KR20210035476A - Method, apparatus and program for evaluating drivers based on accident tendency - Google Patents

Abstract

As a method performed by a computer, a method for evaluating a driver based on an accident tendency comprises the following steps of: obtaining data recorded for a first driver; extracting one or more accident data from the obtained data; determining one or more accident tendencies for the first driver based on the one or more accident data; and obtaining evaluation information about the first driver based on an accident tendency of the first driver.

Description

METHOD, APPARATUS AND PROGRAM FOR EVALUATING DRIVERS BASED ON ACCIDENT TENDENCY}

The present invention relates to a method of evaluating a driver based on an accident tendency.

In car insurance, it is very important to estimate the probability of a driver causing an accident and the amount of compensation accordingly. This is directly related to the profits of insurance products, and a deficit may occur if the forecast is incorrect.

Insurance companies estimate this in various ways and use it to determine insurance payments for each driver, mainly referring to each driver's personal information, vehicle type, and accident status. Recently, in some cases, information on the driver's driving tendency is collected and analysis is performed based on the information.

In addition, insurance companies and insurance management companies use computer systems to manage information on insurance subscribers and insurance products, but if errors or omissions occur in the computer system, a major problem may occur, so it is required to prepare a means to prevent this. do.

Furthermore, there is a need to develop a technology that enables various statistical analysis to be performed based on the information recorded in the computer system, and allows the driver to be evaluated based on standards different from the existing ones.

The problem to be solved by the present invention is to provide a method of evaluating a driver based on an accident tendency.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems that are not mentioned will be clearly understood by those skilled in the art from the following description.

A method of evaluating a driver based on an accident tendency according to an aspect of the present invention for solving the above-described problem comprises: obtaining recorded data for a first driver, and extracting one or more accident data from the obtained data. And determining at least one accident tendency for the first driver based on the at least one accident data, and obtaining evaluation information for the first driver based on the accident tendency of the first driver can do.

In addition, the accident tendency may include information on whether the first driver has an accident, an accident frequency, and a content of the accident.

In addition, the accident propensity includes at least one of the average negligence rate of the first driver, the accident cycle, the amount of damage in the past year, the number of high-income accidents, the number of serious accidents, fine information, the degree of separation from the accident site, and the degree of late-night accidents, and the average The error rate is obtained based on the average of the error rates of accidents that occurred while the first driver is driving, and the accident cycle is obtained based on the number of accidents compared to the contract elapsed period of the first driver, and the elapsed year The amount of damage is obtained based on the total amount of damage compared to the contract elapsed period of the first driver, the number of high-priced accidents is obtained based on the number of high-priced accidents exceeding a preset criterion, and the number of serious accidents is a preset item. Means the number of serious accidents corresponding to, and the penalty information includes the number of fines corresponding to a preset item and its contents, and the accident location separation degree is based on the accident sequence of accidents related to the first driver. It is obtained based on a result of calculating each separation degree, and the late-night accident degree may include information on an accident occurring at a preset late-night time.

In addition, the obtaining of the evaluation information further includes calculating an evaluation score for the first driver based on an accident tendency toward the first driver, wherein the evaluation score is an average of the first driver. A score calculated based on the preset scoring criteria for at least one of negligence rate, accident cycle, elapsed year damage, high-income accidents, serious accidents, fine information, accident location separation and late-night accidents, and the first driver's It may be characterized in that it is obtained by synthesizing scores calculated based on personal information.

In addition, acquiring information on a preset deep warning item, searching whether there is at least one item corresponding to the deep warning item in the information on the first driver, and the step of searching for information on the first driver from the information on the first driver. When one or more items corresponding to the deep warning item are searched, the step of providing a notification corresponding thereto may be further included.

In addition, the obtaining of the evaluation information includes extracting one or more parameters from information on the accident tendency of the first driver, inputting the extracted parameters to a pre-learned artificial intelligence model, and the artificial intelligence model. It may further include the step of calculating a predicted loss amount for the first driver from the output of.

In addition, the calculating of the predicted loss may include predicting the number of accidents of the first driver, predicting the type of accident of the first driver, and calculating the predicted cost according to the type of accident of the first driver. And calculating a predicted amount of damage to the first driver based on the number of accidents and the predicted cost of the first driver.

In addition, obtaining revision information of one or more laws or regulations related to calculating the estimated damage amount, obtaining learning data used for learning the artificial intelligence model, and modifying the contents of the learning data based on the revision information. It may further include augmenting the learning data reflecting the revision information, and retraining the artificial intelligence model using the modified and augmented learning data.

An apparatus according to an aspect of the present invention for solving the above-described problem includes a memory for storing one or more instructions, and a processor for executing the one or more instructions stored in the memory, and the processor executes the one or more instructions. By doing so, a method of evaluating a driver based on an accident tendency according to the disclosed embodiment is performed.

A computer program according to an aspect of the present invention for solving the above-described problems is combined with a computer that is hardware, and a computer-readable record to perform a method of evaluating a driver based on an accident tendency according to the disclosed embodiment. It is stored on the medium.

Other specific details of the present invention are included in the detailed description and drawings.

According to the disclosed embodiment, there is a differentiated effect of providing a method for evaluating a driver based on specific factors such as accident propensity, as well as simple indicators such as the number of accidents, whether or not there is an accident, and the amount of compensation.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the following description.

