JP2023088541A - Program, information processor, and method - Google Patents

Abstract

To provide a program that calculates risk tolerance for a user for a financial product according to a change factor of the risk tolerance.SOLUTION: A program causes a computer to perform processing of: acquiring product information including attribute information of a financial product, similarity information indicating degree of similarity between a fluctuation trend of a financial benchmark referenced by the financial product and a price fluctuation trend of the financial product, and risk information of the financial product calculated on the basis of a price fluctuation of the financial product; assigning, on the basis of the similarity information and the risk information, classification information to the financial product; acquiring, on the basis of on the attribute information, text information associated with the financial product and evaluation information for the text information; causing, on the basis of the attribute information, the text information, and the evaluation information, a user to present a question on the classification information; acquiring preference information from the user that indicates preference of the user for the classification information; and calculating, on the basis of the preference information, risk tolerance of the user for the classification information.SELECTED DRAWING: Figure 4

Description

The present invention relates to a program, an information processing device, and a method.

When selling a financial product purchased for the purpose of asset management, etc., to a user, the risk tolerance of the user is acquired. Risk tolerance is a measure of how much risk (value fluctuations) inherent in each financial product can be tolerated. The risk tolerance is used to determine the management style for the user's asset management. Patent Literature 1 discloses a management style determination system that can change or add a management style according to changes in social conditions and demands.

Japanese Patent Application Laid-Open No. 2021-26408

Risk tolerance may change depending on the changing factors of the user's values, operational experience, and economic conditions. The management style determination system described in Patent Literature 1 acquires answers to questions about risk tolerance from users who have already acquired risk tolerance, and changes the risk tolerance based on the acquired answers. A management style is determined based on the modified risk tolerance. The content of the question asked by the user to obtain the answer is fixed text and does not change. In this case, the question is not asked according to the change factor of the risk tolerance, and it may be difficult to accurately calculate the user's risk tolerance.

SUMMARY OF THE INVENTION Accordingly, an object of the present invention is to provide a program, an information processing apparatus, and a method that enable calculation of a user's risk tolerance for financial products in accordance with a change factor of the risk tolerance.

A program according to an aspect of the present invention provides a computer with attribute information of a financial product, similarity information indicating the degree of similarity between the fluctuation trend of a financial benchmark referenced by the financial product and the price fluctuation trend of the financial product, and financial Product information acquisition processing for acquiring product information including financial product risk information calculated based on product price fluctuations, and classification information allocation for assigning classification information to financial products based on similarity information and risk information a text information acquisition process for acquiring text information associated with a financial product and evaluation information for the text information based on the attribute information; A question text presentation process for presenting a question text to a user, a preference information acquisition process for acquiring preference information indicating the user's preference for classification information from the user, and a user's risk tolerance for classification information calculated based on the preference information. and a risk tolerance calculation process to be executed.

In the above aspect, the product information acquisition process includes acquiring the first product information of the first financial product and the second product information of the second financial product as product information, and the classification information allocation process assigning first classification information to the financial instrument and assigning second classification information to the second financial instrument, wherein the textual information acquisition process is associated with the first financial instrument based on attribute information of the first financial instrument; Acquiring text information and evaluation information, and acquiring text information and evaluation information associated with the second financial product based on the attribute information of the second financial product; Preference information, including presenting to the user a first question about the first classification information based on the evaluation information, and presenting a second question about the second classification information to the user based on the text information and the evaluation information; The acquisition process includes acquiring from the user preference information indicating the user's preferences for the first classification information and the second classification information, and the risk tolerance calculation process calculates the first classification information and the second classification information based on the preference information. It may also include calculating the user's risk tolerance for each of the categories of information.

In the above aspect, the product information acquisition process includes acquiring the first product information of the first financial product and the second product information of the second financial product as product information, and the classification information allocation process including assigning classification information to the product and the second financial product, wherein the textual information acquisition processing includes textual information and evaluation information of the first financial product based on the attribute information of the first financial product and the attribute information of the second financial product; and acquiring the text information and evaluation information of the second financial product, and the question text presentation process is based on the attribute information, the text information and the evaluation information of the first financial product, and generates the first question text related to the classification information. is presented to the user, and based on the attribute information, text information, and evaluation information of the second financial product, the user is presented with a second question about the classification information, and the preference information acquisition process comprises the first question and The risk tolerance calculation process includes obtaining from the user preference information indicating the user's preferences for each of the second question sentences, and calculating the user's risk tolerance for financial instruments included in the classification information based on the preference information. It may include calculating.

In the above aspect, the question presentation process includes presenting a plurality of questions regarding the classification information to the user, and the risk tolerance calculation process calculates the risk tolerance based on the preference information obtained through the plurality of questions. may include calculating

In the above aspect, the computer may be caused to further execute a screen information generation process for generating information for a risk tolerance display screen that presents the user with the risk tolerance for the classification information based on the classification information and the risk tolerance. good.

In the above aspect, the computer may be caused to further execute candidate screen information generation processing for generating information for a product candidate screen that presents to the user financial product candidates included in the classification information that satisfies a predetermined condition in risk tolerance. good.

An information processing apparatus according to an aspect of the present invention includes attribute information of a financial product, similarity information indicating a degree of similarity between a fluctuation trend of a financial benchmark referenced by the financial product and a fluctuation trend of the price of the financial product, and A product information acquisition unit that acquires product information including risk information of financial products calculated based on price fluctuations of financial products, and a classification information allocation unit that allocates classification information to financial products based on similarity information and risk information a text information acquisition unit that acquires text information associated with a financial product and evaluation information for the text information based on the attribute information; and a question about the classification information based on the attribute information, the text information, and the evaluation information. A question sentence presenting unit that presents a sentence to the user, a preference information acquisition unit that acquires preference information indicating the user's preference for the classification information from the user, and a user's risk tolerance for the classification information based on the preference information. and a risk tolerance calculator.

