JP7366355B2 - Loan approval probability calculation device, loan approval probability calculation method and program - Google Patents
Loan approval probability calculation device, loan approval probability calculation method and program Download PDFInfo
- Publication number
- JP7366355B2 JP7366355B2 JP2021155320A JP2021155320A JP7366355B2 JP 7366355 B2 JP7366355 B2 JP 7366355B2 JP 2021155320 A JP2021155320 A JP 2021155320A JP 2021155320 A JP2021155320 A JP 2021155320A JP 7366355 B2 JP7366355 B2 JP 7366355B2
- Authority
- JP
- Japan
- Prior art keywords
- loan
- loan approval
- attribute information
- user
- approval
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000004364 calculation method Methods 0.000 title claims description 51
- 238000013210 evaluation model Methods 0.000 claims description 40
- 238000000034 method Methods 0.000 claims description 24
- 238000010801 machine learning Methods 0.000 claims description 23
- 238000012216 screening Methods 0.000 claims description 12
- 238000012545 processing Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 9
- 230000008569 process Effects 0.000 claims description 9
- 238000013473 artificial intelligence Methods 0.000 description 20
- 238000010586 diagram Methods 0.000 description 15
- 238000003745 diagnosis Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000002349 favourable effect Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
Landscapes
- Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)
Description
本発明は、ローン商品の融資承認確率および借入可能額を算出する融資承認確率算出装置、融資承認確率算出方法及びプログラムに関する。 The present invention relates to a loan approval probability calculation device, a loan approval probability calculation method, and a program for calculating the loan approval probability and borrowable amount of a loan product.
金融機関に住宅ローン等のローン商品の申し込みを行う際に、どのローン商品を選択するのがベストなのか、どのローン商品が最も有利な条件で借り入れできるのか、また金融機関の審査にどれくらいの確率で通るのか、等の情報を事前に知ることができればユーザはローン商品の選択を効率的に行えるようになる。
例えば、特許文献１においては、金融機関に対して融資の申し込みをする前に、事前に審査を行って融資の可否を判断し、融資を受けることが可能な金融商品の融資情報を提供する情報提供装置および方法が開示されている。本装置では、融資の申込者の信用度を算出するための与信モデルに基づき、申込者の財務情報をから融資の可否を判定し、融資が可能な場合に融資を受けることが可能な金融商品の融資情報のレコメンドを行うようにし、これによりユーザに適切な金融商品の融資情報を提供可能としている。
When applying for a loan product such as a home loan from a financial institution, which loan product is the best to choose, which loan product can be borrowed under the most advantageous conditions, and what is the probability of the financial institution's examination? If the user can know in advance information such as whether the loan will be approved or not, the user will be able to select a loan product more efficiently.
For example, in Patent Document 1, before applying for a loan to a financial institution, a preliminary examination is performed to determine whether a loan is available, and information is provided that provides loan information on financial products for which a loan can be obtained. A providing apparatus and method are disclosed. Based on a credit model for calculating the creditworthiness of a loan applicant, this device determines whether or not a loan can be granted based on the applicant's financial information, and if a loan is possible, it determines whether or not a loan can be obtained. Loan information is recommended, thereby making it possible to provide users with loan information on appropriate financial products.
ところで、このようなローン商品の融資条件は金融機関により異なっているため、金融機関毎に審査にどれくらいの確率で通るのかをできる限り正確な情報をユーザに提供することが重要になってくる。
そこで本発明では、蓄積された多くの審査結果等の情報を学習データとしてＡＩ（Artificial Intelligence）に投入して学習モデルを構築して利用することにより、ユーザ属性に応じたローン商品の融資承認確率および借入可能額について高い精度で算出する融資承認確率算出装置及び融資承認確率算出方法を提供する。
By the way, the financing conditions for such loan products differ depending on the financial institution, so it is important to provide users with as accurate information as possible about the probability of passing the examination for each financial institution.
Therefore, in the present invention, by inputting a large amount of accumulated information such as examination results as learning data into AI (Artificial Intelligence) and building and using a learning model, the probability of loan approval for loan products according to user attributes is improved. The present invention provides a loan approval probability calculation device and a loan approval probability calculation method that calculate the borrowable amount with high accuracy.
上記の目的を達成するために、第１発明に係る融資承認確率算出装置は、
ユーザの属性情報からローン商品の融資承認確率および借入可能額を算出する融資承認確率算出装置であって、
前記ユーザの属性情報の入力を受け付けるユーザ属性情報入力部と、
前記ユーザの個人信用情報を取得する個人信用情報取得部と、
前記個人信用情報取得部により取得された前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する個人信用情報判定部と、
前記個人信用情報判定部により前記個人信用情報条件を充足すると判定されたユーザに対して、前記ユーザ属性情報入力部により入力を受け付けた該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を予測する融資承認機械学習部と、
を備え、
前記融資承認機械学習部は、
前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存する融資承認評価モデル作成部と、
前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、前記融資承認評価モデル作成部で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出すると共に、その確率を算出した該属性情報における指標の特徴を出力する融資承認確率算出部と
を有することを特徴とする。
In order to achieve the above object, a loan approval probability calculation device according to the first invention includes:
A loan approval probability calculation device that calculates a loan approval probability and a loanable amount for a loan product from user attribute information, the device comprising:
a user attribute information input unit that accepts input of the user's attribute information;
a personal credit information acquisition unit that acquires personal credit information of the user;
a personal credit information determination unit that determines whether the user's personal credit information acquired by the personal credit information acquisition unit satisfies predetermined personal credit information conditions;
For a user who is determined by the personal credit information determination unit to satisfy the personal credit information conditions, the loan product is approved based on the attribute information of the user whose input is accepted by the user attribute information input unit. a loan approval machine learning unit that predicts the probability and loan amount;
Equipped with
The loan approval machine learning department is
From the attribute information used in the approval examination of the loan product and the result of the approval examination based on the attribute information, a loan approval evaluation model is created that associates the characteristics of each index of the attribute information with the loan approval probability. A loan approval evaluation model creation department to save;
When the attribute information of the user for which the loan approval probability of the loan product is to be calculated is specified, the loan approval evaluation model stored by the loan approval evaluation model creation unit is based on a plurality of indicators of the attribute information. The present invention is characterized by comprising a loan approval probability calculation unit that calculates the loan approval probability and the borrowable amount by and outputs the characteristics of the index in the attribute information for which the probability has been calculated.
第２発明に係る融資承認確率算出装置は、第１発明において、
前記融資承認機械学習部は、
融資承認評価モデル作成部が、複数の前記ローン商品について、融資承認評価モデルを作成して保存し、
前記融資承認確率算出部が、融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、複数の前記ローン商品について、前記融資承認評価モデル作成部で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出し、
前記融資承認確率算出部により融資承認確率が算出された複数の前記ローン商品について、所定の評価項目によりランキングして出力することを特徴とする。
A loan approval probability calculation device according to a second invention includes, in the first invention,
The loan approval machine learning department is
a loan approval evaluation model creation unit creates and saves loan approval evaluation models for the plurality of loan products;
When the attribute information of the user for which the loan approval probability calculation unit is to calculate the loan approval probability is specified, the loan approval evaluation model stored by the loan approval evaluation model creation unit for a plurality of loan products is determined by the loan approval probability calculation unit. Calculate the loan approval probability and loan amount by
The present invention is characterized in that the plurality of loan products for which loan approval probabilities have been calculated by the loan approval probability calculation unit are ranked and output based on predetermined evaluation items.
