WO2019169768A1 - 保险保单集中核单方法、电子装置及可读存储介质 - Google Patents

保险保单集中核单方法、电子装置及可读存储介质 Download PDF

Info

Publication number
WO2019169768A1
WO2019169768A1 PCT/CN2018/089719 CN2018089719W WO2019169768A1 WO 2019169768 A1 WO2019169768 A1 WO 2019169768A1 CN 2018089719 W CN2018089719 W CN 2018089719W WO 2019169768 A1 WO2019169768 A1 WO 2019169768A1
Authority
WO
WIPO (PCT)
Prior art keywords
insurance policy
preset
mean
risk
policy
Prior art date
Application number
PCT/CN2018/089719
Other languages
English (en)
French (fr)
Inventor
刘洪晔
Original Assignee
平安科技（深圳）有限公司
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by 平安科技（深圳）有限公司 filed Critical 平安科技（深圳）有限公司
Publication of WO2019169768A1 publication Critical patent/WO2019169768A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/10—Office automation; Time management
    • G06Q10/103—Workflow collaboration or project management
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q40/00—Finance; Insurance; Tax strategies; Processing of corporate or income taxes
    • G06Q40/08—Insurance

Definitions

  • the present application relates to the field of computer technology, and in particular, to an insurance policy centralized nuclear single method, an electronic device, and a readable storage medium.
  • the insurance policy is manually screened, and the relevant information of the suspected risk items is uploaded to the headquarters by fax, etc., and then the headquarters will have a professional underwriting team to make manual judgments, and then the manual judgment results will be Send it to the local company for processing.
  • the whole process is driven by labor, which is time-consuming and laborious. As the number of risk components increases sharply, it will consume a lot of labor and time costs, greatly reducing the efficiency of claims, resulting in poor customer experience.
  • the purpose of the present application is to provide an insurance policy centralized verification method, an electronic device and a readable storage medium, which aim to improve the efficiency of the insurance policy and the efficiency of claim settlement.
  • a first aspect of the present application provides an electronic device, where the electronic device includes a memory and a processor, where the memory stores an insurance policy centralized core system that can be run on the processor, When the insurance policy centralized verification system is executed by the processor, the following steps are implemented:
  • the insurance policy is assigned to enter the risk component underwriting process according to a preset allocation manner; if it is determined that the insurance policy is not a risk component, the utility directly enters a preset normal claim process.
  • the second aspect of the present application further provides an insurance policy centralized verification method, wherein the insurance policy centralized verification method includes:
  • the insurance policy is assigned to enter the risk component underwriting process according to a preset allocation manner; if it is determined that the insurance policy is not a risk component, the utility directly enters a preset normal claim process.
  • the third aspect of the present application further provides a computer readable storage medium storing an insurance policy centralized nuclear single system, wherein the insurance policy centralized nuclear single system can be executed by at least one processor to enable The at least one processor performs the steps of the insurance policy centralized verification method as described above.
  • the method, system and readable storage medium for the insurance policy centralized insurance policy proposed by the present application by extracting the preset attribute information in the insurance policy uploaded by the salesperson, using the pre-set risk component determination rule to determine whether the insurance policy is If the risk component is judged to be a risk component, the insurance policy is distributed according to a preset allocation method to further underwrite, that is, enter a pre-set risk component workflow: if it is judged that the insurance policy is not a risk component, If it is a normal piece, it will directly bring the policy insurance into the normal claim process.
  • the risk review process can be automated, and the risk and non-risk parts of a large number of insurance policies can be initially distinguished, without the need to manually determine a large amount of insurance.
  • the suspected risk component in the policy will then manually send the suspected risk component to the underwriting center for judgment, saving labor and time costs, and the insurance policy judged as non-risk parts can directly enter the normal claims process, improving the claims efficiency and improving the customer.
  • Experience since the process of risk and non-risk parts in all insurance policies can be unified, the risk review process can be automated, and the risk and non-risk parts of a large number of insurance policies can be initially distinguished, without the need to manually determine a large amount of insurance.
  • the suspected risk component in the policy will then manually send the suspected risk component to the underwriting center for judgment, saving labor and time costs, and the insurance policy judged as non-risk parts can directly enter the normal claims process, improving the claims efficiency and improving the customer.
  • the underwriting center for judgment, saving labor and time costs
  • FIG. 1 is a schematic diagram of an operating environment of a preferred embodiment of an insurance policy centralized nuclear billing system 10 of the present application;
  • FIG. 2 is a schematic flow chart of an embodiment of a method for centralized insurance policy of an insurance policy according to the present application.
  • first, second and the like in the present application are for the purpose of description only, and are not to be construed as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. .
  • features defining “first” and “second” may include at least one of the features, either explicitly or implicitly.
  • the technical solutions between the various embodiments may be combined with each other, but must be based on the realization of those skilled in the art, and when the combination of the technical solutions is contradictory or impossible to implement, it should be considered that the combination of the technical solutions does not exist. Nor is it within the scope of protection required by this application.
  • FIG. 1 is a schematic diagram of an operating environment of a preferred embodiment of the insurance policy centralized verification system 10 of the present application.