1 is a diagram illustrating a system according to an exemplary embodiment.
2 is a flow chart illustrating a contract management of automobile insurance and a method for verifying errors and omissions therefor according to an exemplary embodiment.
3 is a diagram illustrating an example of a screen for verifying an error of an update case.
4 is a diagram illustrating an example of a screen for verifying an error of an installment case.
5 is a diagram illustrating an example of a performance-based verification screen.
6 is a diagram illustrating an example of a screen for verifying whether data is missing.
7 is a flowchart illustrating a statistical analysis method according to an exemplary embodiment.
8 and 9 are diagrams illustrating an example of a screen for searching for statistical information according to a contract type and providing statistical information according to the search.
10 and 11 are diagrams illustrating an example of a screen providing statistical information search and statistical information according to an accident type.
12 is a flowchart illustrating a method of evaluating a driver based on an accident tendency, according to an exemplary embodiment.
13 is a diagram illustrating a method of learning an artificial intelligence model according to an exemplary embodiment.
14 is a block diagram of an apparatus according to an exemplary embodiment.

Advantages and features of the present invention, and a method of achieving them will become apparent with reference to the embodiments described below in detail together with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in a variety of different forms. It is provided to fully inform the skilled person of the scope of the present invention, and the present invention is only defined by the scope of the claims.

The terms used in the present specification are for describing exemplary embodiments and are not intended to limit the present invention. In this specification, the singular form also includes the plural form unless specifically stated in the phrase. As used herein, “comprises” and/or “comprising” do not exclude the presence or addition of one or more other elements other than the mentioned elements. Throughout the specification, the same reference numerals refer to the same elements, and "and/or" includes each and all combinations of one or more of the mentioned elements. Although "first", "second", and the like are used to describe various elements, it goes without saying that these elements are not limited by these terms. These terms are only used to distinguish one component from another component. Therefore, it goes without saying that the first component mentioned below may be the second component within the technical idea of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in the present specification may be used with meanings that can be commonly understood by those of ordinary skill in the art to which the present invention belongs. In addition, terms defined in a commonly used dictionary are not interpreted ideally or excessively unless explicitly defined specifically.

The term "unit" or "module" used in the specification refers to a hardware component such as software, FPGA or ASIC, and the "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. The "unit" or "module" may be configured to be in an addressable storage medium or may be configured to reproduce one or more processors. Thus, as an example, "sub" or "module" refers to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, Includes procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays and variables. Components and functions provided within "sub" or "modules" may be combined into a smaller number of components and "sub" or "modules" or into additional components and "sub" or "modules". Can be further separated.

In the present specification, a computer means all kinds of hardware devices including at least one processor, and may be understood as encompassing a software configuration operating in the corresponding hardware device according to embodiments. For example, the computer may be understood as including all of a smartphone, a tablet PC, a desktop, a laptop, and a user client and application running on each device, and is not limited thereto.

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

Each of the steps described herein is described as being performed by a computer, but the subject of each step is not limited thereto, and at least some of the steps may be performed by different devices according to embodiments.

1 is a diagram illustrating a system according to an exemplary embodiment.

Referring to FIG. 1, a server 100, an insurance company server 10, a management terminal 20, and a user terminal 30 performing a method according to the disclosed embodiment are illustrated.

Methods according to the disclosed embodiment may be performed by the server 100 and the management terminal 20, and at least some or all of the steps included in the method according to the disclosed embodiment are the server 100 and the management terminal 20 It can be carried out separately or can be carried out by dividing different steps.

However, this is disclosed as an example, and the subject of the method according to the disclosed embodiment is not limited thereto.

The management terminal 20 and the server 100 may refer to a terminal and a server operated by a company that brokers and manages automobile insurance. The method according to the disclosed embodiment is performed by the management terminal 20 and the server 100, and relates to a method for brokering and managing automobile insurance, but is not limited thereto.

The insurance company server 10 is an insurance company server that provides insurance services managed by the server 100, and stores information on insurance products, information on insurance subscribers, management information after insurance subscription, and the like.

The server 100 also stores information on insurance products, information on insurance subscribers, management information after insurance subscription, etc., which are obtained from the user terminal 30, the management terminal 20, or the like, or stored in the server 100. It may be created and managed by, or may be acquired and managed from the insurance company server 10.

The server 100 must manage information input from a terminal or acquired from an insurance company server without errors or omissions, and therefore, it is necessary to provide a verification method and an interface for this.

That is, the server 100 has its own computer system separate from the insurance company, and by using this, it is possible to provide a system that is more systematic and easier to manage and read than the insurance company computer system.

However, by operating its own computer system, there may be differences from the insurance company's computer system, and to eliminate or minimize this, errors between the insurance company's computer system and its own computer system are verified at predetermined intervals (e.g., monthly). Accordingly, it can be managed so that errors or omissions do not occur.

In addition, in the case of automobile insurance, since it has a system that is renewed every year, a means to manage renewal cases is required, and changes must be reflected upon renewal. It is necessary to provide a system to prevent errors or omissions in this update step.

Hereinafter, a method of managing and verifying information related to automobile insurance in various ways will be described in detail with reference to FIGS. 2 to 6.

2 is a flow chart illustrating a contract management of automobile insurance and a method for verifying errors and omissions therefor according to an exemplary embodiment.

In step S110, the computer performs an error verification step for the update case.

Referring to FIG. 3, an example of a screen for verifying an error of an update case is shown. Referring to the search unit 210 illustrated in FIG. 3, a search field capable of searching data based on an insurance company, verification presence, error status, and reference year is included.