第３発明に係る融資承認確率算出方法は、
ユーザの属性情報からローン商品の融資承認確率および借入可能額を算出する融資承認確率の算出方法であって、
前記ユーザの属性情報の入力を受け付けるユーザ属性情報入力工程と、
前記ユーザの個人信用情報を取得する個人信用情報取得工程と、
前記個人信用情報取得工程により取得された前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する個人信用情報判定工程と、
前記個人信用情報判定工程により前記個人信用情報条件を充足すると判定されたユーザに対して、前記ユーザ属性情報入力工程により入力を受け付けた該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を予測する融資承認機械学習処理工程と
を備え、
前記融資承認機械学習処理工程は、
予め、前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存しておき、
前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、保存された前記融資承認評価モデルにより融資承認確率および借入可能額を算出すると共に、その確率を算出した該属性情報における指標の特徴を出力することを特徴とする。
The loan approval probability calculation method according to the third invention is as follows:
A method for calculating a loan approval probability for calculating a loan approval probability and a borrowable amount for a loan product from user attribute information, the method comprising:
a user attribute information input step of accepting input of the user's attribute information;
a personal credit information acquisition step of acquiring personal credit information of the user;
a personal credit information determination step of determining whether the user's personal credit information acquired in the personal credit information acquisition step satisfies predetermined personal credit information conditions;
For a user who is determined to satisfy the personal credit information conditions in the personal credit information determination step, the loan product is approved based on the attribute information of the user whose input is accepted in the user attribute information input step. Equipped with a loan approval machine learning process that predicts the probability and loan amount,
The loan approval machine learning processing step includes:
In advance, from the attribute information used in the approval examination of the loan product and the result of the approval examination based on the attribute information, a loan approval evaluation model is created that associates the characteristics of each index of the attribute information with the loan approval probability. and save it.
When the attribute information of the user for which the loan approval probability of the loan product is to be calculated is specified, the loan approval probability and borrowability are calculated based on the stored loan approval evaluation model based on a plurality of indicators of the attribute information. It is characterized by calculating the amount and outputting the characteristics of the index in the attribute information for which the probability has been calculated.
第４発明に係る融資承認確率算出プログラムは、
ユーザの属性情報からローン商品の融資承認確率および借入可能額を算出するプログラムであって、
コンピュータに、
属性情報が入力された前記ユーザの個人信用情報を取得させ、
取得した前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定させ、
個人信用情報条件を充足すると判定された前記ユーザに対して、入力された該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を融資承認機械学習処理により算出させるプログラムであって、
前記融資承認機械学習処理は、
予め、前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存させておき、
前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、保存された前記融資承認評価モデルにより融資承認確率および借入可能額を算出させると共に、その確率を算出した該属性情報における指標の特徴を出力させることを特徴とする。
The loan approval probability calculation program according to the fourth invention is as follows:
A program that calculates the loan approval probability and borrowable amount of a loan product from user attribute information,
to the computer,
obtain personal credit information of the user whose attribute information has been input;
determining whether the acquired personal credit information of the user satisfies predetermined personal credit information conditions;
A program that causes the user who has been determined to satisfy the personal credit information conditions to calculate the loan approval probability and borrowable amount for the loan product using loan approval machine learning processing based on the input attribute information of the user. And,
The loan approval machine learning process is
In advance, from the attribute information used in the approval examination of the loan product and the result of the approval examination based on the attribute information, a loan approval evaluation model is created that associates the characteristics of each index of the attribute information with the loan approval probability. and save it,
When the attribute information of the user for which the loan approval probability of the loan product is to be calculated is specified, the loan approval probability and borrowability are calculated based on the stored loan approval evaluation model based on a plurality of indicators of the attribute information. The present invention is characterized in that the amount is calculated and the characteristics of the index in the attribute information whose probability has been calculated are output.
本発明によれば、ＡＩモデルを構築して利用することにより、ユーザ属性に応じたローン商品の融資承認確率および借入可能額を算出する際の精度の向上を期待することができる。特に、ユーザ属性情報に加え、個人信用情報、審査を行った金融機関、借入希望額、審査結果等の情報を学習データとしてＡＩに投入して学習モデルを構築することにより、蓄積された審査結果データから算出する精度が継続的に向上するとともに、自律的に学習モデルの調整が可能となるため審査基準の変化に柔軟に対応可能となり、さらに算出した予測の説明が可能になる効果も期待することができる。予測の説明ができることにより、ユーザにどのようにすれば融資承認確率を上げ、より有利な条件のローンを借りられるかをアドバイスすることが可能になる。 According to the present invention, by constructing and using an AI model, it is possible to expect an improvement in accuracy when calculating the loan approval probability and loanable amount of a loan product according to user attributes. In particular, in addition to user attribute information, information such as personal credit information, financial institution that conducted the screening, desired loan amount, screening results, etc. is input to AI as learning data to build a learning model, and the accumulated screening results. In addition to continuously improving the accuracy of calculations from data, it is also possible to autonomously adjust the learning model, making it possible to respond flexibly to changes in screening criteria, and also making it possible to explain the calculated predictions. be able to. By being able to explain predictions, it becomes possible to advise users on how to increase the probability of loan approval and obtain a loan with more advantageous terms.
以下、本発明を実施するための形態について図面を参照して詳細に説明する。なお、以下に説明する実施形態は、あくまでも、本発明の理解を容易にするために挙げた一例にすぎず、本発明を限定するものではない。すなわち、本発明は、その趣旨を逸脱しない限りにおいて、以下に説明する実施形態から変更又は改良され得る。 DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments for carrying out the present invention will be described in detail with reference to the drawings. Note that the embodiments described below are merely examples given to facilitate understanding of the present invention, and do not limit the present invention. That is, the present invention may be modified or improved from the embodiments described below without departing from the spirit thereof.
本発明の実施形態に係る融資承認確率算出装置の利用環境について図１を参照しながら説明する。図１は、融資承認確率算出装置を利用するシステム構成の一例を示す模式図である。図１に示すように、融資承認確率算出装置10は、ユーザの属性情報から住宅ローン等のローン商品の融資承認確率および借入可能額を算出する装置であり、利用者端末20とネットワーク（ＮＷ）を経由して接続されている。ユーザは住宅ローン等のローン商品の申し込みを行う前に、利用者端末20から融資承認確率算出装置10にアクセスしユーザ属性情報を入力することにより、融資承認確率算出装置10では、金融機関の審査にどれくらいの確率で通るかという融資承認確率や借入可能額の予測結果を算出し、その予測結果を利用者端末20に表示する。融資承認確率や借入可能額については、金融機関により異なるため、ユーザはその予測結果を基にどの金融機関のローン商品を選択するのが最適か判断してローン商品の申し込みを行うことになる。 The usage environment of the loan approval probability calculation device according to the embodiment of the present invention will be described with reference to FIG. FIG. 1 is a schematic diagram showing an example of a system configuration using a loan approval probability calculation device. As shown in FIG. 1, the loan approval probability calculation device 10 is a device that calculates the loan approval probability and borrowable amount for loan products such as home loans from user attribute information, and is connected to a user terminal 20 and a network (NW). connected via. Before applying for a loan product such as a home loan, the user accesses the loan approval probability calculation device 10 from the user terminal 20 and inputs user attribute information. The prediction result of the loan approval probability and the borrowable amount is calculated, and the prediction result is displayed on the user terminal 20. Since the loan approval probability and loan amount differ depending on the financial institution, the user will judge which financial institution's loan product is best to select based on the prediction result and apply for the loan product.