  • the insurance policy centralized nuclear single system 10 is installed and operated in the electronic device 1.
  • the electronic device 1 may include, but is not limited to, a memory 11, a processor 12, and a display 13.
  • Figure 1 shows only the electronic device 1 with components 11-13, but it should be understood that not all illustrated components may be implemented, and more or fewer components may be implemented instead.
  • the memory 11 is at least one type of readable computer storage medium, which in some embodiments may be an internal storage unit of the electronic device 1, such as a hard disk or memory of the electronic device 1.
  • the memory 11 may also be an external storage device of the electronic device 1 in other embodiments, such as a plug-in hard disk equipped on the electronic device 1, a smart memory card (SMC), and a secure digital device. (Secure Digital, SD) card, flash card, etc.
  • the memory 11 may also include both an internal storage unit of the electronic device 1 and an external storage device.
  • the memory 11 is configured to store application software installed on the electronic device 1 and various types of data, such as program codes of the insurance policy centralized core system 10 and the like.
  • the memory 11 can also be used to temporarily store data that has been output or is about to be output.
  • the processor 12 may be a central processing unit (CPU), a microprocessor or other data processing chip for running program code or processing data stored in the memory 11, for example Execute the insurance policy centralized nuclear order system 10 and the like.
  • CPU central processing unit
  • microprocessor or other data processing chip for running program code or processing data stored in the memory 11, for example Execute the insurance policy centralized nuclear order system 10 and the like.
  • the display 13 in some embodiments may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode) touch sensor, or the like.
  • the display 13 is configured to display information processed in the electronic device 1 and a user interface for displaying visualization, such as various attribute information in an insurance policy, whether the insurance policy is a judgment result of a risk component, and a risk component underwriting waiting Interface, normal claims interface, etc.
  • the components 11-13 of the electronic device 1 communicate with one another via a system bus.
  • the insurance policy centralized verification system 10 includes at least one computer readable instruction stored in the memory 11, the at least one computer readable instruction being executable by the processor 12 to implement various embodiments of the present application.
  • step S1 the insurance policy to be processed is received.
  • the electronic device can be connected to the salesperson terminal system through the network.
  • the salesperson submits the policy-related information to the insurance policy centralized verification system through the claims terminal system.
  • the salesperson can receive an insurance policy and related information sent by a user on a pre-installed client in a mobile phone, a tablet, a handheld claim terminal, a self-service terminal device, or the like, or receive the user on a mobile phone, a tablet, a handheld claims terminal, and a self-service.
  • An insurance policy and related information sent on a browser system in a terminal such as a terminal device.
  • Step S2 extracting preset attribute information in the insurance policy, and determining whether the insurance policy is a risk component based on the preset attribute information and using a preset risk component determination rule.
  • the preset attribute information in the insurance policy may be extracted.
  • the extracted preset attribute information includes but is not limited to: the insured The risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount, and so on. And determining, by using a preset risk component determination rule, whether the insurance policy is a risk component.
  • multiple preset attributes may be extracted (for example, selecting a few important ones)
  • Information such as age, education level, income level, treatment hospital grade, disease level, and reimbursement amount is input as a plurality of insurance checklist factors to a pre-trained deep learning model, and the output of the deep learning model is obtained, and according to The result is output to determine whether the insurance policy is a risk component.
  • the deep learning model includes, but is not limited to, the following models: Convolutional Neural Network (CNN), Recurrent Neural Network RNN and LSTM, Recursive Tensor Neural Network RNTN, Autoencoder Autoencoder, and the like.
  • the deep learning model is pre-trained by the following steps:
  • the sample factor is input as input to a preset deep learning model for training, and the output of the trained deep learning model is obtained;
  • Adjusting the parameters of the trained deep learning model (such as optimizing the values of the weights in the CNN network or adjusting the hidden layer parameters of the model) to minimize the resulting output between the output and the underwriting result of the sample (normal or risky) Error
  • the training is terminated and the trained deep learning model is obtained.
  • step S3 if it is determined that the insurance policy is a risk component, the insurance policy is distributed into the risk component underwriting process according to a preset allocation manner.
  • the manual acquisition of the underwriting task trigger button is received through a preset manual acquisition underwriting task trigger button, and the insurance policy is sent to the user for the user to perform risk checking on the insurance policy; That is, the team manually obtains the task function.
  • the insurance policy is automatically assigned to the corresponding user for the risk component underwriting according to the preset task processing ratio.
  • the risk is automatically verified by assigning to the corresponding user.
  • a plurality of flexible task assignment modes are provided to allocate the risk component underwriting. To a large extent, it avoids the slow processing of the processing personnel due to different working ability or negative completion, and improves the feedback timeliness of the underwriting assessment of risk parts.
  • the insurance policy If it is judged that the insurance policy is not a risk component, it directly enters a preset normal claim process. For example, it can automatically jump or wait for the preset time (5 seconds) without operation and directly enter the preset normal claim operation interface to complete the normal claim process.