In addition, referring to the verification unit 212 illustrated in FIG. 3, a field for verifying the contents of the vehicle number, insurance company, and insurance term is included.

In an embodiment, the computer may verify the wrong part by comparing the contents including the vehicle number, the insurance company, and the insurance type by comparing the contents of the insurance company computer system based on its own computer system.

In addition, the computer can, on the contrary, verify and correct the wrong part by comparing the vehicle number of its own computer system, the insurance company, and the insurance type based on the insurance company computer system.

The most important thing in car insurance renewal is not to miss the end of the insurance, and to accurately identify the license plate number and insurance company of each insurance case.

Therefore, referring to the screen shown in FIG. 3, a configuration for cross-validating the vehicle number, insurance company, and insurance term is shown. In addition, it is possible to perform error verification of an automobile insurance renewal case through cross-validation of various factors. to be.

In step S120, the computer verifies the error of the installment case.

Referring to FIG. 4, an example of a screen for verifying an error of an installment case is shown. Referring to the search unit 220 illustrated in FIG. 4, a search field capable of searching data based on an insurance company, verification presence, error status, and reference year is included.

In addition, referring to the verification unit 222 illustrated in FIG. 4, a field for verifying the contents of the vehicle number, insurance company, round trip and insurance premium is included.

In an embodiment, the computer compares the contents of the insurance company's computer system based on its own computer system to verify the wrong part by comparing the contents including the vehicle number, the insurance company, the round trip, and the insurance premium.

In addition, the computer, on the contrary, compares the vehicle number, insurance company, turnaround and insurance premium of its own computer system based on the insurance company computer system to verify and correct the wrong part.

The most important thing in the installment management of automobile insurance is to manage the installments and premiums, and to accurately identify the vehicle number and insurance company of each insurance case.

Therefore, referring to the screen shown in FIG. 4, a configuration for cross-validating the vehicle number, insurance company, round trip, and insurance premium is shown.In addition, it is possible to perform error verification of an automobile insurance installment case through cross-validation of various factors. Of course.

In step S130, the computer performs verification based on the insurance company performance information.

Referring to FIG. 5, an example of a performance-based verification screen is shown. Referring to the search unit 230 illustrated in FIG. 5, a search field capable of searching data based on an insurance company, verification presence, error status, and reference year is included.

Referring to the screen shown in FIG. 5, a verification field 232 including contract number, insured person, vehicle number, person in charge, original investigation error, status, error criterion and error classification is shown, through which the above-described update, In addition to installment payments, changes, terminations, and other various types of contracts can be compared and verified.

This verification is performed by comparing the contents of the in-house computer system based on the contents of the insurance company's computer system, and in particular, each insurance premium and contract details are well in their own computer system without problems such as errors or omissions based on the performance information of the insurance company computer system. It can be performed by verifying whether or not it has been input.

In step S140, the computer verifies whether other data are omitted.

Referring to FIG. 6, an example of a screen for verifying whether data is missing is shown. Referring to the search unit 240 of the screen shown in FIG. 6, the insured name, holder code, vehicle number, insurance period, business/resident number, vehicle information, basic information, installment total insurance premium, attached documents, driver and special agreement, A search field including items such as collateral content, tenant information, and insurance period is shown.

Further, referring to the verification field 242, an interface for verifying omission of various items is shown.

That is, the screen shown in FIG. 6 is for checking whether there is any missing data among the data to be input and uploaded during contract inquiry such as automobile insurance, vehicle registration certificate, and subscription certificate of customers (insurance subscribers).

These data are preferably entered and uploaded at the time of sign-up, but some data are delivered late, and this information is necessary for insurance management as well as big data analysis using it, so a process of verifying the omission needs to be performed.

The items for omission verification are not limited, but for example, insurance period, total installment premium, division of head office, basic information, vehicle name, displacement, insurance premium by collateral, age, driver limitation, registration card, membership card, rental type, rental gender/type of business, It may include the name of the tenant, the tenant business/resident number, etc.

The computer may perform the verification steps as shown in FIG. 2, which may be performed based on a user's request, but may be performed automatically in the background at a preset period.

Thereafter, the computer may output the verification results of information corresponding to a specific search word based on the user's request, or separately collect and output only verification results with specific items (eg, errors or omissions), and data and verification The search and output method of results is not limited to a specific method.

The computer manages the insurance contract-related information of its own computer system according to the above-described method, and prevents errors or omissions in the management content.

These data are not only used for the management of insurance contracts, but also various statistical analysis can be performed based on big data including them. The statistical analysis results can be used in various fields such as insurance product design, subscriber management, product recommendation, and improvement of profit rate.

Hereinafter, a statistical analysis method of big data according to the disclosed embodiment will be described in detail with reference to FIGS. 7 to 11.

7 is a flowchart illustrating a statistical analysis method according to an exemplary embodiment.

In step S210, the computer may perform a statistical analysis method according to the contract type.

8 and 9, an example of a screen for searching statistical information according to a contract type and providing statistical information according to the search is illustrated.

Referring to FIG. 8, the search unit 310 shows a search field including a classification, a holder code, a business/resident number, a product name, a detailed product name, a month, a summation for each section, and a reference year and month.

In addition, referring to the condition setting unit 312, it includes age, driver limitation, rental type, insurance company, vehicle type, domestic/foreign, special type, vehicle name, displacement, branch, collateral, rental gender, rental type, rental business type, etc. Fields in which one or more criteria to be set can be set are shown.

As shown in FIGS. 8 and 9, one or a plurality of criteria may be set, and the searched and sorted information according to one or more criteria may be viewed.

