables included in the base data corresponding to the fund for financial estimation, extracts a plurality of principal components by applying the PCA technique to the normalized related variable, and extracts a plurality of principal components from the extracted principal components. It is possible to select those whose relevance, explained variance to or greater than a predetermined reference value. For example, the processor 110 may extract 10 principal components from the normalized related variable, and may select only the principal components whose explanatory power for the fund subject to financial estimation is 0.1 or more from among the 10 principal components. If a variable is extracted without normalizing the related variable, since each variable has a different scale and thus distorted variable extraction may occur, the processor 110 will perform the normalization of the related variable. Variables can be extracted.

Alternatively, the processor 110 normalizes the related variable included in the base data corresponding to the fund for financial estimation, extracts a plurality of principal components by applying the PCA technique to the normalized related variable, and applies the extracted principal components to the financial estimation target. It is possible to sort in the order in which the relevance or explanatory power of change to the fund is high, accumulate and sum the relevance or explanatory power of change in the sorted order, and select the principal components up to the point where the cumulative sum exceeds a predetermined reference value.

For example, the processor 110 extracts 10 principal components from the normalized related variable, arranges the 10 principal components in the order of high variance explanatory power, and the first principal component with the highest variance explanatory power to the fourth with the fourth highest variance explanatory power. When the cumulative sum of the explanatory power to the main component exceeds a predetermined reference value of 0.9, only the first to fourth main components may be selected. In this case, from the fifth main component, the relevance or explanatory power of changes to the fund subject to financial estimation is rapidly weakened, and the fourth main component may be an elbow point.

The selected principal components are the principal components that have a high correlation with the volatility of the fund subject to fiscal estimation, and the non-selected principal components are the principal components that have a low correlation with the volatility of the fund subject to fiscal estimation.

Here, the processor 110 may select a related variable that is the subject of the principal component extraction from among the base data received based on a predetermined set value for each fund to be subject to financial estimation or an input from a user.

Next, the processor 110 of the fiscal estimate derivation apparatus 100 predicts related variables in consideration of population characteristics by using a microscopic fiscal estimating model based on a Markov Chain (S309).

The processor 110 may analyze microscopic factors for a fund subject to financial estimation in consideration of population distribution characteristics based on a Markov chain. Markov chain refers to a probability model applied when the state space is finite or countable during the Markov process, and the state at a specific point in time only affects the state at least one or more past points in time before the specific point in time. There is a characteristic of receiving. For example, in the Markov chain, the state of the current (t) time point can only be affected by the state from the past (t-1) time to the past (tn) time point. Likewise, the state of the current (t) point in time in the Markov chain can only be affected by the state of the previous (t-1) point in time. Equation 2 below represents the probability of a Markov chain in which the state of a specific point of time is affected only by the state of the immediately preceding point of time, and X _t denotes the state of point t.

Equation 3 below _{represents the probability that X t +1 is state j when X t} is state i, and a transition probability matrix P as shown in Equation 4 below can be obtained using this.

When using the Markov chain, it is possible to examine how population redistribution occurs through the chain action of population movement, and accordingly, it is possible to predict the population distribution or the amount of population movement in the future. The micro-fiscal estimating model using Markov Chain overcomes the shortcomings of the macroscopic approach of predicting simple trends through accumulated historical data, and predicts changes in the state of each individual over time, allowing precise prediction of the characteristics of the fund. have. That is, the apparatus 100 for deriving a financial estimate may predict a change in a related variable in consideration of population characteristics by using a microscopic financial estimate model, and accordingly, may derive a more accurate financial estimate. The Markov Chain-based microfiscal estimating model can be applied to funds such as Mutual Aid Association, Retirement Pension Fund, Workers' Compensation Insurance Fund, and Employment Insurance Fund, or funds that do not have a large amount of historical data, where changes in the population structure have a significant impact on the estimate results. There is an advantage to be able to.

The processor 110 may classify the data necessary for the micro-financial estimation model among the base data into basic statistical data and statistical assumption data, calculate the transition probability of the Markov chain using the statistical assumption data, and calculate the basic statistical data and One transition probability can be used to predict the relevant variable. In addition, the statistical assumption data may include demographic assumption data and economic assumption data.