  • the pre-set risk component determination rule is used to determine whether the insurance policy is a risk component, and if the insurance policy is determined to be a risk component,
  • the distribution method is used to allocate the insurance policy for further underwriting, that is, to enter a pre-set risk component workflow: if it is judged that the insurance policy is not a risk component, it is a normal piece, and the policy insurance is directly put into the normal claim process. Since the process of risk and non-risk parts in all insurance policies can be unified, the risk review process can be automated, and the risk and non-risk parts of a large number of insurance policies can be initially distinguished, without the need to manually determine a large amount of insurance. The suspected risk component in the policy will then manually send the suspected risk component to the underwriting center for judgment, saving labor and time costs, and the insurance policy judged as non-risk parts can directly enter the normal claims process, improving the claims efficiency and improving the customer. Experience.
  • the method specifically includes:
  • a1, a2...an are the policy attribute value points after the conversion of the first, second...n preset attribute information in the insurance policy
  • A1, A2 to An are the corresponding corresponding to all normal claims in the preset archive database. 1, 2 to n sets of reference attribute value points after conversion of the preset attribute information;
  • Gauss(an, mean(An), mean((an-mean(An)) ⁇ 2))) is the Gaussian of the nth policy attribute value point in the insurance policy in the corresponding nth reference attribute value point set
  • the probability value of occurrence in the distribution probability space, P is the superposition value of the probability of occurrence of the value points of the first, second to n policy attributes in the insurance policy;
  • each insurance policy is the reference attribute in the claim (such as the insured's risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount, etc.) And etc.) are converted into individual numerical points, and within a certain area, there is a Gaussian distribution probability space around each appearing point space in several normal claims of historical data records.
  • the insurance policy centralized verification system after receiving the policy-related information uploaded by the salesperson, the insurance policy centralized verification system first extracts the preset attribute information in the uploaded policy-related information, such as the insured person's Risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount and other information extracted from the insurance personnel and the policy system, and then the extracted default attributes
  • the information is converted into corresponding attribute value points according to the preset segment conversion mode, for example, converted into numerical points in a segmental manner, where the age attribute is taken as an example, for example, the identifier of age 0-16 is 1
  • the sign of 17-22 years old is 2; the mark of 23-35 years old is 3; the mark of 36-50 years old is 4; the mark of 51-65 years old is 5; the mark of 66 years old and above is 7; other attributes are Analogy, I won't go into details here.
  • a1, a2 to an is the first, second to n preset attribute information of the current insurance policy of this nuclear order (for example, selecting several more important attributes such as age, education level, income level, treatment hospital level) , disease level, reimbursement amount)
  • the value of the policy attribute after conversion, A1, A2 to An are the first, second to n preset attribute information corresponding to all normal claims in the default archive database (with the extracted current insurance policy)
  • mean(A1) is the mean of A1, which determines the position of the Gaussian distribution (ie, normal distribution) probability map.
  • mean((a1-mean(A1)) ⁇ 2) is for a1 and mean(A1)
  • the standard deviation is also the magnitude of the distribution of Gaussian distributions (ie, normal distributions).
  • Gauss(an, mean(An), mean((an-mean(An)) ⁇ 2))) is the Gaussian of the nth policy attribute value point in the insurance policy in the corresponding nth reference attribute value point set
  • P is the superposition value of the probability value of the first, second...n policy attribute numerical points in the insurance policy; the final calculated P value is the current insurance of the nuclear order The probability that the policy is a normal claim.
  • a reasonable preset probability threshold may be preset. If the calculated current policy is a probability that the normal claim is a normal claim, that is, the P value is less than the preset probability threshold, then the current insurance policy and the normal claim are not determined. Similarly, if the current insurance policy is a suspected claim risk, it is handed over to the underwriting authority for manual processing; if the P value is greater than the preset probability threshold, it is judged that the current insurance policy is similar to the normal claim, that is, the current insurance policy is normal. Claims, return a claimable token, allowing it to go through the normal claims process.
  • the preset probability threshold may also be adjusted by the user according to the needs of different application scenarios. For example, in a scenario where the risk component is required to be strict, the threshold may be appropriately increased; In a scenario where the judgment is not strict, the threshold can be appropriately lowered. More flexible and practical.
  • the insurance policy is saved in association with the returned underwriting result, and the insurance policy and the returned underwriting result are stored in a preset archive database.
  • an archiving system may also be introduced to store and archive each uploaded insurance policy in association with its final determination result, such as the current insurance policy and the feedback underwriting result (either risk or normal)
  • the association save is performed, and the insurance policy and the feedback underwriting result are stored in the preset archive database.
  • the insurance policy and other claims information and the final approval opinion (which is the risk component or the normal component, and finally the artificial underwriting result of the risk component) are filed, and the audit opinion can be retained, and the same user and similar claims are made in the subsequent process.
  • the situation can be referred to.
  • FIG. 2 is a schematic flowchart of an embodiment of a method for centralized insurance policy of an insurance policy according to an embodiment of the present invention.
  • step S10 the insurance policy to be processed is received.
  • the electronic device can be connected to the salesperson terminal system through the network.
  • the salesperson submits the policy-related information to the insurance policy centralized verification system through the claims terminal system.
  • the salesperson can receive an insurance policy and related information sent by a user on a pre-installed client in a mobile phone, a tablet, a handheld claim terminal, a self-service terminal device, or the like, or receive the user on a mobile phone, a tablet, a handheld claims terminal, and a self-service.