The type of information that can be viewed is not limited, but, for example, information related to an automobile accident may be searched, sorted, and provided.

Referring to FIG. 9, a search unit 320, a reference setting unit 322, and a statistical information providing unit 324 are shown. Referring to the statistical information providing unit 324, the number of effective units, the number of accidents, and Information including accident rate, loss ratio, number of gross negligence, gross negligence rate, number of high value accidents, high value accident rate, loss per case, average error rate, etc. is provided by searching and sorting.

Such information can be used to estimate the risk of traffic accidents and the expected amount of damage for the group corresponding to the set criteria.

The estimated information can be used to design insurance products in the future, and can also be used to calculate insurance premiums or to decide whether to approve insurance subscriptions based on the risk of accidents and expected damages of the group to which the subscriber belongs when insured.

In step S220, the computer may perform a statistical analysis method according to the type of accident.

Referring to FIGS. 10 and 11, an example of a screen providing statistical information search and statistical information according to an accident type is illustrated.

Referring to FIG. 10, a search field including a classification, a holder code, a business/resident number, a product name, a detailed product name, and a period setting is shown in the search unit 330.

In addition, referring to the condition setting unit 332, one or more criteria including accident type, accident type, accident driving order, gross negligence accident, accident difference, accident day, total accident amount, accident time slot, accident month, etc. are set. The possible fields are shown.

As shown in FIGS. 10 and 11, one or more criteria may be set, and information that is searched and sorted according to one or more criteria can be viewed.

The type of information that can be viewed is not limited, but, for example, information related to an automobile accident may be searched, sorted, and provided.

Referring to FIG. 11, a search unit 340, a reference setting unit 342, and a statistical information providing unit 344 are shown. Referring to the statistical information providing unit 344, the number of accidents, accident rate, and Information including loss per case, loss ratio, gross negligence, gross negligence rate, high value accident, high value accident rate, and average error rate, etc. is searched and sorted and provided.

Such information can be used to analyze and predict the number of traffic accidents in the type of accidents corresponding to the set criteria and the amount of damages according to the type of accidents.

The estimated information may be used to design insurance products in the future, but is not limited thereto.

That is, the computer may search for accident records corresponding to one or more criteria and provide statistics on specific contents of each accident record, thereby providing data for later analysis based on the type of accident.

Hereinafter, a method for evaluating each driver using big data that is based on a computer system verified and managed according to the above-described method and managed according to the above-described method will be described in detail.

12 is a flowchart illustrating a method of evaluating a driver based on an accident tendency, according to an exemplary embodiment.

In step S310, the computer acquires the recorded data for the first driver.

In the disclosed embodiment, the first driver may mean a driver who subscribes to automobile insurance or drives a vehicle subscribed to automobile insurance through the system according to the disclosed embodiment, but is not limited thereto.

As an example, the first driver may refer to a driver who drives a rental car covered by auto insurance through the system according to the disclosed embodiment.

In step S320, the computer extracts one or more accident data from the acquired data.

In an embodiment, the computer may obtain information on an accident record associated with the first driver. In some cases, there may be no accident record associated with the first driver.

In step S330, the computer determines one or more accident tendencies for the first driver based on the one or more accident data.

In an embodiment, the accident tendency may include information on whether the first driver has an accident, an accident frequency, and details of the accident.

For example, the driver's accident tendency may include at least one of the average negligence rate of the first driver, the accident cycle, the amount of damage in the past year, the number of high-cost accidents, the number of serious accidents, fine information, the distance from the accident site, and the late-night accident. However, it is not limited thereto.

It is also possible that at least one of the above-described accident propensity determination factors is omitted or other determination factors not disclosed may be added.

In an embodiment, the average error rate may be obtained based on an average of the error rates of accidents that occur while the first driver is driving.

For example, the average negligence rate can only evaluate accidents that occurred during the driving of the first driver, and specific types of accidents such as no-fault, unknown perpetrators, theft, natural disasters, and fires can be excluded from the judgment criteria.

In one embodiment, the computer may assign a score as follows based on the average error rate, but this is disclosed as an example and is not limited thereto.

0% ~ 30% = 1 point

31% to 50% = 2 points

51% ~ 70% = 3 points

71% ~ 90% = 4 points

91% to 100% = 5 points

In an embodiment, the accident cycle may be obtained based on the number of accidents compared to the contract elapsed period of the first driver. Specifically, the contract elapsed period may be obtained based on the contract elapsed month, but is not limited thereto. Here, indemnity cases can be excluded among cases of no-fault accidents and unknown perpetrators.

In addition, when new contracts are evaluated within 6 months, they are evaluated as the median default value, which is evaluated in this way only for one case, and for two or more cases, it is evaluated according to the original criteria.

As an example, points may be given as follows according to the accident cycle, but this is disclosed as an example and is not limited thereto.

0 to 5 months = 5 points

6 months ~ 12 months = 4 points

13 months to 18 months = 3 points

19 months to 24 months = 2 points

25 months or more = 1 point

In one embodiment, the elapsed yearly damage amount may be obtained based on the total amount of damage compared to the elapsed contract period of the first driver. Here, the contract elapsed period may be based on one year, but is not limited thereto.

In this case, the remedy among the no-fault accidents and unknown perpetrators may be excluded from the judgment criteria, but is not limited thereto.

As an example, a score may be given as follows according to the amount of loss in the elapsed year, but this is disclosed as an example and is not limited thereto.