For example, if the fund subject to financial estimation is a retirement pension fund, basic statistical data may include insurance premium subscriber information, retirement pension recipient information, survivor benefits recipient information, etc., and demographic assumption data includes the size of insurance premium payers, retirement rates by industry type, and Survival rate, general mortality rate by industry, (re)employment rate by industry, lump sum/pension selection probability, etc. may be included, and economic assumption data may include wage increase rate, operating rate of return, and base rate. Here, demographic assumption data affects the future development of retirement pension subscribers (death, retirement status, etc.), and economic assumption data affects the amount of retirement pension payments or the size of the retirement pension fund. In this way, the processor 110 may predict the size of the person subject to retirement pension premium payment, the size of the retirement pension premium payment amount, the size of the person subject to retirement pension benefits, the size of the retirement pension benefit amount, and the like using the Markov chain. In this case, statistical assumption data can be obtained from the Korea Labor Welfare Corporation's subscriber information data or related reports. Further, the processor 110 calculates and applies a demand probability value for the subscriber size by industry/gender/age group, wage increase rate, normal termination rate, and intermediate termination rate, etc., through past subscriber data, prior to financial estimation using a micro-financial estimation model. I can.

Next, the processor 110 of the fiscal estimate derivation apparatus 100 derives a fiscal estimate using the macro fiscal estimate results and the micro fiscal estimate results (S311).

The macro-fiscal estimation result may mean the principal component extracted using the macro-fiscal estimation model, and the micro-fiscal estimation result can mean the predicted value of the related variable. Accordingly, the processor 110 may predict the principal components extracted through the predicted values of the related variables, and derive a financial estimate by predicting the fund subject to financial estimation based on the predicted principal components.

Furthermore, the processor 110 may derive a fiscal estimate using a macro fiscal estimate result and a micro fiscal estimate result for each of a plurality of predetermined scenarios. To this end, the processor 110 may convert the impact (or change) of related variables for each past policy into data and reflect this in various scenario tests.

At least some of the steps illustrated in FIG. 3 may be performed in parallel or may be performed in a different order.

4 is a diagram illustrating an example of a method of extracting a principal component using a macroscopic financial estimation model according to an embodiment of the present specification.

Referring to FIG. 4, the processor 110 of the apparatus 100 for deriving a financial estimate may extract a plurality of principal components (PCs) from base data using a macroscopic financial estimate model based on PCA extraction. The processor 110 may extract the principal components in the order of high explained variance by using the macroscopic financial estimation model. The processor 110 may calculate cumulative explained variance with respect to the extracted main components, and select only the main components up to a point in which the calculated accumulated variance explanatory power exceeds a threshold 411.

In the example of FIG. 4, the variance explanatory power of the first main component PC1 is about 0.42, the variance explanatory power of the second main component PC2 is about 0.36, and the variance explanatory power of the third main component PC3 is about 0.17, and the fourth The variance explanatory power of the main component (PC4) is about 0.04. When the cumulative variance explanatory power is calculated in the order of high variance explanatory power, the first cumulative variance explanatory power is about 0.43 (or about 43%), the second cumulative variance explanatory power is about 0.78, and the third cumulative variance explanatory power is about 0.95, The fourth cumulative variance explanatory power is about 0.99. If the threshold value 411 for the cumulative variance explanatory power is 0.9, since the third cumulative variance explanatory power first exceeded the threshold value 411 and 0.9, the processor 110 Only the main component (PC2) and the third main component (PC3) can be selected. In particular, the variance explanatory power of the third main component PC3 is about 0.17, but since the variance explanatory power is less than 0.1 from the fourth main component PC4, the increase in the cumulative variance explanatory power is greatly reduced starting from the third cumulative variance explanatory power. Accordingly, the third main component PC3 becomes the elbow point 421 for the cumulative variance explanatory power.

5A and 5B are diagrams illustrating an example of extracting a main component using a macroscopic financial estimation model for a housing and urban fund according to an embodiment of the present specification.

The Housing City Fund is the main source of financial income and expenditure through issuance of national housing bonds and payments for housing subscriptions. National housing bonds are additive-fired bonds based on housing transactions, and have a structure in which income increases when the housing economy is booming, and income decreases as housing transaction volume decreases when the housing economy is in recession. In addition, if the amount of new issuance is low compared to the repayment amount due to maturity of the previously issued bonds, expenditures are incurred from the Housing and City Fund, and if the issuance amount is higher, income is generated from the Housing and City Fund.