次に、本発明の実施形態に係る融資承認確率算出装置の機能構成について図２を参照しながら説明する。図２は、融資承認確率算出装置を説明するための機能ブロック図である。
図２に示すように、融資承認確率算出装置10は、ユーザ属性情報入力部110、個人信用情報取得部120、個人信用情報判定部130、融資承認機械学習部140、ローン商品記憶部150、ユーザ情報記憶部160、融資承認結果記憶部170、を備えている。さらに融資承認機械学習部140は、融資承認評価モデル作成部141、融資承認確率算出部142、を備えている。以下に、各機能部について詳細に説明する。
Next, the functional configuration of the loan approval probability calculation device according to the embodiment of the present invention will be described with reference to FIG. 2. FIG. 2 is a functional block diagram for explaining the loan approval probability calculation device.
As shown in FIG. 2, the loan approval probability calculation device 10 includes a user attribute information input unit 110, a personal credit information acquisition unit 120, a personal credit information determination unit 130, a loan approval machine learning unit 140, a loan product storage unit 150, a user It includes an information storage section 160 and a loan approval result storage section 170. Furthermore, the loan approval machine learning section 140 includes a loan approval evaluation model creation section 141 and a loan approval probability calculation section 142. Each functional unit will be explained in detail below.
＜ユーザ属性情報入力部＞
ユーザ属性情報入力部110は、ユーザが住宅ローン等のローン商品の申し込みを行うために利用者端末20から入力したユーザ属性情報を受け付ける。このユーザ属性情報は、ローン商品の融資承認確率および借入可能額を算出するために必要な情報であり、ユーザ情報記憶部160に保存される。
<User attribute information input section>
The user attribute information input unit 110 receives user attribute information input by the user from the user terminal 20 in order to apply for a loan product such as a home loan. This user attribute information is information necessary for calculating the loan approval probability and loanable amount of the loan product, and is stored in the user information storage unit 160.
＜個人信用情報取得部＞
個人信用情報取得部120は、ユーザ属性情報入力部110で受け付けたユーザに関する個人信用情報を取得する。個人信用情報を取得する方法としては、例えばＪＩＣＣ（株式会社日本信用情報機構）を利用して取得する方法がある。
<Personal credit information acquisition department>
The personal credit information acquisition unit 120 acquires the personal credit information regarding the user received by the user attribute information input unit 110. As a method of acquiring personal credit information, for example, there is a method of acquiring it using JICC (Japan Credit Information Corporation).
＜個人信用情報判定部＞
個人信用情報判定部130は、個人信用情報取得部120により取得されたユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する。ここで個人信用情報条件とは、異常値をはじくための条件（例えば、ネガ情報等）であり、ユーザのローン申請をこの条件でフィルタリングして、この条件を充足しないローン申請をはじく処理を行っている。
<Personal Credit Information Judgment Department>
The personal credit information determination unit 130 determines whether the user's personal credit information acquired by the personal credit information acquisition unit 120 satisfies predetermined personal credit information conditions. Here, personal credit information conditions are conditions for rejecting abnormal values (for example, negative information, etc.), and the process is to filter user loan applications using these conditions and reject loan applications that do not meet these conditions. ing.
＜融資承認機械学習部＞
融資承認機械学習部140は、個人信用情報判定部130により個人信用情報条件を充足すると判定されたユーザに対して、ユーザ属性情報入力部110により入力を受け付けた該ユーザの属性情報に基づいて、ローン商品の融資承認確率および借入可能額を予測する。
<Loan approval machine learning department>
The loan approval machine learning unit 140 performs a process based on the attribute information of the user whose input is accepted by the user attribute information input unit 110 for the user who is determined by the personal credit information determination unit 130 to satisfy the personal credit information conditions. Predict the approval probability and loan amount for loan products.
なお、この融資承認機械学習部140ではＡＩモデルを利用してローン商品の融資承認確率および借入可能額を予測するため、事前に学習データをＡＩに投入してモデルを構築することが必要になる。この学習データには、ユーザ属性情報に加え、個人信用情報、審査を行った金融機関、借入希望額、審査結果等の情報が含まれる。
また本明細書中に記載のユーザ属性情報および個人信用情報として、例えば、次のような情報が含まれる。
ユーザ属性情報＝｛年齢、年収、業種、職種、雇用形態、勤務先規模、家族構成、自己資金、住宅ローン以外の負債、物件所在地、保有資産・・・｝
個人信用情報＝｛無担保返済額、消費者ローン件数、30日未満延滞回数、30日以上延滞回数、グレー債権数、ネガ債権数、無担保返済残高、有担保返済残高、・・・｝
またＡＩモデルを構築するプラットフォームとして、例えばDataRobot（DataRobot Japan株式会社の製品）を利用する。DataRobotを利用することにより、「モデル作成の自動化により開発時間を短縮できる」「データのモニタリングによりユーザ、審査基準の傾向を把握できる」「予測の説明が可能になる」等のメリットがある。
Note that this loan approval machine learning unit 140 uses an AI model to predict the loan approval probability and loan amount for loan products, so it is necessary to input learning data into the AI in advance to build the model. . In addition to user attribute information, this learning data includes information such as personal credit information, the financial institution that conducted the examination, the desired loan amount, and the examination results.
Further, the user attribute information and personal credit information described in this specification include, for example, the following information.
User attribute information = {Age, annual income, industry, occupation, employment type, workplace size, family structure, own funds, debts other than mortgage, property location, owned assets...}
Personal credit information = {unsecured repayment amount, number of consumer loans, number of delinquencies of less than 30 days, number of delinquencies of 30 days or more, number of gray loans, number of negative receivables, unsecured repayment balance, secured repayment balance,...}
Furthermore, as a platform for building an AI model, for example, DataRobot (a product of DataRobot Japan Inc.) is used. By using DataRobot, there are benefits such as ``automating model creation to shorten development time,'' ``monitoring data to understand trends in users and screening criteria,'' and ``making it possible to explain predictions.''
＜融資承認評価モデル作成部＞
融資承認評価モデル作成部141は、ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存する。
<Loan approval evaluation model creation department>
The loan approval evaluation model creation unit 141 associates the characteristics of each index of the attribute information with the loan approval probability from the attribute information used in the approval screening of the loan product and the result of the approval screening based on the attribute information. Create and save a loan approval evaluation model.