A method according to an aspect of the present invention is a method in which a computer performs attribute information of a financial product, similarity information indicating the degree of similarity between the fluctuation trend of a financial benchmark referenced by the financial product and the fluctuation trend of the price of the financial product; Acquiring product information including risk information of financial products calculated based on product price fluctuations; assigning classification information to financial products based on similarity information and risk information; obtaining text information associated with the financial product and evaluation information for the text information based on the attribute information, text information, and evaluation information, and presenting a question about the classification information to the user based on the attribute information, the text information, and the evaluation information; , obtaining preference information from the user indicating the user's preference for the classification information; and calculating the user's risk tolerance for the classification information based on the preference information.

According to the present invention, it is possible to provide a program, an information processing device, and a method that enable calculation of a user's risk tolerance for financial products in accordance with factors that change the risk tolerance.

BRIEF DESCRIPTION OF THE DRAWINGS It is a figure which shows the outline of the risk-tolerance calculation system which concerns on this embodiment. It is a figure explaining a financial benchmark and a financial product to which a financial product refers. It is a figure explaining the procedure of question sentence generation in the risk-tolerance calculation system which concerns on this embodiment. It is a block diagram which shows the structure of the risk tolerance calculation apparatus which concerns on this embodiment. It is a figure which shows an example of the merchandise information which concerns on this embodiment. It is a figure which shows an example of the financial benchmark information which concerns on this embodiment. It is a figure which shows an example of the risk tolerance map which concerns on this embodiment. It is a figure which shows an example of the classification|category information for every financial product which concerns on this embodiment. It is an example of text information and evaluation information according to the present embodiment. It is an example of co-occurrence expression information according to the present embodiment. 4 is a flowchart of processing in the risk tolerance calculation device according to the present embodiment; It is a figure which shows an example of the financial product classified into the risk-tolerance map which concerns on this embodiment. It is a figure explaining an example of generation of a question sentence and risk tolerance calculation by a risk tolerance calculation device concerning this embodiment. It is a figure explaining an example of generation of a question sentence by the risk tolerance calculation device concerning this embodiment. It is a figure explaining another example of generation of a question sentence by the risk tolerance calculation device concerning this embodiment. It is a figure explaining an example of generation of the question sentence with respect to classification information by the risk-tolerance calculation apparatus which concerns on this embodiment. It is a figure which shows an example of the question sentence presentation screen which concerns on this embodiment. FIG. 4 is a diagram illustrating a procedure for selecting classification information for generating a question sentence by the risk tolerance calculation device according to the present embodiment; FIG. 4 is a diagram illustrating a procedure for selecting classification information for generating a question sentence by the risk tolerance calculation device according to the present embodiment; It is a figure which shows an example of the risk tolerance which concerns on this embodiment. It is a figure which shows an example of the risk tolerance display screen which concerns on this embodiment. It is a figure which shows an example of the goods candidate screen which concerns on this embodiment.

Preferred embodiments of the present invention will be described with reference to the accompanying drawings. It should be noted that, in each figure, the same reference numerals have the same or similar configurations.

FIG. 1 shows a schematic diagram of a risk tolerance calculation system 10 according to this embodiment. The risk tolerance calculation system 10 includes a risk tolerance calculation device 101, user terminals 102a, 102b . . . Note that the user terminals 102a to 102n are referred to as user terminals 102 when there is no particular need to distinguish between them.

The risk tolerance calculation device 101 is an information processing device, such as a server, that executes a predetermined program to perform processing to be described later and calculates a user's risk tolerance. The user terminal 102 is an information processing device such as a computer or a smart phone operated by a user who calculates his or her own risk tolerance. The text information providing device 103 is, for example, a server used by an operator who provides economic news to distribute news.

The risk tolerance calculation device 101 is connected to the user terminal 102 and the text information providing device 103 through the network N so as to be communicable.

With reference to FIG. 2, the respective risks of financial benchmarks and financial products and the degree of similarity between financial benchmarks and financial products according to this embodiment will be described. Here, a financial benchmark is an index calculated based on commodity prices in the financial market. For example, in a stock market, a financial benchmark is a value calculated based on the price of an issue selected from the issues listed on the market and the composition ratio of each issue. A financial benchmark is also calculated in the bond market according to the bond price and the composition ratio of the bonds that constitute the financial benchmark. Financial products for which financial benchmarks are calculated are not limited to stocks and bonds, and include, for example, financial products related to gold, oil, and agricultural products, financial products related to futures trading on underlying assets, and financial products related to crypto assets. Indices that are referenced during operation, such as ESG indices related to the environment and indices related to diversity, can also be used as financial benchmarks.

A financial benchmark is a value independently calculated by each institution that calculates the financial benchmark, and there are multiple types of financial benchmarks. A financial benchmark may be calculated, for example, as an average of stock prices of constituent stocks, or may be calculated as an index independently provided by an institution that calculates a financial benchmark.

In asset management, financial products purchased by users include, for example, investment trusts, ETFs (Exchange Traded Funds), and bonds. A financial benchmark is associated with these financial products, and their investment performance is measured according to price fluctuations relative to the financial benchmark.

FIG. 2 shows time fluctuations in values of financial benchmarks BMa and BMb and time fluctuations in prices of financial products FPa, FPb, FPc and FPd.

A risk is calculated for each of the financial benchmarks BMa, BMb and the financial instruments FPa, FPb, FPc, FPd. Financial benchmark risk is expressed as the standard deviation of the financial benchmark over a given period of time. Higher risk indicates greater volatility in the value of the financial benchmark. In FIG. 2, the financial benchmark BMa has a higher risk than the financial benchmark BMb. Similarly, for financial instruments, high risk indicates high fluctuations in the price of the financial instrument. In FIG. 2, the risks of financial instruments FPa and FPb are small compared to the risks of financial instruments FPc and FPd.

For financial products, we introduce the concept of similarity to financial benchmarks referred to by each financial product. The degree of similarity is information calculated based on the relationship between a financial benchmark and financial products that refer to the financial benchmark. The degree of similarity is information calculated using, for example, fluctuation trends of the financial benchmark and the financial product, component stocks of the financial benchmark and the financial product, and the like. The degree of similarity is information indicating how similar fluctuation trends, constituent stocks, etc. are.