5A and 5B, when the target fund for financial estimation is the housing city fund, the processor 110 of the financial estimation plan derivation device 100 is a factor affecting the housing city fund, and the subscription saving balance 520_1, Subscription income (520_2), subscription principal repayment amount (520_3), subscription accounts (520_4), interest rate difference (520_5), national real estate sales price (520_6), Seoul real estate sales price (520_7), nationwide area unit transaction volume (520_8), The main component can be extracted from the Seoul area unit transaction volume (520_9), the Seoul apartment area unit transaction volume (520_10), the nationwide real estate charter price (520_11), and the Seoul real estate charter price (520_12).

The processor 110 of the device 100 for deriving a financial estimate may extract a plurality of main components (511 to 516, etc.) from a plurality of factors related to the housing and urban fund, and each of the extracted main components (511 to 516, etc.) It can be expressed as a weighted sum of factors.

5A and 5B, the first main component 511 is a main component that has a high weight for real estate area unit transaction volume, has a variance explanatory power of about 0.2610, and has a nationwide area unit transaction volume (520_8), and a Seoul area unit transaction volume (520_9). ) And Seoul apartment area unit transaction volume (520_10) appear to be relatively high compared to other factors.

The second main component 512 is a main component in which the weight of the real estate sales price and real estate charter price is set high, and has a variance explanatory power of about 0.2309, and the national real estate sales price (520_6), the Seoul real estate sales price (520_7), and the national real estate charter price The weights for (520_11) and Seoul real estate charter price (520_12) appear relatively higher than those of other factors.

The third principal component 513 is a principal component whose weight for the subscription savings-related index is set high, has a variance explanatory power of about 0.1961, and has a weight for the subscription savings balance (520_1), subscription income (520_2), and the number of subscription accounts (520_4). It appears relatively high compared to other factors.

Accordingly, the first main component 511 can be viewed as a main component representing real estate transaction volume, the second principal component 512 can be viewed as a main component representing real estate price, and the third main component 513 is It can be seen as the main component.

In addition, the first principal component 511 has a negative correlation with the real estate transaction volume, the second principal component 512 has a positive correlation with the real estate sale price and the real estate jeonse price, and the third principal component 513 is a subscription. It can be seen that there is a positive correlation with the indicators related to saving. If the income or expenditure of the housing city fund changes due to the first main component 511, the change factor of the housing city fund can be regarded as the real estate transaction volume.

As a result of the principal component extraction, when the variance explanatory power for the first principal component 511, the second principal component 512, and the third principal component 513 is accumulated, the cumulative variance explanatory power is about 68.8%. On the other hand, from the fourth main component 514, it can be confirmed that the variance explanatory power is less than 0.1, and the influence on the housing and city fund as a fund subject to financial estimation is significantly lower than that of the first to third main components 511 to 513. 3 The main component 513 can be regarded as an elbow point.

As a result, it can be seen that the real estate transaction volume has a greater influence on the housing city fund than the real estate transaction price, and it can be seen that the real estate transaction volume in Seoul has a greater effect on the housing city fund than the real estate transaction volume in other regions.

In one embodiment, the processor 110 of the fiscal estimate derivation apparatus 100 extracts a plurality of main components (511 to 516, etc.) from the factors (520_1 to 520_12) related to the fund to be estimated financially, and extracts the main components (511). To 516, etc.), only the main components 511, 512, and 513 having a variance explanatory power exceeding a preset reference value of 0.1 may be selected.

6 is a diagram illustrating an example of predicting the number of housing subscription subscribers using a microscopic financial estimation model for a housing and city fund according to an embodiment of the present specification.

Housing and City Fund is one of the main sources of income from housing subscription payments, and housing subscription payments are the closest to the population structure. This is because the majority of subscription savings subscribers do not use subscription savings several times in their life cycle. Therefore, related variables closely related to changes in the population structure, such as the number of subscription savings subscribers, can be effectively predicted using a microscopic fiscal estimation model.

Referring to FIG. 6, for example, a male in his twenties residing in Seoul who has not subscribed to a housing subscription has a 30% probability of signing up for a housing subscription the following year, and an 80% probability of maintaining the subscription for one year after signing up. Assume %. In this case, if there are 100 males in their 20s residing in Seoul who have not subscribed to the housing subscription in N year (610), 30 people subscribe to the housing subscription in (N+1) year (620), and 70 still subscribe to housing. You can predict that you will not subscribe to. And, in (N+2) year (630), out of 30 people who signed up for housing in (N+1) year (620), it is predicted that 24 will still maintain the housing subscription and 6 will cancel the housing subscription I can.