この融資承認評価モデルは、ＡＩモデルを利用してローン商品の融資承認確率および借入可能額を予測するために作成するものであり、事前に学習データをＡＩに投入してモデルを構築する。この学習データには、上記で説明した通り、ユーザ属性情報に加え、個人信用情報、審査を行った金融機関、借入希望額、審査結果等の情報が含まれる。
また今回のモデルを構築するにあたり、自社の保有する約１万件の学習データを使用しているため、蓄積された審査結果データから融資承認確率および借入可能額を高い精度で算出することができる。さらにＡＩに投入する学習データが増えるほど精度が継続的に向上することが自社の分析から証明されている。また自律的にモデルの調整が可能となるため審査基準の変化に柔軟に対応できる効果が期待される。
This loan approval evaluation model is created in order to predict the loan approval probability and borrowable amount of a loan product using an AI model, and the model is constructed by inputting learning data into the AI in advance. As explained above, in addition to the user attribute information, this learning data includes information such as personal credit information, the financial institution that conducted the examination, the desired loan amount, and the examination results.
In addition, in building this model, we used approximately 10,000 training data that we own, so we are able to calculate the loan approval probability and loan amount with high accuracy from the accumulated examination result data. . Furthermore, our own analysis has proven that the more training data we feed into AI, the more accuracy it continuously improves. Furthermore, since it will be possible to adjust the model autonomously, it is expected to have the effect of being able to respond flexibly to changes in screening standards.
＜融資承認確率算出部＞
融資承認確率算出部142は、ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、融資承認評価モデル作成部141で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出すると共に、その確率を算出した該属性情報における指標の特徴を出力する。
<Loan approval probability calculation department>
When the attribute information of the user for which the loan approval probability of a loan product is to be calculated is specified, the loan approval probability calculation unit 142 calculates the loan approval evaluation model creation unit 141 based on a plurality of indicators of the attribute information. The loan approval probability and the loanable amount are calculated using the stored loan approval evaluation model, and the characteristics of the index in the attribute information for which the probability was calculated are output.
この融資承認評価モデルにより融資承認確率および借入可能額を算出する際に、その予測結果になった理由について特徴量を用いて説明することができる。図４は、融資承認評価モデルを利用して融資承認確率の予測の説明を表示する一例を示したイメージ図である。図４の例では、ユーザ毎に融資承認確率の予測結果と、その予測結果を説明する理由として３個の特徴量（説明１，説明２，説明３）を用いて表示するようにしている。一行目の表示例では、予測結果８０％と高い確率であり、その理由として「説明１：＋＋＋自己資金」「説明２：＋＋＋勤務先規模」「説明３：＋＋年収」が表示されている。また最終行の表示例では、予測結果２０％と低い確率であり、その理由として「説明１：－－－年齢」「説明２：－－勤務先規模」「説明３：－返済比率」が表示されている。ここで、予測結果にポジティブな影響を与えた項目には「＋」、ネガティブな影響を与えた項目には「－」が付与され、「＋」「－」の個数は影響度合いの大きさを示している。 When calculating the loan approval probability and the borrowable amount using this loan approval evaluation model, it is possible to explain the reason for the predicted result using the feature amount. FIG. 4 is an image diagram showing an example of displaying an explanation of prediction of loan approval probability using the loan approval evaluation model. In the example of FIG. 4, the prediction result of the loan approval probability and the reason for explaining the prediction result are displayed using three feature amounts (Explanation 1, Explanation 2, Explanation 3) for each user. In the display example in the first line, the prediction result is a high probability of 80%, and the reasons for this are "Explanation 1: +++ own funds", "Explanation 2: +++ workplace size", and "Explanation 3: ++ annual income". In addition, in the display example on the last line, the prediction result is a low probability of 20%, and the reasons for this are "Explanation 1: ---Age," "Explanation 2: ---Work size," and "Explanation 3: --Repayment ratio." has been done. Here, items that had a positive impact on the prediction results are given a "+", and items that had a negative impact are given a "-", and the number of "+" and "-" indicates the degree of influence. It shows.
また融資承認確率算出部142で算出した融資承認確率の予測結果は、当該ローン商品の申し込みを行ったユーザの利用者端末20に表示される。図５は、融資承認確率の予測結果を表示する一例を示したイメージ図である。図５の例では、銀行名と、その銀行における融資承認確率の予測結果８０％と、ローン条件（金利、毎月返済額、団体信用生命保険、等）の情報をユーザの利用者端末20に表示するようにしている。さらにAI診断結果として融資承認確率を受けてユーザが取るべき行動（ローンを申し込むべきかどうか）やAIアドバイスとして融資承認確率を上げるためのアドバイスが提示される。これらのインストラクションに従うことでユーザはより有利な条件のローンを自然に選ぶことができる。 Further, the predicted result of the loan approval probability calculated by the loan approval probability calculation unit 142 is displayed on the user terminal 20 of the user who applied for the loan product. FIG. 5 is an image diagram showing an example of displaying the prediction result of the loan approval probability. In the example shown in FIG. 5, the bank name, the predicted loan approval probability of 80% at that bank, and information on loan conditions (interest rate, monthly repayment amount, group credit life insurance, etc.) are displayed on the user terminal 20 of the user. I try to do that. Furthermore, the user is presented with the actions they should take (whether or not to apply for a loan) based on the probability of loan approval as an AI diagnosis result, as well as advice on how to increase the probability of loan approval as AI advice. By following these instructions, users can naturally choose loans with more favorable terms.
また融資承認確率算出部142で算出した借入可能額の予測結果は、当該ローン商品の申し込みを行ったユーザの利用者端末20に表示される。図６は、借入可能額の予測結果を表示する一例を示したイメージ図である。図６の例では、借入可能額と、適用金利の情報をユーザの利用者端末20に表示するようにしている。 Further, the predicted loan amount calculated by the loan approval probability calculation unit 142 is displayed on the user terminal 20 of the user who applied for the loan product. FIG. 6 is an image diagram showing an example of displaying the prediction result of the borrowable amount. In the example of FIG. 6, information on the borrowable amount and the applicable interest rate is displayed on the user terminal 20 of the user.
また融資承認確率算出部142では、複数のローン商品の融資承認確率および借入可能額について所定の評価項目によりランキングして出力することもできる。
すなわち、融資承認評価モデル作成部141が、複数のローン商品について、融資承認評価モデルを作成して保存し、融資承認確率算出部142が、融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、複数のローン商品について、融資承認評価モデル作成部141で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出し、融資承認確率算出部142により融資承認確率が算出された複数のローン商品について、所定の評価項目によりランキングして出力する。
Furthermore, the loan approval probability calculation unit 142 can also rank and output the loan approval probabilities and loanable amounts of a plurality of loan products based on predetermined evaluation items.
That is, the loan approval evaluation model creation unit 141 creates and stores loan approval evaluation models for a plurality of loan products, and the loan approval probability calculation unit 142 calculates the attribute information of the user whose loan approval probability is to be calculated. When specified, the loan approval probability and borrowable amount are calculated for multiple loan products using the loan approval evaluation model stored in the loan approval evaluation model creation unit 141, and the loan approval probability calculation unit 142 calculates the loan approval probability. A plurality of calculated loan products are ranked and output based on predetermined evaluation items.