The degree of similarity is calculated using, for example, dynamic time warping (DTW), which is one of techniques for calculating the degree of similarity of time-series data. In this case, the degree of similarity is calculated using time-series data of financial products and time-series data of financial benchmarks. For example, assume that financial products FPa and FPb are financial products that refer to a financial benchmark BMa. At this time, as shown in FIG. 2, the financial product FPa undergoes price fluctuations similar to those of the financial benchmark BMa with a time lag. Therefore, the degree of similarity between the financial product FPa and the financial benchmark BMa is high. Also, the financial product FPb undergoes price fluctuations different from those of the financial benchmark BMa. In this case, from the viewpoint of risk, the financial product FPb has a large risk like the financial benchmark BMa. become smaller.

Similarly, assume that the financial products FPc and FPd are financial products that refer to the financial benchmark BMb. In this case, the financial product FPc has a small risk like the financial benchmark BMb, but the prices fluctuate differently from each other, so the degree of similarity between the financial product FPc and the financial benchmark BMb decreases. The financial product FPd has a small risk like the financial product FPd, but its price fluctuations are somewhat similar to those of the financial benchmark BMb, so the degree of similarity is moderate.

The degree of similarity between a financial instrument and a financial benchmark is calculated based on the degree of overlap in constituents of the financial instrument and the financial benchmark, in addition to the degree of similarity based on the degree of similarity in risk and the degree of similarity based on time-series data. may In this case, the similarity is calculated using the Jaccard coefficient, the Dice coefficient, or the Simpson coefficient, which are scales indicating the similarity of sets. In this embodiment, as the similarity between the financial product and the financial benchmark, a similarity other than the similarity based on the degree of similarity of risks calculated by the above method is used.

Financial instruments can be classified according to the risk of each financial instrument and the degree of similarity to the financial benchmarks to which each financial instrument refers. For example, the financial product FPa in FIG. 2 is classified as a financial product with high risk and high similarity.

A procedure for generating a question sentence in the risk tolerance calculation system will be described with reference to FIG. In step (A), financial instruments and financial benchmarks are classified according to risk and similarity to financial benchmarks. As shown in FIG. 3, the financial instruments are classified according to their degree of similarity with the benchmark and their risks, corresponding to a risk tolerance map M having similarity with the benchmark on the horizontal axis and risk information on the vertical axis. be. Here, the risk information is information indicating risk quantiles based on risk values. In the classification of financial benchmarks, financial benchmarks are classified in order of risk in the leftmost column of the risk tolerance map M, with the degree of similarity to itself being the same (extremely high). There may be multiple financial instruments classified with common similarities and risks. That is, a given similarity and risk information includes multiple financial instruments classified as having that similarity and risk information.

In step (B), the financial product is associated with text information such as news based on information on the financial product, such as information indicating the type of business of the brand included in the financial product, information on the region of the brand, and the like. In step (C), the risk tolerance calculation device 101 calculates the risk tolerance for financial products classified by similarity to benchmarks and risks based on a certain financial product and text information associated with the financial product. A question sentence for doing is generated.

The question sentence is generated as a sentence containing several sentences and accompanied by options, as shown in FIG. 3, for example. The questions are presented to the user, and the user selects an option that matches his or her operational tendency. For example, in FIG. 3, for a certain theme X, a question sentence including options of "A: There are high expectations for utilization", "B: There are great concerns about regulations", and "C: I don't know" is presented. For example, when the user selects option A, the user has a favorable view of financial instruments classified into the similarity and risk, and has a higher risk tolerance for those financial instruments. is judged to be acceptable. In the case of option B, on the contrary, the person has an unfavorable view of financial products, and it is judged that risk tolerance should be lowered. In the case of option C, since the risk cannot be judged as a specific numerical value, it is judged that safety is taken and the risk tolerance is zero, that is, no risk is allowed.

By classifying financial products in this way and generating questions related to a certain classification, financial products belonging to a common Get a more comprehensive risk tolerance for your product. In addition, since the question sentence is generated based on text information such as news related to financial products that can be a factor of change in the user's risk tolerance, it is possible to obtain the risk tolerance by asking a question that reflects the factor of change.

Each part of the risk tolerance calculation device 101 will be described with reference to FIG. FIG. 4 shows a block diagram of the risk tolerance calculation device 101. As shown in FIG. The risk tolerance calculation device 101 includes a storage unit 401, a communication unit 402, a product information acquisition unit 403, a classification information allocation unit 404, a text information acquisition unit 405, a question presentation unit 406, a preference information acquisition unit 407, and a risk tolerance calculation. A screen information generation unit 409 and a candidate screen information generation unit 410 are provided.

Information processing in each part of the risk tolerance calculation device 101 is realized, for example, by using a storage area such as a memory in the risk tolerance calculation device 101 or executing a program stored in the storage area by a processor. can do.

The storage unit 401 stores various types of information used for processing in the risk tolerance calculation device 101 . The storage unit 401 has a financial product DB 4011, a benchmark DB 4012, and a question sentence generation information DB 4013.

As shown in FIG. 5, the financial product DB 4011 stores product information of financial products used for calculation of risk tolerance by the risk tolerance calculation device 101 .

The product information has items of "financial product name", "financial benchmark", "financial product ID", "financial benchmark ID", "attribute information", "similarity information", "risk", and "risk information". .

“Financial product name” is information indicating the name of the financial product. Financial products include, for example, “A investment trust (emerging countries)” (first financial product) and “B investment trust (United States)” (second financial product). There are financial products, such as “C investment trust (theme: semiconductors)” and “C investment trust (theme: automobiles)”, for which the industrial theme of the brand is provided. In addition, there are financial products such as “D investment trust (global stocks)” that do not limit countries or regions, and financial products that handle bonds as assets such as “E bond (country X)”. In addition, there are financial products such as "F investment trust (index type)" and "G investment trust (emerging country (index) type)" whose purpose is to fluctuate in price according to the benchmarks they refer to. The financial products are not limited to the above examples, and may include real estate investment REITs in addition to investment trusts and bonds. Also, financial instruments may include various types of financial instruments, as long as they have corresponding financial benchmarks. For example, financial products related to gold, oil, and agricultural products, financial products related to futures transactions targeting underlying assets, and financial products related to crypto assets may be included. It may also include financial instruments associated with, for example, ESG indices, indices related to diversity, and the like.

The item "financial benchmark" is information indicating a financial benchmark referred to by each financial product. Financial benchmarks may be common or different between financial instruments.