Although not shown in Fig. 6, 70 men in their 20s residing in Seoul who did not subscribe to a housing subscription in (N+1) year (620), 30% or 21 people who lived in Seoul in (N+2) year (630). It can be predicted that 49 people, 70% of whom have subscribed to the subscription, will still not subscribe to the housing subscription.

As described above, the number of housing subscription subscribers can be predicted in consideration of population distribution characteristics using a microscopic financial estimation model, and the predicted number of housing subscription subscribers can be used to derive a financial estimate.

7A and 7B are diagrams illustrating an example of deriving a financial estimate for a retirement pension fund according to an embodiment of the present specification. Specifically, FIG. 7A shows the distribution of subscribers and terminators by year of retirement pensions for business employees of less than 50 employees, predicted by using a microscopic financial estimation model, in order to derive a financial estimate for the retirement pension fund. In addition, FIG. 7B shows a financial estimate derived by reflecting the results of macroscopic and microscopic financial estimates analyzed for subscribers of businesses with less than 50 employees for the retirement pension fund.

As a result of fiscal estimates, the size of the fiscal balance for employees of businesses with fewer than 50 of the retirement pension subscribers in 2029 is about 1.5 trillion won at the median level, and the total retirement pension fund is estimated to be about 7.3 trillion won.

8 is a view showing an example of clustering several funds in consideration of the characteristics of the funds according to an embodiment of the present specification. Referring to FIG. 8, the apparatus 100 for deriving a financial estimate may extract a plurality of main components for a fund to be subject to financial estimates using a macroscopic financial estimate model, and the extracted main ingredients represent characteristics of the fund to be subject to financial estimates. . Accordingly, the apparatus 100 for deriving a financial estimate may map each fund to a feature space in consideration of the characteristics of the fund, and map a plurality of funds to a plurality of clusters based on the proximity in the feature space. ) Can be clustered. Accordingly, the apparatus 100 for deriving a financial estimate may provide information on the similarity or correlation between various funds to a user.

An embodiment of the present invention may also be implemented in the form of a recording medium including instructions executable by a computer, such as a program module 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. Further, the computer-readable medium may include all computer storage media. 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 illustrative purposes only, and those of ordinary skill in the art to which the present invention pertains will be able to understand that other specific forms can be easily modified 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 and non-limiting in all respects. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as being distributed may also be implemented in a combined form.

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

100: device for deriving financial estimates
110: processor 120: communication unit
130: memory 140: input
150: output
151: speaker 152: display
200: database 300: terminal

Claims (17)