図７は、複数のローン商品についてランキングして表示する一例を示したイメージ図である。図７の例は、ユーザの利用者端末20に表示する表示例を示したものであり、複数のローン商品毎に、銀行名と、その銀行における融資承認確率の予測結果（Ａ銀行：５０％、Ｂ銀行：８０％、Ｃ銀行：８０％）と、ローン条件（金利、毎月返済額、団体信用生命保険、等）の情報についてランキング表示するようにしている。ランキングの順位付けについては、例えば、各銀行の金利（金利の低い順等）、団体信用生命保険の充実度、融資承認確率、等の情報を指標とした評価項目に応じて適宜に更新するようにしている。すなわち、様々な要因により各銀行の金利等の条件が更新される場合があり、その際にはタイムリーに更新するようにしている。また、融資承認確率の表示については、例えば、２０％から８０％の範囲で１０％刻みに表示するようにしている。さらにAI診断結果として融資承認確率を受けてユーザが取るべき行動（ローンを申し込むべきかどうか）やAIアドバイスとして融資承認確率を上げるためのアドバイスが提示される。これらのインストラクションに従うことでユーザはより有利な条件のローンを自然に選ぶことができる。 FIG. 7 is an image diagram showing an example of ranking and displaying a plurality of loan products. The example in FIG. 7 shows a display example displayed on the user terminal 20 of the user, and shows the bank name and the predicted loan approval probability for that bank (Bank A: 50%) for each of multiple loan products. , Bank B: 80%, Bank C: 80%) and information on loan conditions (interest rate, monthly repayment amount, group credit life insurance, etc.) are displayed in rankings. The rankings will be updated as appropriate according to evaluation items using information such as each bank's interest rate (lowest interest rate, etc.), degree of group credit life insurance coverage, loan approval probability, etc. I have to. That is, conditions such as interest rates of each bank may be updated due to various factors, and in such cases, we try to update them in a timely manner. Furthermore, the loan approval probability is displayed in 10% increments within the range of 20% to 80%, for example. Furthermore, the user is presented with the actions they should take (whether or not to apply for a loan) based on the probability of loan approval as an AI diagnosis result, as well as advice on how to increase the probability of loan approval as AI advice. By following these instructions, users can naturally choose loans with more favorable terms.
以上から、金融機関に住宅ローン等のローン商品の申し込みを行う際に、ユーザの利用者端末から事前にローン商品について高い精度の融資承認確率および借入可能額を知ることができるようになり、また複数のローン商品のランキング表示により、ユーザは、どのローン商品が最も有利な条件で借り入れできるのか、あるいは金融機関の審査にどれくらいの確率で通るのか、さらにローンを申し込むべきかどうか、また、どうすればより高い確率でより有利な条件のローン商品が借りられるか、等について高い精度の情報を事前に知ることが可能になりローン商品の選択を効率的に行えるようになる。 From the above, when applying for a loan product such as a home loan to a financial institution, it is now possible to know the highly accurate loan approval probability and loan amount for the loan product from the user's user terminal, and By displaying the rankings of multiple loan products, users can find out which loan product has the most advantageous terms, how likely it is that they will be approved by a financial institution, and whether or not they should apply for a loan. It becomes possible to know in advance highly accurate information such as whether a loan product with more advantageous conditions can be borrowed with a high probability, and it becomes possible to efficiently select a loan product.
次に本発明の実施形態に係る融資承認確率算出装置のデータ処理手順について、図３のフローチャート図を参照しながら説明する。上記で説明した通り、ユーザは住宅ローン等のローン商品の申し込みを行う前に、利用者端末20から融資承認確率算出装置10にアクセスしユーザ属性情報を入力することにより、融資承認確率算出装置10では、金融機関の審査にどれくらいの確率で通るかという融資承認確率や借入可能額の予測結果を算出し、その予測結果を利用者端末20に表示する。融資承認確率や借入可能額については、金融機関により異なるため、ユーザはその予測結果を基にどの金融機関のローン商品を選択するのが最適か判断してローン商品の申し込みを行うことになる。 Next, the data processing procedure of the loan approval probability calculation device according to the embodiment of the present invention will be explained with reference to the flowchart of FIG. As explained above, before applying for a loan product such as a home loan, the user accesses the loan approval probability calculation device 10 from the user terminal 20 and inputs user attribute information. Then, the prediction result of the loan approval probability and the borrowable amount is calculated, and the prediction result is displayed on the user terminal 20. Since the loan approval probability and loan amount differ depending on the financial institution, the user will judge which financial institution's loan product is best to select based on the prediction result and apply for the loan product.
＜データ処理手順＞
ステップS110において、ユーザが住宅ローン等のローン商品の申し込みを行うために利用者端末20から入力したユーザ属性情報を受け付ける。
<Data processing procedure>
In step S110, user attribute information input by the user from the user terminal 20 in order to apply for a loan product such as a home loan is accepted.
次にステップS120において、ステップS110で入力されたユーザに関する個人信用情報を取得する。個人信用情報を取得する方法としては、例えばＪＩＣＣ（株式会社日本信用情報機構）を利用して取得する方法がある。 Next, in step S120, the personal credit information regarding the user input in step S110 is acquired. As a method of acquiring personal credit information, for example, there is a method of acquiring it using JICC (Japan Credit Information Corporation).
次にステップS130において、ステップS120により取得されたユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する。ここで個人信用情報条件とは、異常値をはじくための条件（例えば、ネガ情報等）であり、ユーザのローン申請をこの条件でフィルタリングして、この条件を充足しないローン申請をはじく処理を行っており、この条件をクリアした場合に次のステップへ進む。ここで、図３のフローチャートには図示していないが、ローン申請がはじかれた場合には、その情報がユーザの利用者端末20に送信される。あるいは、すべての銀行の融資承認確率を一律20%としたランキングが、ユーザの利用者端末20に表示されてもよい。 Next, in step S130, it is determined whether the user's personal credit information acquired in step S120 satisfies predetermined personal credit information conditions. Here, personal credit information conditions are conditions for rejecting abnormal values (for example, negative information, etc.), and the process is to filter user loan applications using these conditions and reject loan applications that do not meet these conditions. If this condition is cleared, proceed to the next step. Although not shown in the flowchart of FIG. 3, if the loan application is rejected, the information is sent to the user terminal 20 of the user. Alternatively, a ranking may be displayed on the user terminal 20 of the user, with the loan approval probability of all banks being uniformly 20%.
次にステップS135において、ＡＩモデルを利用してローン商品の融資承認確率および借入可能額を予測するにあたり、次のステップ140である融資承認機械学習処理に投入するデータを整備する。この投入するデータには、ステップS110で受け付けたユーザ属性情報、ステップS120で取得した個人信用情報に加えて、返済比率、年収倍率、金融機関の情報、等の情報（これらは独自のノウハウ情報であってもよい）が含まれる。
またユーザ属性情報および個人信用情報として、例えば、次のような情報が含まれる。
ユーザ属性情報＝｛年齢、年収、業種、職種、雇用形態、勤務先規模、家族構成、自己資金、住宅ローン以外の負債、物件所在地、保有資産・・・｝
個人信用情報＝｛無担保返済額、消費者ローン件数、30日未満延滞回数、30日以上延滞回数、グレー債権数、ネガ債権数、無担保返済残高、有担保返済残高、・・・｝
Next, in step S135, data to be input into the loan approval machine learning process, which is the next step 140, is prepared in order to predict the loan approval probability and borrowable amount of the loan product using the AI model. In addition to the user attribute information received in step S110 and the personal credit information acquired in step S120, this input data includes information such as repayment ratio, annual income ratio, and financial institution information (these are unique know-how information). ) may be included.