The items “financial product ID” and “financial benchmark ID” are information for identifying financial products and financial benchmarks.

"Attribute information" includes various information related to financial products. The attribute information has, for example, items of "asset class", "issue", "industry", "region", "country" and "theme". The item of "asset class" is information indicating the type of asset of financial products, and information such as stocks and bonds is recorded. The item "brand name" records information indicating the brand name of the stock in which the investment trust invests. For example, in the AA investment trust, it is recorded that issues such as "A001" and "A002" are investment targets. In the item of "industry", for example, when the investment target is limited to a certain industry of financial products, the industry to be invested in is recorded. For example, in the C investment trust, since the theme is semiconductors, the information "manufacturing" is recorded in the type of industry. The items of "Region" and "Country" store information indicating the region and country in which stocks, etc., in which financial products are invested are located. For example, for investment trust A, the region is recorded as "Asia" and the country as "Thailand, Malaysia, Indonesia, etc.". The item "theme" records information indicating an industrial theme in which financial products are invested. For example, in C investment trust, the theme is recorded as "semiconductor". As another example, the theme may be any theme such as automobiles or medical care. Although not shown, attribute information is also recorded for other financial products. Also, when a new financial product is added, attribute information corresponding to the added financial product is recorded. Also, the attribute information items themselves may be added as appropriate depending on the financial product.

The “similarity information” is information indicating the level of similarity according to the similarity between the financial product and the financial benchmark calculated based on the DTW distance, Jaccard coefficient, or the like. For example, the standardized similarity is divided into several categories, and similarity information for each financial product is associated according to the similarity between each financial product and each financial benchmark. The example of FIG. 5 shows a case where five similarity divisions (BM, 1, 2, 3, 4) are provided. When the similarity information is BM, it means that the similarity between the financial product and the financial benchmark is very high and the price of the financial product changes in the same manner as the benchmark (Bench Mark: BM). As the similarity decreases, the similarity information of "4" is associated with the similarity information of "1". The smaller the numerical value of the similarity information, the more different the financial product price changes from the financial benchmark.

The item "risk" is information indicating the risk of a financial product in a predetermined period. The predetermined period is, for example, one year. The item of "risk information" is information on the level of risk corresponding to the value of risk. Risk information has several categories. For example, when setting five categories, if the risk is denoted as r, "1" if 0% < r ≤ 3%, "2" if 3% < r ≤ 5%, 5% < r The risk information is classified as "3" when ≦7%, "4" when 7%<r≦10%, and "5" when 10%<r.

Each information recorded in "similarity information", "risk", and "risk information" may be calculated and recorded by the risk tolerance calculation device 101, and information processing different from the risk tolerance calculation device 101 may be performed. Information calculated by the device may be acquired and recorded by the risk tolerance calculation device 101 .

As shown in FIG. 6, the benchmark DB 4012 stores risk information associated with the risk of the financial benchmark so as to have the same classification as the risk of the financial benchmark referred to by each financial product in the product information and the product information. is recorded. Moreover, various information regarding each benchmark is recorded in the benchmark DB 4012 . For example, information that shows the constituent stocks of each benchmark and their composition ratio, the region covered by the benchmark, the industry ratio of the constituent stocks, or the factor ratio based on the value and growth judged for each constituent stock, etc. Information is recorded.

Based on the "similarity information" and "risk information" associated with the financial instrument, the financial instrument can be categorized into any area of the risk tolerance map M as shown in FIG. That is, the risk tolerance map M has a region specified by classification information=(similarity information, risk information). FIG. 7 shows an example in which both similarity information and risk information are divided into 5 categories. Note that the number of categories of similarity information and risk information can be increased or decreased as needed. In addition, the risk tolerance map M is provided with divisions in which the risk information is "0" without division with respect to the similarity information. The risk information “0” category is a category provided for deposits as broadly defined financial products when the risk tolerance is zero. In addition to being provided so as to be displayed integrally with the risk tolerance map M, the division in which the risk information is "0" is separated from the risk tolerance map M, for example, a separation having one square It may be provided as a map with

FIG. 8 shows a state in which classification information is assigned to each financial product based on the similarity information and risk information in FIG. The classification information is stored in the storage unit 401 in association with the financial product ID or the like. Financial instruments are associated with regions on the risk tolerance map M based on their classification information. For example, A investment trust, which is a financial product with a financial product ID “ASSET001”, has classification information (similarity information, risk information)=(BM, 5) (first classification information), so the risk tolerance map M is exactly Associated with the upper left region. In addition, the B investment trust, which is the financial product with the financial product ID “ASSET002”, has classification information (similarity information, risk information) = (BM, 2) (second classification information), so the left side of the risk tolerance map M Associated with the middle region.

The question text generation information DB 4013 records text information and evaluation information used for question text generation by the risk tolerance calculation device 101 . Information recorded in the question sentence generation information DB 4013 has items of "sentence information", "sentence ID", "keyword information", and "evaluation information".

The item “text information” is text information acquired by the risk tolerance calculation device 101 from the text information providing device 103 . The item "sentence ID" is information for identifying each sentence. The item of “keyword information” is, for example, information indicating keywords in sentences extracted by natural language processing by the sentence information acquisition unit 405, which will be described later.

The “evaluation information” item is information (sentiment information) generated by natural language processing by the text information acquisition unit 405 and indicating the economic evaluation of the text. Evaluation information has three values, for example, "positive", "negative", and "neutral". A deep learning model such as BERT (Bidirectional Encoder Representations from Transformers) pre-learned using learning data in which the input is text and the output is evaluation information can be used to generate the evaluation information. Positive evaluation information means that the text is economically positive news, and negative evaluation information means that the text is economically unfavorable news. Neutral rating information means that the sentence is not suitable for any rating.

For example, the text with the text ID “S001” is the text “Economic expansion trend in Southeast Asian countries”, and the keyword information is “Southeast Asia” and “economic expansion”. In this case, the evaluation information is "positive". On the other hand, the sentence with the sentence ID "S002" is a sentence "consumption in Malaysia tends to be restrained", and the keyword information is "Malaysia", "consumption" and "restraint". In this case, the evaluation information is "negative". In the question sentence generation information DB 4013, keyword information and evaluation information are similarly associated and recorded for other sentences.