0 ~ 1.5 million won = 1 point

1.15 million won ~ 3 million won = 2 points

3,100,000 won ~ 5 million won = 3 points

5,100,000 won ~ 8 million won = 4 points

8,100,000 won or more = 5 points

In an embodiment, the number of high-cost accidents may be obtained based on the number of high-cost accidents exceeding a preset criterion.

For example, a case of 10 million won or more is considered as one, and additional points are given per case according to the amount. This can be accumulated and calculated for each case.

As an example, a score may be given as follows according to the number of high-cost accidents, but this is disclosed as an example and is not limited thereto.

10 million won ~ 20 million won = 1 point added per case

2001 million won ~ 30 million won = 2 points added per case

KRW 3,10,000 ~ KRW 40,000 = 3 points added per case

KRW 400,000 ~ KRW 50 million = 4 points added per case

510 thousand won or more = plus 5 points

In an embodiment, the number of serious accidents may mean the number of serious accidents corresponding to a preset item. For example, a serious accident may mean a malicious accident such as alcohol, unlicensed, hit and run, insurance fraud.

As an example, points may be given as follows according to the number of serious accidents, but this is disclosed as an example and is not limited thereto.

Drinking, unlicensed, insurance fraud = 5 points

Hit and run = 4 points

12 heavy fruits = 3 points

In an embodiment, the penalty information may include the number of fines corresponding to a preset item and contents thereof. For example, the penalty information includes information on the violation of traffic rules, which is expressed as removing a ticket, for example, 1 point for violations within 10-20km/h and 2 points for violations of 20km/h or more in case of a speed violation. , In the case of violation of the exclusive bus lane, 1 point is awarded, and points may be awarded based on the penalty information.

In an embodiment, the accident location separation degree may be obtained based on a result of calculating each separation degree according to an accident occurrence order of accidents associated with the first driver.

That is, when there is a plurality of accident information associated with the first driver, the distance between each accident information is sequentially calculated, and the average of the accident location distance may be calculated, but is not limited thereto.

As an example, points may be given as follows according to the degree of separation of the accident location, but this is disclosed as an example and is not limited thereto.

0km ~ 50km = 1 point

51km ~ 100km = 2 points

101km ~ 150km = 3 points

151km ~ 200km = 4 points

More than 201km = 5 points

The late-night accident degree may include information on an accident occurring at a preset late-night time. For example, for an accident that occurs between 10 pm and 4 am, 2 points may be additionally assigned per incident when more than two times occur, but the present disclosure is not limited thereto.

In an embodiment, whether or not an accident has occurred due to multiple drivers may also be considered as a factor in determining an accident propensity. For example, a higher additional point may be given to a total of two accidents, once each by two different drivers, than two accidents caused by one driver during the rental period. In this case, it may be determined that the vehicle is driven by multiple drivers, and according to existing statistics, it is confirmed that the accident rate or loss rate of the vehicle driven by multiple drivers is higher, and additional points can be given to this.

That is, in the disclosed embodiment, the content mentioned as the accident tendency of the first driver may mean the accident tendency of the first driver himself, but may mean the accident tendency of the insurance contract contracted by the first driver or the insurance contracted vehicle. May be. The accident propensity analysis method according to the disclosed embodiment may be applied to an individual driver or to a specific vehicle or insurance contract.

In particular, in the case of corporate vehicles, a plurality of drivers (eg, employees) are often driven, so in this case, it may be more meaningful to analyze the accident tendency of the vehicle itself rather than the accident tendency analysis for each driver.

In this case, the accident tendency for the driver who caused the accident or the individual driver who is confirmed to be operating the vehicle or who is likely to drive the vehicle may be synthesized and used as data to determine the accident tendency for the vehicle, It is also possible to determine the accident tendency of the vehicle by analyzing the accident tendency of the accidents that have occurred in the vehicle itself.

The determination result may be used to determine whether to subscribe to insurance or to calculate the insurance amount when renewing the insurance contract of the vehicle, but is not limited thereto.

Depending on the embodiment, it may be determined that the accident tendency of a specific vehicle is the accident tendency of a corporation or group that has purchased insurance for the vehicle. In this case, even if the corporation or group wants to subscribe to other insurance, it may decide whether to subscribe to insurance or calculate the insurance amount based on the result of the determination.

In step S340, the computer acquires evaluation information on the first driver based on the accident tendency of the first driver.

That is, the computer does not acquire information for car insurance management based on the driving tendency of the first driver, but acquires the accident tendency of the first driver from the accident record of the first driver, and based on this, Its main feature is to obtain evaluation information.

In obtaining evaluation information, the computer may calculate an evaluation score for the first driver based on an accident tendency toward the first driver, and the evaluation score is the first driver as described in connection with the step S330. Average negligence rate, accident cycle, elapsed year damage, high-cost accidents, serious accidents, fine information, accident location separation, and late-night accidents. I can. According to an embodiment, the score calculated based on the personal information of the first driver may be considered together in the evaluation score calculation.

For example, if the driver's age is less than 26 or 70 or older, an additional point may be given to the score, but is not limited thereto.

According to existing statistics, the accident rate of those under the age of 26 and those over the age of 70 was found to be high. Accordingly, the trend can be continuously updated by extracting and analyzing the accident rate statistics for those under the age of 26 and those over the age of 70.