Further, the user attribute information and personal credit information include, for example, the following information.
User attribute information = {Age, annual income, industry, occupation, employment type, workplace size, family structure, own funds, debts other than mortgage, property location, owned assets...}
Personal credit information = {unsecured repayment amount, number of consumer loans, number of delinquencies of less than 30 days, number of delinquencies of 30 days or more, number of gray loans, number of negative receivables, unsecured repayment balance, secured repayment balance,...}
次にステップS140において、ステップS135で整備したデータを、事前に構築されたＡＩモデルに投入して融資承認機械学習処理を行い融資承認確率および借入可能額を算出する。ここでＡＩモデルは、ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルとして作成されるものであり、事前に学習データをＡＩに投入してモデルが構築される。この学習データには、上記で説明した通り、ユーザ属性情報に加え、個人信用情報、審査を行った金融機関、借入希望額、審査結果等の情報が含まれる。 Next, in step S140, the data prepared in step S135 is input into an AI model built in advance, and loan approval machine learning processing is performed to calculate the loan approval probability and the borrowable amount. Here, the AI model uses the attribute information used in the approval screening of loan products and the results of the approval screening based on the attribute information to evaluate the loan approval by associating the characteristics of each index of the attribute information with the loan approval probability. It is created as a model, and the model is constructed by inputting learning data into AI in advance. As explained above, in addition to the user attribute information, this learning data includes information such as personal credit information, the financial institution that conducted the examination, the desired loan amount, and the examination results.
次にステップS180において、ステップS140で算出されたローン商品の融資承認確率および借入可能額について、返済比率ディスカウント等の情報（これらは独自のノウハウ情報であってもよい）を加味したパラメータ調整を実行する。パラメータ調整をしない場合には、ステップS180を飛ばして次のステップS200に進んでもよい。 Next, in step S180, parameter adjustment is performed on the loan approval probability and borrowable amount of the loan product calculated in step S140, taking into account information such as repayment ratio discount (these may be unique know-how information). do. If parameter adjustment is not to be performed, step S180 may be skipped and the process may proceed to the next step S200.
次にステップS200において、ステップS180でパラメータ調整されたデータ（ローン商品の融資承認確率および借入可能額）についてローン商品の申し込みを行ったユーザの利用者端末20に送信する（ステップS180を飛ばした場合には、ステップS140で算出されたローン商品の融資承認確率および借入可能額についてローン商品の申し込みを行ったユーザの利用者端末20に送信する）。図５は、融資承認確率の予測結果をユーザの利用者端末20に表示する一例を示したイメージ図であり、図５の例では、銀行名と、その銀行における融資承認確率の予測結果８０％と、ローン条件（金利、毎月返済額、団体信用生命保険、等）の情報、融資承認確率を受けてユーザが取るべきアクションと融資承認確率を上げるために何をしなければならないかを表示するようにしている。また図６は、借入可能額の予測結果をユーザの利用者端末20に表示する一例を示したイメージ図であり、図６の例では、借入可能額と、適用金利の情報が表示するようにしている。 Next, in step S200, the data whose parameters were adjusted in step S180 (loan approval probability and borrowable amount) is sent to the user terminal 20 of the user who applied for the loan product (if step S180 is skipped) (The loan approval probability and loan amount calculated in step S140 are transmitted to the user terminal 20 of the user who applied for the loan product.) FIG. 5 is an image diagram showing an example of displaying the prediction result of the loan approval probability on the user's user terminal 20. In the example of FIG. 5, the bank name and the prediction result of the loan approval probability for that bank are 80%. , information on loan conditions (interest rate, monthly repayment amount, group credit life insurance, etc.), actions that the user should take in response to the loan approval probability, and what to do to increase the loan approval probability are displayed. I have to. Further, FIG. 6 is an image diagram showing an example of displaying the prediction result of the borrowable amount on the user terminal 20 of the user. In the example of FIG. 6, information on the borrowable amount and the applicable interest rate is displayed. There is.
また図３のフローチャートには図示していないが、複数のローン商品の融資承認確率および借入可能額について所定の評価項目によりランキングして出力することもできる。すなわち、ユーザの属性情報が指定されたときに、複数のローン商品について融資承認評価モデルにより融資承認確率および借入可能額を算出し、複数のローン商品について、所定の評価項目によりランキングして出力する。 Although not shown in the flowchart of FIG. 3, it is also possible to rank and output the loan approval probabilities and loanable amounts of a plurality of loan products based on predetermined evaluation items. That is, when user attribute information is specified, the loan approval probability and borrowable amount are calculated for multiple loan products using a loan approval evaluation model, and the multiple loan products are ranked and output based on predetermined evaluation items. .
図７は、複数のローン商品についてランキングして表示する一例を示したイメージ図であり、図７の例は、ユーザの利用者端末20に表示する表示例を示したものであり、複数のローン商品毎に、銀行名と、その銀行における融資承認確率の予測結果（Ａ銀行：５０％、Ｂ銀行：８０％、Ｃ銀行：８０％）と、ローン条件（金利、毎月返済額、団体信用生命保険、等）の情報についてランキング表示するようにしている。ランキングの順位付けについては、例えば、各銀行の金利（金利の低い順等）、団体信用生命保険の充実度、融資承認確率、等の情報を指標とした評価項目に応じて適宜に更新するようにしている。すなわち、様々な要因により各銀行の金利等の条件が更新される場合があり、その際にはタイムリーに更新するようにしている。また、融資承認確率の表示については、例えば、２０％から８０％の範囲で１０％刻みに表示する。さらに融資承認確率を受けてユーザが取るべきアクション（ローンに申し込むべきかどうか等）をAI診断結果として表示すると共に融資承認確率を上げるために何をしなければならないかをAIアドバイスとして表示する。 FIG. 7 is an image diagram showing an example of ranking and displaying a plurality of loan products. For each, the name of the bank, the prediction result of the loan approval probability at that bank (Bank A: 50%, Bank B: 80%, Bank C: 80%), and loan conditions (interest rate, monthly repayment amount, group credit life insurance). , etc.) are displayed in rankings. The rankings will be updated as appropriate according to evaluation items using information such as each bank's interest rate (lowest interest rate, etc.), degree of group credit life insurance coverage, loan approval probability, etc. I have to. That is, conditions such as interest rates of each bank may be updated due to various factors, and in such cases, we try to update them in a timely manner. Furthermore, the loan approval probability is displayed in 10% increments within the range of 20% to 80%, for example. Furthermore, the actions that the user should take in response to the loan approval probability (such as whether or not to apply for a loan) are displayed as AI diagnosis results, and AI advice on what to do to increase the loan approval probability is displayed.
なお、上記で説明したデータ処理手順については、コンピュータによって実行される方法として実現されてもよいし、またコンピュータに実行されるためのプログラムとして実現されてもよい。 Note that the data processing procedure described above may be realized as a method executed by a computer, or may be realized as a program executed by a computer.