  • An insurance policy and related information sent on a browser system in a terminal such as a terminal device.
  • Step S20 extracting preset attribute information in the insurance policy, and determining whether the insurance policy is a risk component based on the preset attribute information and using a preset risk component determination rule.
  • the preset attribute information in the insurance policy may be extracted.
  • the extracted preset attribute information includes but is not limited to: the insured The risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount, and so on. And determining, by using a preset risk component determination rule, whether the insurance policy is a risk component.
  • multiple preset attributes may be extracted (for example, selecting a few important ones)
  • Information such as age, education level, income level, treatment hospital grade, disease level, and reimbursement amount is input as a plurality of insurance checklist factors to a pre-trained deep learning model, and the output of the deep learning model is obtained, and according to The result is output to determine whether the insurance policy is a risk component.
  • the deep learning model includes, but is not limited to, the following models: Convolutional Neural Network (CNN), Recurrent Neural Network RNN and LSTM, Recursive Tensor Neural Network RNTN, Autoencoder Autoencoder, and the like.
  • the deep learning model is pre-trained by the following steps:
  • the sample factor is input as input to a preset deep learning model for training, and the output of the trained deep learning model is obtained;
  • Adjusting the parameters of the trained deep learning model (such as optimizing the values of the weights in the CNN network or adjusting the hidden layer parameters of the model) to minimize the resulting output between the output and the underwriting result of the sample (normal or risky) Error
  • the training is ended, and the trained deep learning model is obtained.
  • step S30 if it is determined that the insurance policy is a risk component, the insurance policy is distributed into the risk component underwriting process according to a preset allocation manner.
  • the manual acquisition of the underwriting task trigger button is received through a preset manual acquisition underwriting task trigger button, and the insurance policy is sent to the user for the user to perform risk checking on the insurance policy; That is, the team manually obtains the task function.
  • the insurance policy is automatically assigned to the corresponding user for the risk component underwriting according to the preset task processing ratio.
  • the risk is automatically verified by assigning to the corresponding user.
  • a plurality of flexible task assignment modes are provided to allocate the risk component underwriting. To a large extent, it avoids the slow processing of the processing personnel due to different working ability or negative completion, and improves the feedback timeliness of the underwriting assessment of risk parts.
  • the insurance policy If it is judged that the insurance policy is not a risk component, it directly enters a preset normal claim process. For example, it can automatically jump or wait for the preset time (5 seconds) without operation and directly enter the preset normal claim operation interface to complete the normal claim process.
  • the pre-set risk component determination rule is used to determine whether the insurance policy is a risk component, and if the insurance policy is determined to be a risk component,
  • the distribution method is used to allocate the insurance policy for further underwriting, that is, to enter a pre-set risk component workflow: if it is judged that the insurance policy is not a risk component, it is a normal piece, and the policy insurance is directly put into the normal claim process. Since the process of risk and non-risk parts in all insurance policies can be unified, the risk review process can be automated, and the risk and non-risk parts of a large number of insurance policies can be initially distinguished, without the need to manually determine a large amount of insurance. The suspected risk component in the policy will then manually send the suspected risk component to the underwriting center for judgment, saving labor and time costs, and the insurance policy judged as non-risk parts can directly enter the normal claims process, improving the claims efficiency and improving the customer. Experience.
  • the step S20 specifically includes:
  • a1, a2 to an are the policy attribute value points after the conversion of the first, second to n preset attribute information in the insurance policy, and A1, A2 to An are corresponding to all normal claims in the preset archive database. 1, 2 to n sets of reference attribute value points after conversion of the preset attribute information;
  • Gauss(an, mean(An), mean((an-mean(An)) ⁇ 2))) is the Gaussian of the nth policy attribute value point in the insurance policy in the corresponding nth reference attribute value point set
  • the probability value of occurrence in the distribution probability space, P is the superposition value of the probability of occurrence of the value points of the first, second to n policy attributes in the insurance policy;
  • each insurance policy is the reference attribute in the claim (such as the insured's risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount, etc.) And etc.) are converted into individual numerical points, and within a certain area, there is a Gaussian distribution probability space around each appearing point space in several normal claims of historical data records.
  • the insurance policy centralized verification system after receiving the policy-related information uploaded by the salesperson, the insurance policy centralized verification system first extracts the preset attribute information in the uploaded policy-related information, such as the insured person's Risk, gender, age, education level, work industry, income level, treatment hospital grade, disease level, duration, reimbursement amount and other information extracted from the insurance personnel and the policy system, and then the extracted default attributes
  • the information is converted into corresponding attribute value points according to the preset segment conversion mode, for example, converted into numerical points in a segmental manner, where the age attribute is taken as an example, for example, the identifier of age 0-16 is 1
  • the sign of 17-22 years old is 2; the mark of 23-35 years old is 3; the mark of 36-50 years old is 4; the mark of 51-65 years old is 5; the mark of 66 years old and above is 7; other attributes are Analogy, I won't go into details here.