According to an embodiment, accident rate statistics for each age may be extracted and analyzed at a predetermined period. Accordingly, the above-described additional point age range set to less than 26 years old and 70 years old or more can be adjusted up and down, and a plurality of different age ranges are set according to embodiments, and different additional points or Penalties may be given. For example, additional points may be given to an age or age range whose accident rate is equal to or greater than a preset criterion, and deduction points may be given to an age or age range whose accident rate is less than a preset criterion, but are not limited thereto.

In an embodiment, the computer may obtain information on a preset depth warning item, and search whether there is at least one item corresponding to the depth warning item in the information on the first driver.

When one or more items corresponding to the in-depth warning item are searched for from the information on the first driver, the computer may provide a notification (alarm) corresponding thereto.

The in-depth warning item is a separate classification of matters to be noted for the driver, and if there is a content corresponding to this, an additional alert can be used to guide the manager to conduct an in-depth review of the driver.

For example, the in-depth survey items may include the following items, but are not limited thereto.

1. Cases where the date of receipt and the date of the accident are different

2. Accident occurred within 10 days of endorsement

3. Late-night accident (22:00 ~ 5:00)

4. Accident with 4 or more adults

5. Over 5 million won for unsuccessful athlete re vegan for foreign cars

13 is a diagram illustrating a method of learning an artificial intelligence model according to an exemplary embodiment.

In the disclosed embodiment, the artificial intelligence model may mean a model learned based on machine learning technology, but is not limited thereto.

In an embodiment, the computer may extract one or more parameters from information on the accident tendency of the first driver.

The type, unit, and specific value of the parameter are not limited, and for example, the parameter may be configured in the form of a vector that can be used as an input of an artificial intelligence model, but is not limited thereto.

For example, the extracted parameter may include a score for each item of an accident tendency or a value for each item of an accident tendency, but is not limited thereto. For example, a parameter obtained by vectorizing the scores for each accident item described with reference to FIG. 12 may be obtained.

In an embodiment, the computer may input the extracted parameters into a pre-learned artificial intelligence model. The shape of the artificial intelligence model is not limited, but as described above, it may mean an artificial intelligence model previously learned based on machine learning.

In an embodiment, the computer may calculate an estimated amount of damage to the first driver from the output of the artificial intelligence model.

In an embodiment, the artificial intelligence model may directly output the expected damage amount according to the parameter input, but may also output information for calculating the expected damage amount. In this case, the artificial intelligence model may be learned based on training data including a parameter representing the driver's accident tendency and information about the driver's loss amount (eg, loss amount for one year). Here, the time point at which the driver's accident propensity information is collected and the time point at which the driver's loss amount is collected may be different. For example, if it is assumed that the time when accident propensity information was collected is until 2017, the time when the driver's damages are collected is set to be for the next year, 2018, and based on the existing accident propensity information. Learning data can be configured to predict the amount of damage over a year (contract period).

After that, when information about the driver in question is collected for one year in 2019, based on the collected information, the damage amount information is updated to the information in 2019, and the artificial intelligence model is used using the accident propensity information up to 2018 and the damage amount information in 2019. You can make it learn.

In one embodiment, information on an accident tendency may be accumulated and learned, and according to an embodiment, only one year of information may be reflected, or it may be synthesized.

For example, an artificial intelligence model may be trained based on information on accident propensity for the year 2018 and loss information for the year 2019. As another example, an artificial intelligence model may be trained based on accumulated accident propensity information up to 2018 and loss amount information for the year 2019. Depending on the embodiment, training data according to the damage information base year and the cumulative number of years of accident propensity information (for example, information for one year, cumulative information for two years, cumulative information for three years, etc.) for the same driver. You can also augment in various ways. According to an embodiment, the accumulated data may be averaged and used for learning.

According to an embodiment, information for calculating the estimated damage amount may be obtained through a plurality of artificial intelligence models, and the estimated damage amount may be calculated based on the information.

In an embodiment, the estimated damage amount may be calculated in units of a predetermined contract period, and may mean, for example, an estimated damage amount for one year. Herein, the amount of damage may mean an estimated amount that must be paid by the insurance company due to an accident, etc. when purchasing insurance.

In an embodiment, the computer may predict the number of accidents of the first driver and the type of accidents expected for the first driver in order to calculate the estimated amount of damage.

For example, the computer is a first artificial intelligence that is learned based on learning data including a parameter representing the driver's existing accident tendency and information on the number of accidents within a predetermined period (eg, 1 year) after the driver. The number of accidents of the first driver can be predicted using the intelligent model. According to an embodiment, the number of accidents may be obtained in the form of a probability of occurrence of an accident, or may be obtained in the form of the number of accidents expected within a predetermined period (eg, one year).

The computer may predict the predicted number of accidents in the next year of the first driver by using the learned first artificial intelligence model.

In addition, the computer may predict the types of accidents that the first driver may cause by using the second artificial intelligence model. The second artificial intelligence model may be learned through parameters representing the drivers' existing accident tendencies and information on the types of accidents that have occurred within a predetermined period (eg, 1 year) of the corresponding driver.

The computer may estimate the expected damage amount of the accident based on the predicted type of accident and information on the vehicle the first driver drives.

The computer may calculate the estimated damage amount of the first driver by combining the expected number of accidents and the estimated damage amount.

In one embodiment, due to the revision of laws or regulations, a difference may occur between existing data and reality. For example, due to the revision of the law, the ratio of negligence for the same situation may be changed, gross negligence may be applied to matters that were not gross negligence in the past, or matters that were not previously illegal may be illegal.

In addition, the method of calculating the amount of compensation may be different due to changes in the insurance company's internal regulations.