10 融資承認確率算出装置
20 利用者端末
110 ユーザ属性情報入力部
120 個人信用情報取得部
130 個人信用情報判定部
140 融資承認機械学習部
141 融資承認評価モデル作成部
142 融資承認確率算出部
150 ローン商品記憶部
160 ユーザ情報記憶部
170 融資承認結果記憶部
10 Loan approval probability calculation device
20 User terminal
110 User attribute information input section
120 Personal credit information acquisition department
130 Personal credit information evaluation department
140 Loan Approval Machine Learning Department
141 Loan approval evaluation model creation department
142 Loan approval probability calculation department
150 Loan product storage section
160 User information storage unit
170 Loan approval result storage unit
Claims (4)
前記ユーザの属性情報の入力を受け付けるユーザ属性情報入力部と、
前記ユーザの個人信用情報を取得する個人信用情報取得部と、
前記個人信用情報取得部により取得された前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する個人信用情報判定部と、
前記個人信用情報判定部により前記個人信用情報条件を充足すると判定されたユーザに対して、前記ユーザ属性情報入力部により入力を受け付けた該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を予測する融資承認機械学習部と、
を備え、
前記融資承認機械学習部は、
前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存する融資承認評価モデル作成部と、
前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、前記融資承認評価モデル作成部で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出すると共に、その確率を算出した該属性情報における指標の特徴を出力する融資承認確率算出部と
を有することを特徴とする融資承認確率算出装置。 A loan approval probability calculation device that calculates a loan approval probability and a loanable amount for a loan product from user attribute information, the device comprising:
a user attribute information input unit that accepts input of the user's attribute information;
a personal credit information acquisition unit that acquires personal credit information of the user;
a personal credit information determination unit that determines whether the user's personal credit information acquired by the personal credit information acquisition unit satisfies predetermined personal credit information conditions;
For a user who is determined by the personal credit information determination unit to satisfy the personal credit information conditions, the loan product is approved based on the attribute information of the user whose input is accepted by the user attribute information input unit. a loan approval machine learning unit that predicts the probability and loan amount;
Equipped with
The loan approval machine learning department is
From the attribute information used in the approval examination of the loan product and the result of the approval examination based on the attribute information, a loan approval evaluation model is created that associates the characteristics of each index of the attribute information with the loan approval probability. A loan approval evaluation model creation department to save;
When the attribute information of the user for which the loan approval probability of the loan product is to be calculated is specified, the loan approval evaluation model stored by the loan approval evaluation model creation unit is based on a plurality of indicators of the attribute information. A loan approval probability calculation device comprising: a loan approval probability calculation unit that calculates a loan approval probability and a borrowable amount, and outputs characteristics of an index in the attribute information for which the probability has been calculated.
前記融資承認機械学習部は、
融資承認評価モデル作成部が、複数の前記ローン商品について、融資承認評価モデルを作成して保存し、
前記融資承認確率算出部が、融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、複数の前記ローン商品について、前記融資承認評価モデル作成部で保存された融資承認評価モデルにより融資承認確率および借入可能額を算出し、
前記融資承認確率算出部により融資承認確率が算出された複数の前記ローン商品について、所定の評価項目によりランキングして出力することを特徴とする融資承認確率算出装置。 The loan approval probability calculation device according to claim 1,
The loan approval machine learning department is
a loan approval evaluation model creation unit creates and saves loan approval evaluation models for the plurality of loan products;
When the attribute information of the user for which the loan approval probability calculation unit is to calculate the loan approval probability is specified, the loan approval evaluation model stored by the loan approval evaluation model creation unit for a plurality of loan products is determined by the loan approval probability calculation unit. Calculate the loan approval probability and loan amount by
A loan approval probability calculation device, characterized in that the plurality of loan products for which loan approval probabilities have been calculated by the loan approval probability calculation unit are ranked and output based on predetermined evaluation items.
ユーザ属性情報入力部が、前記ユーザの属性情報の入力を受け付けるユーザ属性情報入力工程と、
個人信用情報取得部が、前記ユーザの個人信用情報を取得する個人信用情報取得工程と、
個人信用情報判定部が、前記個人信用情報取得工程により取得された前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定する個人信用情報判定工程と、
前記個人信用情報判定工程により前記個人信用情報条件を充足すると判定されたユーザに対して、融資承認機械学習部が、前記ユーザ属性情報入力工程により入力を受け付けた該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を予測する融資承認機械学習処理工程と
を備え、
前記融資承認機械学習処理工程は、
融資承認評価モデル作成部が、予め、前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存しておき、
融資承認確率算出部が、前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、保存された前記融資承認評価モデルにより融資承認確率および借入可能額を算出すると共に、その確率を算出した該属性情報における指標の特徴を出力することを特徴とする融資承認確率の算出方法。 A method for calculating a loan approval probability for calculating a loan approval probability and a borrowable amount for a loan product from user attribute information, the method comprising:
a user attribute information input step in which the user attribute information input unit receives input of the user's attribute information;
a personal credit information acquisition step in which the personal credit information acquisition unit acquires personal credit information of the user;
a personal credit information determining step in which the personal credit information determining unit determines whether the user's personal credit information acquired in the personal credit information acquiring step satisfies predetermined personal credit information conditions;
For a user who is determined to satisfy the personal credit information conditions in the personal credit information determination step, the loan approval machine learning unit performs a loan approval process based on the attribute information of the user whose input was accepted in the user attribute information input step. , a loan approval machine learning processing step for predicting the loan approval probability and borrowable amount of the loan product,
The loan approval machine learning processing step includes:
The loan approval evaluation model creation unit corresponds in advance the characteristics of each index of the attribute information and the loan approval probability from the attribute information used for the approval screening of the loan product and the result of the approval screening based on the attribute information. Create and save a loan approval evaluation model with
When the attribute information of the user for which the loan approval probability calculation unit should calculate the loan approval probability of the loan product is specified, the loan approval probability calculation unit calculates the saved loan approval evaluation model based on a plurality of indicators of the attribute information. A method for calculating a loan approval probability, comprising calculating a loan approval probability and a borrowable amount, and outputting characteristics of an index in the attribute information for which the probability is calculated.