  • a1, a2 to an is the first, second to n preset attribute information of the current insurance policy of this nuclear order (for example, selecting several more important attributes such as age, education level, income level, treatment hospital level) , disease level, reimbursement amount)
  • the value of the policy attribute after conversion, A1, A2 to An are the first, second to n preset attribute information corresponding to all normal claims in the default archive database (with the extracted current insurance policy)
  • mean(A1) is the mean of A1, which determines the position of the Gaussian distribution (ie, normal distribution) probability map.
  • mean((a1-mean(A1)) ⁇ 2) is for a1 and mean(A1)
  • the standard deviation is also the magnitude of the distribution of Gaussian distributions (ie, normal distributions).
  • Gauss(an, mean(An), mean((an-mean(An)) ⁇ 2))) is the Gaussian of the nth policy attribute value point in the insurance policy in the corresponding nth reference attribute value point set
  • P is the superposition value of the probability value of the value points of the first, second to n policy attribute points in the insurance policy; the final calculated P value is the current insurance of the current nuclear order.
  • the probability that the policy is a normal claim.
  • a reasonable preset probability threshold may be preset. If the calculated current policy is a probability that the normal claim is a normal claim, that is, the P value is less than the preset probability threshold, then the current insurance policy and the normal claim are not determined. Similarly, if the current insurance policy is a suspected claim risk, it is handed over to the underwriting authority for manual processing; if the P value is greater than the preset probability threshold, it is judged that the current insurance policy is similar to the normal claim, that is, the current insurance policy is normal. Claims, return a claimable token, allowing it to go through the normal claims process.
  • the preset probability threshold may also be adjusted by the user according to the needs of different application scenarios. For example, in a scenario where the risk component is required to be strict, the threshold may be appropriately increased; In a scenario where the judgment is not strict, the threshold can be appropriately lowered. More flexible and practical.
  • the method further includes:
  • the insurance policy is saved in association with the returned underwriting result, and the insurance policy and the returned underwriting result are stored in a preset archive database.
  • an archiving system may also be introduced to store and archive each uploaded insurance policy in association with its final determination result, such as the current insurance policy and the feedback underwriting result (either risk or normal)
  • the association save is performed, and the insurance policy and the feedback underwriting result are stored in the preset archive database.
  • the insurance policy and other claims information and the final approval opinion (which is the risk component or the normal component, and finally the artificial underwriting result of the risk component) are filed, and the audit opinion can be retained, and the same user and similar claims are made in the subsequent process.
  • the situation can be referred to.