FIG. 10 shows an example of a co-occurrence expression of each financial product recorded in the question sentence generation information DB 4013. As shown in FIG. A co-occurrence expression is a set of words generated based on attribute information of financial products. The number of words to be combined in a co-occurrence expression can be arbitrarily determined, and a combination of three-word abnormalities may be used as a co-occurrence expression. For example, co-occurrence expressions such as (Southeast Asia, economy), (Thailand, economy), and (Malaysia, consumption) are associated in advance and recorded for the financial product with the financial product ID “ASSET001”. For example, the sentence information acquisition unit 405 associates sentences with financial products based on these co-occurrence expressions and keyword information. In FIG. 10, texts with text IDs “S001” and “S002” are associated with the financial product with the financial product ID “ASSET001”. Texts with text IDs "S003" and "S004" are associated with the financial product with the financial product ID "ASSET002", which is a financial product targeting the United States.

Various information recorded in the question text generation information DB 4013 may be generated by the risk tolerance calculation device 101, or information generated by another information processing device is acquired by the risk tolerance calculation device 101, may be recorded.

The communication unit 402 controls communication between the risk tolerance calculation device 101 and the user terminal 102 and text information providing device 103 .

The product information acquisition unit 403 acquires product information by referring to the financial product DB 4011, for example, when there is a request to create a risk tolerance map M. FIG.

The classification information allocation unit 404 allocates classification information such as that shown in FIG. 8 to financial products based on similarity information and risk information included in the acquired product information.

The text information acquisition unit 405, for example, refers to the question generation information DB 4013 and acquires text information associated with the financial product and evaluation information for the text information based on the attribute information.

The question presentation unit 406 generates a question regarding the classification information of financial products based on the text information and the evaluation information, and presents it to the user. The process of generating a question sentence will be described later.

The preference information acquisition unit 407 acquires from the user preference information indicating the user's preference for the question text. Preference information is, for example, information indicating options that the user has answered to be options that the user prefers in a question including options.

Based on the preference information, the risk tolerance calculation unit 408 calculates the user's risk tolerance for the category information for which the question sentence is generated. For example, the risk tolerance calculation unit 408 sets the risk tolerance to 100 for the category information corresponding to the option that the user prefers, and sets the risk tolerance to 0 for the category information corresponding to the other option. When receiving a plurality of user selections for the same classification information, the risk tolerance calculation unit 408 sets the average value of the risk tolerances as the risk tolerance for the classification information. For example, if the risk tolerance for certain classification information is 100 due to the first preference information and the risk tolerance is 0 due to the second preference information, the average value of 50 is used as the risk tolerance. and When calculating the risk tolerance, weights based on responses to other classification information may be taken into consideration.

Acquisition of sentence information and evaluation information by the sentence information acquisition unit 405, presentation of question sentences by the question sentence presentation unit 406, acquisition of preference information by the preference information acquisition unit 407, and risk The calculation of the risk tolerance by the tolerance calculation unit 408 is repeated to calculate the risk tolerance for all the classification information of the risk tolerance map.

Based on the classification information and the risk tolerance of the classification information, the screen information generation unit 409 generates information for a risk tolerance display screen that presents the user with the risk tolerance for the classification information.

The candidate screen information generation unit 410 generates information of a product candidate screen that presents to the user financial product candidates included in the classification information whose risk tolerance satisfies a predetermined condition. The risk tolerance display screen and product candidate screen will be described later.

Processing by the risk tolerance calculation device 101 will be described with reference to FIG. 11 .

In step S<b>1101 , the risk tolerance calculation device 101 acquires a risk tolerance calculation request from the user through the user terminal 102 .

In step S1102, the product information acquisition unit 403 refers to the financial product DB 4011 and acquires product information of the financial product.

In step S1103, the classification information assigning unit 404 assigns classification information to the financial product based on the attribute information of the financial product. Financial instruments to which classification information has been assigned can each be classified into a risk tolerance map M, as shown in FIG. 12, for example.

In step S1104, the question text presenting unit 406 determines classification information for which a question text is to be generated. Since question sentences are generated for a plurality of pieces of classification information, presentation of question sentences, acquisition of preference information for the presented question sentences, and calculation of risk tolerance are repeatedly performed. When the question is generated in the first stage when the question is not generated for any of the classification information, the classification information and similarity information that maximize the similarity information and the risk information are BM. The classification information with the minimum risk information is determined. For example, as indicated by diagonal lines in FIG. 13A, at the first time, (similarity information, risk information)=(BM, 5) classification information and (similarity information, risk information)=(BM, 1 ) is determined as the classification information for which the question sentence is to be generated. Determination of classification information in subsequent loops will be described later.

In step S1105, the question text presenting unit 406 acquires text information and evaluation information associated with a financial product included in certain classification information. For example, in the case of the first time, the question sentence presentation unit 406 is associated with the A investment trust to which this classification information is assigned for the classification information of (similarity information, risk information) = (BM, 5) in FIG. Acquire text information and evaluation information. In addition, the question text presenting unit 406, regarding the classification information of (similarity information, risk information)=(BM, 1) in FIG. to get

In step S1106, the question text presenting unit 406 generates a question about the classification information as shown in FIG. 13(b) and transmits it to the user terminal. For example, for the classification information (BM, 5), the question sentence presenting unit 406 generates a question sentence in steps as shown in FIG. 14(a).