コンピュータに、
属性情報が入力された前記ユーザの個人信用情報を取得させ、
取得した前記ユーザの個人信用情報が所定の個人信用情報条件を充足するか判定させ、
個人信用情報条件を充足すると判定された前記ユーザに対して、入力された該ユーザの前記属性情報に基づいて、前記ローン商品の融資承認確率および借入可能額を融資承認機械学習処理により算出させるプログラムであって、
前記融資承認機械学習処理は、
予め、前記ローン商品の承認審査に使用された属性情報と該属性情報による該承認審査の結果とから、該属性情報の各指標における特徴と融資承認確率とを対応付けた融資承認評価モデルを作成して保存させておき、
前記ローン商品の融資承認確率の算出を行うべき前記ユーザの属性情報が指定されたときに、該属性情報の複数の指標に基づいて、保存された前記融資承認評価モデルにより融資承認確率および借入可能額を算出させると共に、その確率を算出した該属性情報における指標の特徴を出力させることを特徴とするプログラム。 A program that calculates the loan approval probability and borrowable amount of a loan product from user attribute information,
to the computer,
obtain personal credit information of the user whose attribute information has been input;
determining whether the acquired personal credit information of the user satisfies predetermined personal credit information conditions;
A program that causes the user who has been determined to satisfy the personal credit information conditions to calculate the loan approval probability and borrowable amount for the loan product using loan approval machine learning processing based on the input attribute information of the user. And,
The loan approval machine learning process is
In advance, from the attribute information used in the approval examination of the loan product and the result of the approval examination based on the attribute information, a loan approval evaluation model is created that associates the characteristics of each index of the attribute information with the loan approval probability. and save it,
When the attribute information of the user for which the loan approval probability of the loan product is to be calculated is specified, the loan approval probability and borrowability are calculated based on the stored loan approval evaluation model based on a plurality of indicators of the attribute information. A program characterized in that it causes an amount to be calculated and outputs characteristics of an index in the attribute information for which the probability thereof has been calculated.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2021155320A JP7366355B2 (en) | 2021-09-24 | 2021-09-24 | Loan approval probability calculation device, loan approval probability calculation method and program |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| JP2021155320A JP7366355B2 (en) | 2021-09-24 | 2021-09-24 | Loan approval probability calculation device, loan approval probability calculation method and program |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| JP2023046621A JP2023046621A (en) | 2023-04-05 |
| JP7366355B2 true JP7366355B2 (en) | 2023-10-23 |
Family
ID=85778483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| JP2021155320A Active JP7366355B2 (en) | 2021-09-24 | 2021-09-24 | Loan approval probability calculation device, loan approval probability calculation method and program |
Country Status (1)
| Country | Link |
|---|---|
| JP (1) | JP7366355B2 (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117995199A (en) * | 2024-03-22 | 2024-05-07 | 中邮消费金融有限公司 | Loan contract voice signing method, device and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014225253A (en) | 2013-04-26 | 2014-12-04 | スルガ銀行株式会社 | Recommended credit frame calculation device |
| JP2017049673A (en) | 2015-08-31 | 2017-03-09 | 株式会社三菱総合研究所 | Information processing apparatus and information processing method |
| JP2020102013A (en) | 2018-12-21 | 2020-07-02 | 株式会社メルカリ | Method for processing information, program, and information processor |
| US20200265512A1 (en) | 2019-02-20 | 2020-08-20 | HSIP, Inc. | System, method and computer program for underwriting and processing of loans using machine learning |
| US20200334719A1 (en) | 2019-04-19 | 2020-10-22 | Capital One Services, Llc | Augmented intelligence leading to higher deal acceptance for dealerships |
| JP2022167355A (en) | 2021-04-23 | 2022-11-04 | Bhi株式会社 | Information processing apparatus, information processing method, and program |
-
2021
- 2021-09-24 JP JP2021155320A patent/JP7366355B2/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014225253A (en) | 2013-04-26 | 2014-12-04 | スルガ銀行株式会社 | Recommended credit frame calculation device |
| JP2017049673A (en) | 2015-08-31 | 2017-03-09 | 株式会社三菱総合研究所 | Information processing apparatus and information processing method |
| JP2020102013A (en) | 2018-12-21 | 2020-07-02 | 株式会社メルカリ | Method for processing information, program, and information processor |
| US20200265512A1 (en) | 2019-02-20 | 2020-08-20 | HSIP, Inc. | System, method and computer program for underwriting and processing of loans using machine learning |
| US20200334719A1 (en) | 2019-04-19 | 2020-10-22 | Capital One Services, Llc | Augmented intelligence leading to higher deal acceptance for dealerships |
| JP2022167355A (en) | 2021-04-23 | 2022-11-04 | Bhi株式会社 | Information processing apparatus, information processing method, and program |
Non-Patent Citations (1)
| Title |
|---|
| "ローン比較 スマホで手軽に クラウドローン、地銀と連携"，月刊ｅ・コロンブス，第４７巻，東方通信社，2021年05月31日，第１９－２０ページ |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2023046621A (en) | 2023-04-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kozodoi et al. | Fairness in credit scoring: Assessment, implementation and profit implications | |
| US8515862B2 (en) | Computer-implemented systems and methods for integrated model validation for compliance and credit risk | |
| US7392216B1 (en) | Methods and apparatus for utilizing a proportional hazards model to evaluate loan risk | |
| US8401950B2 (en) | Optimizing portfolios of financial instruments | |
| US6951008B2 (en) | Evidential reasoning system and method | |
| US20140200930A1 (en) | Methods and Systems for Determining the Importance of Individual Variables in Statistical Models | |
| US20070124236A1 (en) | Credit risk profiling method and system | |
| Höglund | Tax payment default prediction using genetic algorithm-based variable selection | |
| CN107798607A (en) | Asset Allocation strategy acquisition methods, device, computer equipment and storage medium | |
| CA3081569A1 (en) | Entity segmentation for analysis of sensitivities to potential disruptions | |
| US20230252569A1 (en) | Computer Implemented Method for Compiling a Portfolio of Assets | |
| KR102092461B1 (en) | Method for providing investment success rate | |
| Eddy et al. | Credit scoring models: Techniques and issues | |
| JP7366355B2 (en) | Loan approval probability calculation device, loan approval probability calculation method and program | |
| KR102174608B1 (en) | Apparatus for predicting loan defaults based on machine learning and method thereof | |
| US20150269670A1 (en) | Method, system, and apparatus for semi-automatic risk and automatic targeting and action prioritization in loan monitoring applications | |
| Teles et al. | Classification methods applied to credit scoring with collateral | |
| Yaroshchuk et al. | Credit scoring model for microfinance organizations | |
| CN113487440A (en) | Model generation method, health insurance claim settlement determination method, device, equipment and medium | |
| Selau et al. | A systematic approach to construct credit risk forecast models | |
| Makuch | The basics of a better application score | |
| JP7331944B2 (en) | Learning system, learning method, appropriate interest rate prediction system, appropriate interest rate prediction method, program, and loan matching system | |
| CN113362179B (en) | Method, apparatus, device, storage medium and program product for predicting transaction data | |
| CN117151819B (en) | Transaction user risk recommendation method based on data analysis | |
| US20240020615A1 (en) | System and method for evaluating digital service |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| A621 | Written request for application examination |
Free format text: JAPANESE INTERMEDIATE CODE: A621 Effective date: 20220708 |
|
| RD04 | Notification of resignation of power of attorney |
Free format text: JAPANESE INTERMEDIATE CODE: A7424 Effective date: 20221031 |
|
| A977 | Report on retrieval |
Free format text: JAPANESE INTERMEDIATE CODE: A971007 Effective date: 20230728 |
|
| A131 | Notification of reasons for refusal |
Free format text: JAPANESE INTERMEDIATE CODE: A131 Effective date: 20230801 |
|
| A521 | Request for written amendment filed |
Free format text: JAPANESE INTERMEDIATE CODE: A523 Effective date: 20230830 |
|
| TRDD | Decision of grant or rejection written | ||
| A01 | Written decision to grant a patent or to grant a registration (utility model) |
Free format text: JAPANESE INTERMEDIATE CODE: A01 Effective date: 20230905 |
|
| A61 | First payment of annual fees (during grant procedure) |
Free format text: JAPANESE INTERMEDIATE CODE: A61 Effective date: 20230928 |
|
| R150 | Certificate of patent or registration of utility model |
Ref document number: 7366355 Country of ref document: JP Free format text: JAPANESE INTERMEDIATE CODE: R150 |