  • the present application further provides a computer readable storage medium storing an insurance policy centralized nuclear single system, the insurance policy centralized nuclear single system being executable by at least one processor to enable the The at least one processor performs the steps of the insurance policy centralized verification method in the above-mentioned embodiment, and the specific implementation processes of the steps S10, S20, and S30 of the insurance policy centralized verification method are as described above, and are not described herein again.
  • the foregoing embodiment method can be implemented by means of software plus a necessary general hardware platform, and can also be implemented by hardware, but in many cases, the former is A better implementation.
  • the technical solution of the present application which is essential or contributes to the prior art, may be embodied in the form of a software product stored in a storage medium (such as ROM/RAM, disk,
  • the optical disc includes a number of instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to perform the methods described in various embodiments of the present application.

Landscapes

  • Business, Economics & Management (AREA)
  • Engineering & Computer Science (AREA)
  • Strategic Management (AREA)
  • Human Resources & Organizations (AREA)
  • General Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Marketing (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Physics & Mathematics (AREA)
  • Finance (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Technology Law (AREA)
  • Data Mining & Analysis (AREA)
  • Development Economics (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • Financial Or Insurance-Related Operations Such As Payment And Settlement (AREA)

Abstract

本申请涉及一种保险保单集中核单方法、电子装置及可读存储介质，该方法包括：接收待处理的保险保单；提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。本申请节约了人工和时间成本，而且判断为非风险件的保险保单可直接自动进入正常理赔流程，提高了理赔效率，提升客户体验。

Description

保险保单集中核单方法、电子装置及可读存储介质
优先权申明
本申请基于巴黎公约申明享有2018年3月6日递交的申请号为CN 201810182943.9、名称为“保险保单集中核单方法、电子装置及可读存储介质”中国专利申请的优先权，该中国专利申请的整体内容以参考的方式结合在本申请中。
技术领域
本申请涉及计算机技术领域，尤其涉及一种保险保单集中核单方法、电子装置及可读存储介质。
背景技术
随着人们对保险行业的认识的加深，投保人数量也越来越多，随之保险保单理赔过程中风险件的数量也大量增加。现有技术中对疑似风险件的处理过程通常如下：
出险人员收集相关资料后，对保险保单进行人工甄别，将其中疑似风险件的相关资料通过传真等方式上传到总部，然后总部会有专业的核保团队进行人工判断，之后将人工判断结果再下发给当地公司进行处理。整个过程全由人工来推动流程，费时费力，随着风险件数量的急剧增加，会消耗大量人工和时间成本，大大降低了理赔效率，造成客户体验不佳。
发明内容
本申请的目的在于提供一种保险保单集中核单方法、电子装置及可读存储介质，旨在提高保险保单的核单效率及理赔效率。
为实现上述目的，本申请第一方面提供一种电子装置，所述电子 装置包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的保险保单集中核单系统，所述保险保单集中核单系统被所述处理器执行时实现如下步骤：
接收待处理的保险保单；
提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；
若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。
此外，为实现上述目的，本申请第二方面还提供一种保险保单集中核单方法，所述保险保单集中核单方法包括：
接收待处理的保险保单；
提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；
若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。
本申请第三方面还提供一种计算机可读存储介质，所述计算机可读存储介质存储有保险保单集中核单系统，所述保险保单集中核单系统可被至少一个处理器执行，以使所述至少一个处理器执行如上述的保险保单集中核单方法的步骤。
本申请提出的保险保单集中核单方法、系统及可读存储介质，通过提取出业务员上传的保险保单中的预设属性信息，利用预先设定好 的风险件判定规则判断该保险保单是否为风险件，若判断该保险保单是风险件，则按预设分配方式对该保险保单进行分配以进一步核保，即进入预先设定好的风险件工作流：若判断该保险保单不是风险件即是正常件，则直接使该保单保险进入正常理赔流程。由于能将所有保险保单中风险件和非风险件的流程统一化，将风险件的审核流程自动化，可将大量保险保单中的风险件和非风险件进行初步的区分判定，无需人工判定大量保险保单中的疑似风险件再将疑似风险件人工发送至核保中心判定，节约了人工和时间成本，而且判断为非风险件的保险保单可直接自动进入正常理赔流程，提高了理赔效率，提升客户体验。
附图说明
图1为本申请保险保单集中核单系统10较佳实施例的运行环境示意图；
图2为本申请保险保单集中核单方法一实施例的流程示意图。