In this way, the computer may obtain information on amendments to one or more laws or regulations related to the calculation of the estimated damage amount.

The computer can acquire the training data used for training the artificial intelligence model, and can modify the content of the training data based on the revision information. For example, the driver's accident record information corresponding to each learning data may be acquired, and information on the driver's error rate, gross negligence, and compensation amount may be corrected based on this.

In addition, the computer can augment the modified learning data. For example, data augmentation may be performed by generating a plurality of learning data having contents and values similar to those of the existing learning data, based on the corrected learning data and the revision information used for the correction.

In an embodiment, the computer may retrain the artificial intelligence model using the modified and augmented learning data.

Through this, the computer can update the artificial intelligence model so that amendments to the laws or regulations can be reflected in the artificial intelligence model and output results according to the current laws or regulations.

Hereinafter, another embodiment applicable to the present invention will be described.

In an embodiment, the computer may estimate a time zone and a location in which each driver has a high probability of causing an accident, based on the accident tendency of each driver. When it is determined that the driver is driving in a situation corresponding to at least one of the time zone and location, the computer may transmit a push message to the user terminal through the insurance management app to remind attention to the occurrence of an accident.

According to an embodiment, the computer may acquire information on one or more regions and time zones with a high probability of occurrence of a high-cost accident based on the accident tendency of each driver and the accident records of other drivers. Likewise, the computer can provide push notifications to drivers who are driving in areas and times of high probability of accidents.

However, there may be a situation in which the user terminal 30 temporarily or basically cannot receive the push message. For example, since the user terminal 30 is not a smartphone, it may not be possible to install an application for performing the method according to the disclosed embodiment.

As another example, the application may not be installed on the user terminal 30.

As another example, although the application is installed on the user terminal 30, it may not be in a state of being operated in the background to receive a push message.

The state in which the user terminal 30 cannot receive the push message is not limited to the above examples.

In an embodiment, when the user terminal 30 cannot receive the push message, the server may send a text message to the user terminal 30 through another terminal.

Likewise, if the other terminal cannot receive the push message, the server may send a text message to the other terminal using the user terminal 30.

In one embodiment, the terminals capable of sending text messages to each other by the server may be terminals previously registered with the server based on an application from each terminal or an application and approval.

For example, if the second user terminal is registered in the server based on mutual application and approval for the first user terminal, the first user terminal is a terminal that cannot receive a push message or a push message cannot be received. When in the state, the server may transmit a push message to the second user terminal, and the second user terminal may generate a text message corresponding to the push message and transmit it to the first user terminal.

Message transmission between users may incur no cost, or may be less expensive than corporate message transmission. Therefore, when the first user terminal cannot receive the push message, the server sends a text message to the first user terminal through the second user terminal, and instead, the server can provide a predetermined point to the second user terminal. . The provided points may be exchanged for cash or used for online shopping, but are not limited thereto.

As another example, the terminals capable of sending text messages to each other by the server may be terminals registered in a specific pool by approving a request from the server.

For example, unlike the one-to-one registration as in the above-described example, according to the present embodiment, a plurality of user terminals may be registered as a pool in the server based on prior approval.

The server transmits an inquiry message to check whether or not the push message can be received to the first user terminal, and when a reply to the inquiry message is received from the first user terminal, the server transmits the push message to the first user terminal, and 1 If a reply is not received from the user terminal, select at least one terminal capable of receiving a push message from among the user terminals registered in the pool, and request the first user terminal to send a text message while sending a push message to the terminal. I can.

In one embodiment, the server may pay predetermined points to user terminals registered in the pool. The frequency, degree and method of point payments are not limited.

In order to perform the above-described embodiment, the user terminal 30 receiving a text message transmission approval request from the server to the first user terminal, transmitting approval information for the approval request to the server, the first user terminal 1 Storing information on the first user terminal including a phone number of the user terminal, receiving a first push message including message transmission request information for the first user terminal from the server, the Acquiring content included in a first push message, generating a first text message using content included in the first push message, acquiring a first phone number included in the first push message , Comparing the first phone number with the phone number of the first user terminal, when the first phone number and the phone number of the first user terminal match, the first text message with the first phone number Transmitting, transmitting a transmission result of the first text message to the server, acquiring a point corresponding to the transmission result of the first text message from the server, requesting registration for a first pool from the server Receiving, transmitting approval information for the registration request to the server, receiving an encryption key corresponding to the first pool from the server, a second user included in the first pool from the server Receiving a second push message including message transmission request information to the terminal, acquiring content included in the second push message, and sending a second text message using content included in the second push message. Generating, encrypting the second text message using the encryption key, obtaining a second phone number included in the second push message, and converting the encrypted second text message to the second phone number. Transmitting to the server, transmitting the transmission result of the second text message to the server, and obtaining a point corresponding to the transmission result of the second text message from the server. Can follow the steps.

In addition, the user terminal 30 receives a third text message from a third user terminal, confirms whether the third text message is encrypted, and when the third text message is encrypted, the encryption key Decrypting the third text message by using, comparing the content of the push message with the content included in the third text message, and the content of the push message and the content included in the third text message are matched. If so, transmitting the fact of acquiring the third text message to the server may be performed.

14 is a block diagram of an apparatus according to an exemplary embodiment.

The processor 102 may include one or more cores (not shown) and a graphic processing unit (not shown) and/or a connection path (eg, a bus) for transmitting and receiving signals with other components. .