In step (1) of FIG. 14( a ), the question sentence presentation unit 406 refers to the question sentence generation information DB 4013 , and among the sentences associated with investment trust A, the sentences with positive evaluation information and the negative Get a sentence that is The question sentence presenting unit 406 refers to the keyword information of each sentence, and generates sentences with blank parts of words so that sentences can be created by keywords. In step (2) of FIG. 14( a ), the question sentence presenting unit 406 asks the question sentence presenting unit 406 to make a positive statement regarding the increase in asset value, such as “While there is an expectation of” at the end of the sentence based on the keyword of the sentence having positive evaluation information. and add a negative expression regarding the rise in asset value of "there is concern about" at the end of the sentence based on the keyword of the sentence with negative evaluation information. It should be noted that the set of positive sentences and negative sentences generated by the question sentence presenting unit 406 does not necessarily need to use the templates of "...has expectations", "on the other hand", and "...has concerns". For example, templates such as "... is doing well," "on the other hand," and "... is bad," as well as templates such as "... will be relaxed," "in a few years," and "... will be regulated." There may be Also, expressions between keywords are complemented using the question sentence presenting unit 406 . For example, the question sentence presenting unit 406 uses learning data in which the input is a combination of arbitrary words and the output is the head word, and a deep learning model such as BERT that has been trained to solve the fill-in-the-blank problem of particles or compound verbs is used. , to perform completion. In step (3) of FIG. 14(a), the question sentence presenting unit 406 generates a sentence in which each word is combined. As shown in FIG. 14(b), the question sentence presenting unit 406 can similarly generate a sentence for other investment trusts, such as the B investment trust.

In addition, the question text presenting unit 406 can also generate a question text based on attribute information of a certain financial product or information on a financial benchmark as another method of generating a question text. For a given financial product, a question is generated based on attribute information such as the constituent stocks of the financial product, the region targeted by the financial product, or the industrial field of the constituent stocks. For example, if there are X financial instruments, is it possible to tolerate concerns about investment expectations? generated. The question sentence A is generated by combining affirmative sentences and negative sentences based on the attribute information of X's financial product. As another example, for example, regarding constituent company A1 of the financial benchmark "benchmark BM1" shown in FIG. can also be generated. Also, based on the text information and evaluation information regarding Company A2, which is another component of the financial benchmark "Benchmark BM1", it is also possible to generate a sentence containing positive expressions and negative expressions. By allowing the user to make a selection by comparing the two, it is possible to know the brand that the user prefers.

In addition, the question text presenting unit 406 can also generate a question text for obtaining preferences for classification information including financial products having a certain technical theme, as another method of generating a question text. For example, regarding the classification information including investment trust C, a question such as that shown in FIG. 15(b) can be generated in order to acquire the risk tolerance for investment trust C itself. At this time, the question text presenting unit 406 generates a headline "Semiconductor technology in the United States" based on the region, country, and theme of the attribute information of the C investment trust. and the negative word "regulation". For example, the question sentence presentation unit 406, based on information in which positive words such as "utilization" and negative words such as "regulation" are associated with fields such as "semiconductor technology", sentences can be generated. The question sentence presenting unit 406 adds a positive expression regarding the rise in asset value, such as "while there is an expectation of" to the end of sentences containing positive words, and adds "of," to the end of sentences containing negative words. A question sentence is generated by adding a negative expression regarding the increase in asset value, such as "I have concerns."

For example, as shown in FIG. Information, Risk Information)=(BM, 1) A question sentence paired with a sentence (second question sentence) for the classification information is transmitted to the user terminal 102 . Specifically, the question text presenting unit 406 sets the classification information of (similarity information, risk information) = (BM, 5) and the risk tolerance for the classification information of (similarity information, risk information) = (BM, 1). In order to obtain the degree, we asked, "When there are products A and B, which one suits your tastes in terms of investment expectations and concerns? A: (question based on A investment trust), B: (F A question sentence based on investment trusts), C: I don't know." can be generated. Here, the option "C: I don't know" is an option provided corresponding to the deposit.

In step S1107, the preference information acquisition unit 407 acquires the user's preference for the classification information from the user terminal 102 through the answer to the question.

In step S1108, the risk tolerance calculator 408 calculates the user's risk tolerance for the classification information based on the preference information. For example, the risk tolerance calculation unit 408 calculates the risk tolerance for the classification information by summing the values of 100 when the classification information is preferred and 0 when the classification information is not selected. Calculate the average value for the number of responses as the risk tolerance.

FIG. 13(c) shows the risk tolerance map M in which the classification information for which the risk tolerance has been calculated is filled. When the user answers that the product of A matches his taste in response to the question in FIG. 13(b), (similarity information, risk information)=(BM , 5) is set to 100 (=100/1). Also, the risk tolerance for the classification information of (similarity information, risk information)=(BM, 1) is set to 0 (=0/1). Also, the risk tolerance for a deposit with risk information 0 is 0 (=0/1). If risk tolerance is calculated for the classification information of (similarity information, risk information)=(BM, 1) in subsequent iterations, the question sentence based on this classification information is more likely for the user to If it is answered that it suits the taste, the risk tolerance is set at 50 (=100/2).

In step S1109, the risk tolerance calculation unit 408 determines whether or not calculation of the risk tolerance for each piece of classification information in the risk tolerance map has been completed. For example, if a negative determination is made in step S1108, the process returns to step S1104, and the question text presenting unit 406 acquires text information and evaluation information associated with financial products included in the classification information for other classification information.

A description will be given of repetitive processing for subsequent risk tolerance calculation when the initial question sentence generation and risk tolerance calculation have been completed.

In step S1104, the question text presenting unit 406 determines classification information for which a question text is to be generated. For example, (1) Among the classification information with common similarity information and different risk information, for classification information with uncalculated risk tolerance, the classification information with the maximum risk tolerance and the classification information with the minimum risk tolerance are targeted. Determined as classification information. (2) In the answer to the first question, among the classified information having common risk information and different similarity information with the selected classified information selected as matching the taste, the similarity information is not calculated. Regarding the classification information, the classification information with the maximum similarity information and the classification information with the minimum similarity information are determined as the target classification information. In addition, (3) regarding the classification information that is diagonally positioned in the risk tolerance map M and whose risk tolerance has not yet been calculated for the selected classification information that has been selected as matching the taste in the answer to the first question , the classification information having the maximum distance and the minimum distance from the selected classification information on the risk tolerance map M are determined as the target classification information.

For example, in response to the question in FIG. 13(b), if the user answers that the item A suits his/her taste, the determination method (1) above gives the following as shown in FIG. 16(a): The classification information of (similarity information, risk information) = (BM, 4) and the classification information of (similarity information, risk information) = (BM, 2) are determined as the classification information for which the question sentence is to be generated. be. Further, according to the determination method of (2) above, as shown in FIG. The classification information (1, 5) is determined as the classification information for which the question sentence is to be generated. Further, according to the determination method of (3) above, as shown in FIG. The classification information (4, 4) is determined as the classification information for which the question sentence is to be generated. Here, it is assumed that the determination of the classification information in (1) to (3) above is collectively performed, and thereafter the calculation of the risk tolerance for each of the classification information is also collectively performed. It should be noted that the determination of the classification information and the calculation of the risk tolerance for the classification information may be repeated individually rather than collectively.