具体实施方式
为了使本申请的目的、技术方案及优点更加清楚明白，以下结合附图及实施例，对本申请进行进一步详细说明。应当理解，此处所描述的具体实施例仅用以解释本申请，并不用于限定本申请。基于本申请中的实施例，本领域普通技术人员在没有做出创造性劳动前提下所获得的所有其他实施例，都属于本申请保护的范围。
需要说明的是，在本申请中涉及“第一”、“第二”等的描述仅用于描述目的，而不能理解为指示或暗示其相对重要性或者隐含指明所指示的技术特征的数量。由此，限定有“第一”、“第二”的特征 可以明示或者隐含地包括至少一个该特征。另外，各个实施例之间的技术方案可以相互结合，但是必须是以本领域普通技术人员能够实现为基础，当技术方案的结合出现相互矛盾或无法实现时应当认为这种技术方案的结合不存在，也不在本申请要求的保护范围之内。
本申请提供一种保险保单集中核单系统。请参阅图1，是本申请保险保单集中核单系统10较佳实施例的运行环境示意图。
在本实施例中，所述的保险保单集中核单系统10安装并运行于电子装置1中。该电子装置1可包括，但不仅限于，存储器11、处理器12及显示器13。图1仅示出了具有组件11-13的电子装置1，但是应理解的是，并不要求实施所有示出的组件，可以替代的实施更多或者更少的组件。
所述存储器11为至少一种类型的可读计算机存储介质，所述存储器11在一些实施例中可以是所述电子装置1的内部存储单元，例如该电子装置1的硬盘或内存。所述存储器11在另一些实施例中也可以是所述电子装置1的外部存储设备，例如所述电子装置1上配备的插接式硬盘，智能存储卡(Smart Media Card,SMC)，安全数字(Secure Digital,SD)卡，闪存卡(Flash Card)等。进一步地，所述存储器11还可以既包括所述电子装置1的内部存储单元也包括外部存储设备。所述存储器11用于存储安装于所述电子装置1的应用软件及各类数据，例如所述保险保单集中核单系统10的程序代码等。所述存储器11还可以用于暂时地存储已经输出或者将要输出的数据。
所述处理器12在一些实施例中可以是一中央处理器(Central Processing Unit,CPU)，微处理器或其他数据处理芯片，用于运行所 述存储器11中存储的程序代码或处理数据，例如执行所述保险保单集中核单系统10等。
所述显示器13在一些实施例中可以是LED显示器、液晶显示器、触控式液晶显示器以及OLED(Organic Light-Emitting Diode，有机发光二极管)触摸器等。所述显示器13用于显示在所述电子装置1中处理的信息以及用于显示可视化的用户界面，例如保险保单中的各个属性信息、保险保单是否为风险件的判断结果、风险件核保等待界面、正常理赔界面等。所述电子装置1的部件11-13通过系统总线相互通信。
保险保单集中核单系统10包括至少一个存储在所述存储器11中的计算机可读指令，该至少一个计算机可读指令可被所述处理器12执行，以实现本申请各实施例。
其中，上述保险保单集中核单系统10被所述处理器12执行时实现如下步骤：
步骤S1，接收待处理的保险保单。
本实施例中，电子装置可通过网络与业务员的理赔终端系统对接，在需要进行保险保单核单时，业务员通过理赔终端系统将保单相关信息提交至保险保单集中核单系统。例如，可接收用户在手机、平板电脑、手持理赔终端、自助终端设备等终端中预先安装的客户端上发送来的保险保单及相关信息，或接收用户在手机、平板电脑、手持理赔终端、自助终端设备等终端中的浏览器系统上发送来的保险保单及相关信息。
步骤S2，提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否 为风险件。
在接收到待处理的保险保单后，可提取所述保险保单中的预设属性信息，例如，当所述保险保单是医疗保险保单时，提取的预设属性信息包括但不限于：被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额，等等。并利用预先设定的风险件判定规则判断所述保险保单是否为风险件，例如，在一种可选的实施方式中，可将提取的多个预设属性(例如，选择几个较为重要的属性如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)的信息作为若干保险核单因子输入至预先训练好的深度学习模型，获取所述深度学习模型的输出结果，并根据输出结果来判断所述保险保单是否为风险件。其中，该深度学习模型包括但不限于以下几种模型：卷积神经网络(CNN)、递归神经网络RNN及LSTM、递归张量神经网络RNTN、自动编码器Autoencoder等等。该深度学习模型由以下步骤预先训练得到：
从过往历史数据中已确定的保险保单中分别提取出预设数量(50万件)的正常件和风险件作为样本，并提取出每一件样本中的多个预设属性(例如，选择几个较为重要的属性如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)的信息作为若干样本因子；
将样本因子作为输入投入至预设的深度学习模型进行训练，得到训练的深度学习模型的输出；
调整训练的深度学习模型的参数(如优化CNN网络内各权重的值或调整模型的隐层参数)，以最小化得到的所述输出与样本的核保结果(正常件或风险件)之间的误差；
若误差满足预设条件(如小于预设误差阈值)，则结束训练，得 到训练好的深度学习模型。
步骤S3，若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程。例如，通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；即提供了团队手动获取任务功能。或者，按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保。或者，根据所述保险保单的保险上报类型或者保险上报地域多维度自动分配至对应的用户进行风险件核保。本实施例中提供多种灵活的任务分配模式来分配进行风险件核保。在很大程度上避免了处理人员由于工作能力不同或者消极怠工等原因造成的处理缓慢的情况，提高了对风险件的核保评估的反馈时效。
若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。例如，可自动跳转或等待预设时间(5秒)无操作后直接进入预先设定的正常理赔操作界面，以完成正常理赔流程。
本实施例通过提取出业务员上传的保险保单中的预设属性信息，利用预先设定好的风险件判定规则判断该保险保单是否为风险件，若判断该保险保单是风险件，则按预设分配方式对该保险保单进行分配以进一步核保，即进入预先设定好的风险件工作流：若判断该保险保单不是风险件即是正常件，则直接使该保单保险进入正常理赔流程。由于能将所有保险保单中风险件和非风险件的流程统一化，将风险件的审核流程自动化，可将大量保险保单中的风险件和非风险件进行初步的区分判定，无需人工判定大量保险保单中的疑似风险件再将疑似 风险件人工发送至核保中心判定，节约了人工和时间成本，而且判断为非风险件的保险保单可直接自动进入正常理赔流程，提高了理赔效率，提升客户体验。
在一可选的实施例中，在上述图1的实施例的基础上，所述保险保单集中核单系统10被所述处理器12执行实现所述步骤S2时，具体包括：
提取所述保险保单中的预设属性信息，并将提取的各个预设属性信息分别按预设分段转换方式转换成对应的属性数值点；并将转换得到的各个属性数值点代入如下公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))
其中，a1，a2…an为所述保险保单中第1，2…n个预设属性信息转换后的保单属性数值点，A1，A2至An为预设归档数据库中所有正常理赔件对应的第1，2至n个预设属性信息转换后的参考属性数值点的集合；
Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1，2至n个保单属性数值点的出现概率值的叠加值；