In the device for deriving a financial estimate,
A communication unit that communicates with a database or a terminal;
Input unit; And
Receives user identification information input through the terminal or the input unit through the communication unit, determines a fund for financial estimation based on the user identification information or a user input through the input unit, and responds to the financial estimation target fund Receive the underlying data, extract at least one principal component from the first related variable included in the base data using a macroscopic financial estimation model, and predict a second related variable considering population characteristics using a microscopic financial estimation model And a processor for deriving a financial estimate using the extracted principal component and the predicted second related variable
Including,
The processor selects a principal component in which one of the relevance, variation explanatory power, and variance explanatory power to the fund to be estimated from among the extracted at least one principal component is equal to or greater than a preset threshold, and
The processor divides the base data necessary for the micro-financial estimation model among the base data into basic statistical data and statistical assumption data, and calculates a transition probability of a Markov Chain based on the statistical assumption data, An apparatus for predicting the second related variable based on the calculated transition probability and the basic statistical data. The method of claim 1,
The processor is
An apparatus for extracting the at least one principal component from the first related variable by using the macroscopic financial estimation model based on principal component analysis (PCA). The method of claim 2,
The processor is
The first related variable is normalized, and a plurality of principal components are extracted in the order of high variance explanatory power for the fund subject to financial estimation from the normalized first related variable, and variance explanatory power for the fund subject to financial estimation for each of the principal components A device for deriving a financial estimate for selecting some of the plurality of main components in consideration of. The method of claim 3,
The processor is
A device for deriving a financial estimate, selecting a main component whose variance explanatory power exceeds a first reference value from among the plurality of main components. The method of claim 3,
The processor is
The apparatus for deriving a financial estimate, calculating the cumulative variance explanatory power in the order of the high variance explanatory power, and selecting a principal component from the plurality of principal components until the cumulative variance explanatory power first exceeds a second reference value. The method of claim 3,
The processor is
An apparatus for deriving a fiscal estimate by predicting the second related variable in consideration of the population characteristics using the Markov chain-based micro fiscal estimating model, in which a state at a specific point in time is determined based on a state at a point in time immediately before. The method of claim 6,
The processor is
An apparatus for deriving a financial estimate for analyzing the similarity between the plurality of funds in consideration of the calculated financial estimate for a plurality of funds. The method of claim 7,
The processor is
Mapping the plurality of funds to a feature space in consideration of the characteristics of each of the plurality of funds, and clustering the plurality of funds based on an adjacent degree in the feature space. In the fiscal estimate derivation method performed by the fiscal estimate derivation device,
Receiving user identification information;
Determining a fund for financial estimation based on the user identification information or a user's input;
Receiving base data corresponding to the fund subject to financial estimation;
Extracting at least one principal component from a first related variable included in the base data using a macroscopic financial estimation model;
Predicting a second related variable in consideration of population characteristics using a microscopic financial estimation model; And
Deriving a financial estimate using the extracted principal component and the predicted second related variable
Including,
The step of extracting the at least one or more active ingredients
Including the step of selecting a principal component from among the extracted at least one or more principal components in which one of the relevance, variation explanatory power, and variance explanatory power with respect to the financial estimation target fund is equal to or greater than a preset threshold,
Predicting the second related variable comprises:
Among the base data, the base data required for the microscopic financial estimation model is classified into basic statistical data and statistical assumption data, and a Markov chain transition probability is calculated based on the statistical assumption data, and the calculated transition probability and the And predicting the second related variable based on basic statistical data. The method of claim 9,
The step of extracting the at least one or more active ingredients
A method of extracting the at least one principal component from the first related variable by using the macroscopic financial estimation model based on principal component analysis. The method of claim 10,
The step of extracting the at least one or more active ingredients
Normalizing the first related variable, and extracting a plurality of principal components from the normalized first related variable in an order of high variance explanatory power for the fund subject to financial estimation; And
Selecting some of the plurality of principal components in consideration of the variance explanatory power of each of the principal components for the fund subject to financial estimation
Including, a method of deriving a financial estimate. The method of claim 11,
Selecting some of the plurality of main components
A method for deriving a financial estimate, wherein a main component whose variance explanatory power exceeds a first reference value is selected from among the plurality of main components. The method of claim 11,
Selecting some of the plurality of main components
A method of deriving a financial estimate, wherein the cumulative variance explanatory power is calculated in the order of the high variance explanatory power, and a principal component until the cumulative variance explanatory power first exceeds a second reference value is selected from among the plurality of principal components. The method of claim 11,
Predicting the related variable
A method for deriving a fiscal estimate by predicting the second related variable in consideration of the population characteristics using the Markov chain-based micro fiscal estimating model in which a state at a specific point in time is determined based on a state at a point in time immediately before. The method of claim 14,
Analyzing the degree of similarity between the plurality of funds in consideration of the calculated financial estimates for the plurality of funds.
A method for deriving a financial estimate further comprising a. The method of claim 15,
Analyzing the similarity between the plurality of funds
Mapping the plurality of funds to a feature space in consideration of characteristics of each of the plurality of funds; And
Clustering the plurality of funds based on the proximity in the feature space
Including, a method of deriving a financial estimate. In a computer program stored on a recording medium comprising a sequence of instructions for deriving a financial estimate,
When the computer program is executed by a computing device,
Receive user identification information,
Determine a fund for financial estimation based on the user identification information or the user's input,
Receive base data corresponding to the fund for the financial estimation,
Extracting at least one principal component from the first related variable included in the base data using a macroscopic financial estimation model,
Predicting the second related variable considering the characteristics of the population using a microscopic financial estimation model,
A sequence of instructions for deriving a financial estimate using the extracted principal component and the predicted second related variable,
The computer program selects a principal component in which one of the relevance, variation explanatory power, and variance explanatory power to the fund to be estimated from among the extracted at least one principal component is equal to or greater than a preset threshold,
Among the base data, the base data required for the microscopic financial estimation model is classified into basic statistical data and statistical assumption data, and a Markov chain transition probability is calculated based on the statistical assumption data, and the calculated transition probability and the A computer program stored on a recording medium comprising a sequence of instructions for predicting the second related variable based on basic statistical data.

Images