In step S1105, the question text presenting unit 406 acquires text information and evaluation information associated with the financial product included in the classification information for each classification information.

In step S1106, the question text presenting unit 406 generates a question about each piece of classification information and transmits it to the user terminal. For example, the question sentence S1 about the group of classification information shown in FIG. 16(a), the question sentence S2 about the group of classification information shown in FIG. 16(b), the group of classification information shown in FIG. 16(c) Generate a question sentence S3 about.

Further, in step S1106, the question text presenting unit 406 transmits to the user terminal 102 information for displaying on the user terminal 102 a question text group SG including a plurality of question texts S1 to S3. A question sentence presentation screen 1701 as shown in FIG. 17 is displayed on the user terminal 102 . By collectively transmitting a plurality of question sentences S1 to S4 in this way, it is possible to reduce the user's answer burden and improve convenience.

In step S1107, the preference information acquisition unit 407 acquires the user's preference information for the classification information corresponding to each question from the user terminal 102 based on each question.

In step S1108, the risk tolerance for each category information is calculated based on each preference information. Since it is possible to acquire the risk tolerance for a plurality of classification information by presenting the screen once, the number of presented questions can be reduced. As a result, it is possible to shorten the number of repetitions until a sufficient answer is obtained in the risk tolerance map M, thereby improving convenience for the user.

In step S1109, the risk tolerance calculation unit 408 determines whether or not calculation of the risk tolerance for each piece of classification information in the risk tolerance map has been completed. For example, if a negative determination is made in step S1108, the process returns to step S1104 and the same processes as those described above are repeated. For example, it is assumed that the risk tolerance of each classification information is calculated based on the answers to the questions in FIGS. 16(a) to (c). In this case, in the repeated processing, the classification information is determined by the methods (1) to (3) above from the initial state shown in FIG. Question sentence generation and risk tolerance for the classified information are calculated. Note that when only one piece of classification information is determined as shown in FIG. 18B, the question sentence presentation unit 406 asks, "When there is a product of A, which investment expectation or concern suits my taste?" Or, A: (question based on financial products belonging to the target classification information), B: I don't know" is generated. With this, the risk tolerance can be calculated even when there is only one target classification information.

As a result of repeating the processing up to step S1108, FIG. 18(e) shows the risk tolerance map M in which the calculation of the risk tolerance for the classification information shown in FIGS. 18(b) to 18(d) is completed. .

FIG. 19 shows an example in which it is determined in step S1109 that further repetition processing is required. That is, the classification information is determined by the methods (1) to (3) from the same initial state shown in FIG. 19(a), and the questions for the classification information shown in FIGS. Sentence production and risk tolerance are calculated. The risk tolerance map M in this case is shown in FIG. 19(e).

In this way, by determining the classification information for which the question sentence is generated and the risk tolerance is calculated by the above methods (1) to (3), the risk tolerance map M for each classification information is efficiently Risk tolerance can be calculated.

In addition, in repeating the above three times, while it is possible to obtain 10 answers for the risk tolerance for deposits, including the first time and the repeated nine times, the risk tolerance for each classification information is Calculated only once.

In such a case, for example, risk tolerance calculation may be performed again from the beginning, and risk tolerance calculation may be performed multiple times for each classification information.

In addition, when certain classification information is selected by the user as being preferable, the risk tolerance of the classification information adjacent to the classification information in the risk tolerance map may be increased while a certain weight is applied.

For example, if the user prefers a question based on the classification information of (similarity information, risk information) = (BM, 5), the surrounding (similarity information, risk information) = (BM, 4), ( 4, 5) and (4, 4) may be weighted while the risk tolerance is calculated. For example, when 100 is added to the risk tolerance total value of (similarity information, risk information) = (BM, 5) classification information, (similarity information, risk information) = (BM, 4) classification information Fifty may be added to the total risk tolerance. At this time, the classification information for which the risk tolerance is calculated by weighting may be used to generate a separate question sentence and calculate the risk tolerance in subsequent processing. By weighting, it is possible to obtain the risk tolerance for a plurality of classification information with one question sentence, so the number of presented questions can be reduced and the calculation speed of the risk tolerance is improved.

Note that the position on the risk tolerance map of the classification information from which the preference information is to be acquired may be displayed together on the question sentence presentation screen 1701 . As a result, the user can grasp which classification information he or she is responding to, and by making more careful or appropriate judgments, the accuracy of risk tolerance calculation is improved.

It is assumed that acquisition of risk tolerance is repeatedly performed and an affirmative determination is made in step S1109. FIG. 20 shows an example of risk tolerance for classification information at this time.

In step S1110, the screen information generation unit 409 generates information for generating the risk tolerance display screen 2101 as shown in FIG. 21 based on the classification information and the risk tolerance. The risk tolerance display screen 2101 has a visualized risk tolerance map 21011 , commentary information column 21012 , and tolerance extension suggestion information column 21013 . The visualized risk tolerance map 21011 is displayed so that each region of the risk tolerance map is color-coded according to the user's risk tolerance, for example, so that the degree of risk tolerance can be understood. This enables the user to grasp the trend of his/her own risk tolerance at a glance.

In the explanatory information column 21012, for example, in the visualized risk tolerance map 21011, according to the risk tolerance for the classification information with the highest similarity to the financial benchmark and the lowest risk, the range of the risk tolerance is displayed. An explanatory text will be displayed. The text displays information such as the range of risk tolerance being "slightly small", "small", "intermediate position", "slightly large", and "large". Also, in the commentary information column 21012, a chart visually showing the range of risk tolerance may be displayed.

The tolerance expansion suggestion information column 21013 displays information about the keyword or theme of the financial instrument selected on the condition that the risk tolerance satisfies a predetermined condition. A financial product including a keyword or theme displayed in the tolerance expansion suggestion information column 21013 is suggested and displayed as a possibility that the user may tolerate more risk. This allows users to review and consult sellers on financial products that can be incorporated to achieve targeted revenue.