The processor 102 according to an embodiment executes one or more instructions stored in the memory 104 to perform the method described with reference to FIGS. 1 to 13.

Meanwhile, the processor 102 temporarily and/or permanently stores a signal (or data) processed inside the processor 102, a RAM (Random Access Memory, not shown) and a ROM (Read-Only Memory). , Not shown) may further include. In addition, the processor 102 may be implemented in the form of a system on chip (SoC) including at least one of a graphic processing unit, RAM, and ROM.

The memory 104 may store programs (one or more instructions) for processing and controlling the processor 102. Programs stored in the memory 104 may be divided into a plurality of modules according to functions.

The steps of a method or algorithm described in connection with an embodiment of the present invention may be implemented directly in hardware, implemented as a software module executed by hardware, or a combination thereof. Software modules include Random Access Memory (RAM), Read Only Memory (ROM), Erasable Programmable ROM (EPROM), Electrically Erasable Programmable ROM (EEPROM), Flash Memory, hard disk, removable disk, CD-ROM, or It may reside on any type of computer-readable recording medium well known in the art to which the present invention pertains.

Components of the present invention may be implemented as a program (or application) and stored in a medium in order to be combined with a computer that is hardware to be executed. Components of the present invention may be implemented as software programming or software elements, and similarly, embodiments include various algorithms implemented with a combination of data structures, processes, routines or other programming elements, including C, C++. , Java, assembler, or the like may be implemented in a programming or scripting language. Functional aspects can be implemented with an algorithm running on one or more processors.

In the above, embodiments of the present invention have been described with reference to the accompanying drawings, but those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features. You will be able to understand. Therefore, the embodiments described above are illustrative in all respects, and should be understood as non-limiting.

Claims (10)

In the method performed by a computer,
Acquiring recorded data for the first driver;
Extracting one or more accident data from the acquired data;
Determining at least one accident tendency for the first driver based on the at least one accident data; And
Acquiring evaluation information for the first driver based on the accident tendency of the first driver; Containing,
A method of evaluating drivers based on their accident tendencies. The method of claim 1,
The above thinking tendencies are,
It characterized in that it includes information on the accident status of the first driver, the accident frequency, and the contents of the accident,
A method of evaluating drivers based on their accident tendencies. The method of claim 2,
The above thinking tendencies are,
Including at least one of the average negligence rate of the first driver, the accident cycle, the amount of damage in the past year, the number of high-income accidents, the number of serious accidents, fine information, the accident location separation degree, and the late-night accident degree,
The average fruit rate is,
The first driver is obtained based on the average of the negligence rate of accidents that occurred while driving,
The accident cycle is,
It is obtained based on the number of accidents compared to the contract elapsed period of the first driver,
The amount of damage in the past year is:
It is acquired based on the total amount of damage compared to the contract elapsed period of the first driver,
The number of high-cost accidents above is,
It is acquired based on the number of high-cost accidents exceeding the preset criteria,
The number of serious accidents above is,
It means the number of serious accidents corresponding to a preset item,
The above penalty information is:
Including the number of fines corresponding to the preset items and their contents,
The above accident location separation is:
It is obtained based on the result of calculating each separation degree according to the accident occurrence order of accidents related to the first driver,
The above late-night accident is:
Including information on accidents that occurred at a preset late-night time,
A method of evaluating drivers based on their accident tendencies. The method of claim 3,
The step of obtaining the evaluation information,
Calculating an evaluation score for the first driver based on the accident tendency toward the first driver; Including more,
The above evaluation score is,
The score calculated based on the preset scoring criteria for at least one of the average negligence rate of the first driver, accident cycle, elapsed year damage, high-cost accidents, serious accidents, fine information, accident location separation, and late-night accidents, and , It characterized in that obtained by synthesizing the score calculated based on the personal information of the first driver,
A method of evaluating drivers based on their accident tendencies. The method of claim 1,
Obtaining information on a preset deep warning item;
Searching whether there is at least one item corresponding to the in-depth warning item in the information on the first driver; And
When one or more items corresponding to the in-depth warning item are searched from the information on the first driver, providing a notification corresponding thereto; Further comprising,
A method of evaluating drivers based on their accident tendencies. The method of claim 1,
The step of obtaining the evaluation information,
Extracting one or more parameters from information on the accident tendency of the first driver;
Inputting the extracted parameters into a pre-learned artificial intelligence model; And
Calculating a predicted amount of damage to the first driver from the output of the artificial intelligence model; Further comprising,
A method of evaluating drivers based on their accident tendencies. The method of claim 6,
The step of calculating the estimated damage amount,
Predicting the number of accidents of the first driver;
Predicting an accident type of the first driver;
Calculating an estimated cost according to the type of accident of the first driver; And
Calculating a predicted amount of damage to the first driver based on the number of accidents and the predicted cost of the first driver; Containing,
A method of evaluating drivers based on their accident tendencies. The method of claim 6,
Acquiring revision information of one or more laws or regulations related to calculating the estimated amount of damage;
Acquiring training data used for training the artificial intelligence model;
Modifying the content of the learning data based on the revision information;
Augmenting learning data reflecting the revision information; And
Retraining the artificial intelligence model using the modified and augmented learning data; Further comprising,
A method of evaluating drivers based on their tendency to accident. A memory for storing one or more instructions; And
A processor that executes the one or more instructions stored in the memory,
The processor executes the one or more instructions,
An apparatus for performing the method of claim 1. A computer program combined with a computer as hardware and stored in a recording medium readable by a computer to perform the method of claim 1.

Images