In step S1121, candidate screen information generating section 410 generates information for displaying a product candidate screen that presents the user with financial product candidates whose risk tolerance satisfies a predetermined condition. FIG. 22 shows a product candidate screen 2201 as an example. As an example, the product candidate screen 2201 displays information on financial products assigned to classification information whose risk tolerance satisfies a predetermined condition.

Note that the risk tolerance display screen 2101 and the product candidate screen 2201 in FIG. 21 may be displayed as one screen to make it easier for the user to check the results. Alternatively, when the user selects a square of the visualized risk tolerance map 21011 on the user terminal 102, financial product information corresponding to the classification information of the selected square may be displayed like a candidate screen. good.

The embodiments described above are for facilitating understanding of the present invention, and are not intended to limit and interpret the present invention. Each element included in the embodiment and its arrangement, conditions, shape, size, etc. are not limited to those illustrated and can be changed as appropriate. Also, it is possible to partially replace or combine the configurations shown in different embodiments.

DESCRIPTION OF SYMBOLS 10... Risk tolerance calculation system 101... Risk tolerance calculation apparatus 102... User terminal 103... Text information provision apparatus 403... Product information acquisition part 404... Classification information allocation part 405... Text information acquisition part 406 Question statement presentation unit 407 Preference information acquisition unit 408 Risk tolerance calculation unit 409 Screen information generation unit 410 Candidate screen information generation unit

Claims (8)

to the computer,
Calculated based on attribute information of a financial product, similarity information indicating the degree of similarity between the fluctuation trend of a financial benchmark referenced by the financial product and the price fluctuation trend of the financial product, and the price fluctuation of the financial product. a product information acquisition process for acquiring product information including risk information of the financial product;
a classification information allocation process for allocating classification information to the financial product based on the similarity information and the risk information;
Text information acquisition processing for acquiring text information associated with the financial product and evaluation information for the text information based on the attribute information;
a question sentence presentation process for presenting a question sentence related to the classification information to the user based on the sentence information and the evaluation information;
Preference information acquisition processing for acquiring from the user preference information indicating the user's preference for the classification information;
and a risk tolerance calculation process for calculating the user's risk tolerance for the classification information based on the preference information. The program according to claim 1,
The product information acquisition process includes acquiring, as the product information, first product information of a first financial product and second product information of a second financial product,
The classification information assignment process includes assigning first classification information to the first financial instrument and assigning second classification information to the second financial instrument;
The text information acquisition process acquires the text information and the evaluation information associated with the first financial product based on the attribute information of the first financial product, and acquires the text information and the evaluation information associated with the first financial product, obtaining the textual information and the valuation information associated with the second financial instrument based on
The question sentence presentation process presents the user with a first question sentence related to the first classification information based on the sentence information and the evaluation information, and presents the second classification sentence based on the sentence information and the evaluation information. Presenting a second question about information to the user;
The preference information acquisition process includes acquiring from the user preference information indicating the user's preferences for the first classification information and the second classification information,
The program, wherein the risk tolerance calculation process includes calculating the user's risk tolerance for each of the first classification information and the second classification information based on the preference information. The program according to claim 1,
The product information acquisition process includes acquiring first product information of a first financial product and second product information of a second financial product as the product information,
The classification information assignment process includes assigning the classification information to the first financial instrument and the second financial instrument,
The text information acquisition process includes the text information and the evaluation information of the first financial product and the second financial product based on the attribute information of the first financial product and the attribute information of the second financial product. including acquiring the text information and the evaluation information of
The question presentation process presents a first question about the classification information to the user based on the attribute information, the text information and the evaluation information of the first financial product, and presents the first question about the second financial product to the user. Presenting a second question about the classification information to the user based on the attribute information, the text information, and the evaluation information;
The preference information acquisition process includes acquiring from the user preference information indicating the user's preferences for each of the first question sentence and the second question sentence,
The program, wherein the risk tolerance calculation process includes calculating the user's risk tolerance for the financial product included in the classification information based on the preference information. The program according to any one of claims 1 to 3,
The question text presentation process includes presenting to the user a plurality of question texts related to the classification information,
The program, wherein the risk tolerance calculation process includes calculating the risk tolerance based on the preference information acquired through the plurality of question sentences. The program according to any one of claims 1 to 4,
to the computer;
a screen information generating process for generating information for a risk tolerance display screen that presents the risk tolerance for the classification information to the user, based on the classification information and the risk tolerance. The program according to any one of claims 1 to 5,
to the computer;
a candidate screen information generating process for generating information of a product candidate screen for presenting to the user financial product candidates included in the classification information whose risk tolerance satisfies a predetermined condition; Calculated based on attribute information of a financial product, similarity information indicating the degree of similarity between the fluctuation trend of a financial benchmark referenced by the financial product and the price fluctuation trend of the financial product, and the price fluctuation of the financial product. a product information acquisition unit that acquires product information including risk information of the financial product;
a classification information allocation unit that allocates classification information to the financial product based on the similarity information and the risk information;
a text information acquisition unit that acquires text information associated with the financial product and evaluation information for the text information based on the attribute information;
a question sentence presentation unit that presents a question sentence about the classification information to the user based on the sentence information and the evaluation information;
a preference information acquisition unit configured to acquire, from the user, preference information indicating the user's preference for the classification information;
and a risk tolerance calculation unit configured to calculate the user's risk tolerance for the classification information based on the preference information. the computer
Calculated based on attribute information of a financial product, similarity information indicating the degree of similarity between the fluctuation trend of a financial benchmark referenced by the financial product and the price fluctuation trend of the financial product, and the price fluctuation of the financial product. obtaining product information including risk information of the financial product;
assigning classification information to the financial instruments based on the similarity information and the risk information;
Acquiring text information associated with the financial product and evaluation information for the text information based on the attribute information;
presenting a user with a question regarding the classification information based on the text information and the evaluation information;
obtaining preference information from the user indicating the user's preference for the classification information;
calculating the user's risk tolerance for the classification information based on the preference information.

Images