若P大于预设概率阈值，则判断所述保险保单不是风险件；
若P小于或等于预设概率阈值，则判断所述保险保单是风险件。
本实施例中，由于理赔风险件和正常理赔件比起来一定会具有较 大差异，而正常理赔件的情况都存在相似性。因此，若将各个保险保单即理赔件中的各个参考属性(如被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额等等)转换成各个数值点，则在一定区域内，在历史数据记录的若干正常理赔件中每个出现的点空间内周围存在着一个高斯分布的概率空间。基于此原理，本实施例中，保险保单集中核单系统在接收到业务员上传的保单相关信息后，首先，提取出上传的保单相关信息中的预设属性信息，如可将被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额等信息从出险人员上报资料以及保单系统中提取出来，然后将提取的各个预设属性信息分别按预设分段转换方式转换成对应的属性数值点，例如，以分段的方式来转换成数值点，在此以年龄属性为例说明，如年龄为0-16岁的标识为1；17-22岁的标识为2；23-35岁的标识为3；36-50岁的标识为4；51-65岁的标识为5；66岁及以上的标识为7；其他属性以此类推，在此不再赘述。
在过往历史数据的预设归档数据库中保存有曾经进行过风险件判断的历史案件信息，找出历史案件中的所有正常理赔件。在根据上传的保险保单相关信息判断当前保险保单是否为风险件时，有如下公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))
其中，a1，a2至an为本次核单的当前保险保单中第1，2至n个预设属性信息(例如，选择几个较为重要的属性如年龄、文化程度、 收入水平、治疗医院等级、疾病等级、报销金额)转换后的保单属性数值点，A1，A2至An为预设归档数据库中所有正常理赔件对应的第1，2至n个预设属性信息(与提取的当前保险保单中相同的属性，如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)转换后的参考属性数值点的集合。公式中，mean(A1)为求A1的均值，决定了高斯分布(即正态分布)概率图的位置，mean((a1-mean(A1))^2)为求a1与mean(A1)的标准差，也是高斯分布(即正态分布)的分布的幅度。Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1，2…n个保单属性数值点的出现概率值的叠加值；最终计算得到的P值即为本次核单的当前保险保单为正常理赔件的概率。
由于上文中提到，在一定区域内,在历史数据记录的若干正常理赔件中每个出现的点空间内周围存在着一个高斯分布的概率空间，即正常理赔件间存在相似性，可利用当前要判断的保险保单是否与历史数据记录的正常理赔件相似来判断当前保险保单是否为风险件。也即当前要判断的保险保单中各个数值点出现在历史数据记录的若干正常理赔件中每个点附近空间位置的概率越高，则当前要判断的保险保单与正常理赔件的相似度越高。具体在公式中，各个点的概率值叠加得到的参数P越高，则当前要判断的保险保单为正常理赔件的可能性越大。
因此，本实施例中可预先设定一个合理的预设概率阈值，若计算得到的当前保单为正常理赔件的概率即P值小于该预设概率阈值，则 判断当前保险保单与正常理赔件不相似，即判断当前保险保单为疑似理赔风险件，则交由核保中心人工处理；若P值大于该预设概率阈值，则判断当前保险保单与正常理赔件相似，即判断当前保险保单为正常理赔件，则返回可理赔标记，从而允许其走正常理赔流程。
在预先设定概率阈值时，可将历史数据记录的若干理赔件(包括理赔风险件和正常理赔件)利用公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))来不断训练及验证概率阈值的合理性，在达到一定准确性后，即可设定好一合理的预设概率阈值。当然，在后续实际应用过程中，该预设概率阈值也可以被用户根据不同应用场景的需要进行调整，如在要求风险件判断较严格的场景中，可适当调高该阈值；在对风险件判断不太严格的场景中，可适当调低该阈值。更加灵活、实用。
在一可选的实施例中，所述保险保单集中核单系统10被所述处理器12执行时，还可实现如下步骤：
将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
本实施例中，还可引入归档系统，将每一个上传的保险保单与其最终的判定结果进行关联保存并归档，如将当前的所述保险保单与反馈的核保结果(是风险件或正常件)进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。即将保险保单等理赔资料与最终的审批意见(是风险件或正常件，以及最终对风险件的人 工核保结果)归档，可对审核意见留证，而且在后续过程中对相同用户以及类似理赔情况可给予参考。
如图2所示，图2为本申请保险保单集中核单方法一实施例的流程示意图，该保险保单集中核单方法包括以下步骤：
步骤S10，接收待处理的保险保单。
本实施例中，电子装置可通过网络与业务员的理赔终端系统对接，在需要进行保险保单核单时，业务员通过理赔终端系统将保单相关信息提交至保险保单集中核单系统。例如，可接收用户在手机、平板电脑、手持理赔终端、自助终端设备等终端中预先安装的客户端上发送来的保险保单及相关信息，或接收用户在手机、平板电脑、手持理赔终端、自助终端设备等终端中的浏览器系统上发送来的保险保单及相关信息。
步骤S20，提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件。
在接收到待处理的保险保单后，可提取所述保险保单中的预设属性信息，例如，当所述保险保单是医疗保险保单时，提取的预设属性信息包括但不限于：被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额，等等。并利用预先设定的风险件判定规则判断所述保险保单是否为风险件，例如，在一种可选的实施方式中，可将提取的多个预设属性(例如，选择几个较为重要的属性如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)的信息作为若干保险核单因子输入至 预先训练好的深度学习模型，获取所述深度学习模型的输出结果，并根据输出结果来判断所述保险保单是否为风险件。其中，该深度学习模型包括但不限于以下几种模型：卷积神经网络(CNN)、递归神经网络RNN及LSTM、递归张量神经网络RNTN、自动编码器Autoencoder等等。该深度学习模型由以下步骤预先训练得到：
从过往历史数据中已确定的保险保单中分别提取出预设数量(50万件)的正常件和风险件作为样本，并提取出每一件样本中的多个预设属性(例如，选择几个较为重要的属性如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)的信息作为若干样本因子；
将样本因子作为输入投入至预设的深度学习模型进行训练，得到训练的深度学习模型的输出；
调整训练的深度学习模型的参数(如优化CNN网络内各权重的值或调整模型的隐层参数)，以最小化得到的所述输出与样本的核保结果(正常件或风险件)之间的误差；
若误差满足预设条件(如小于预设误差阈值)，则结束训练，得到训练好的深度学习模型。
步骤S30，若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程。例如，通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；即提供了团队手动获取任务功能。或者，按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保。或者，根据所述保险保单的保险上报类型或者保险上报地域多维度自动分配至对应的用户进行风险件核保。本实施例中提供多种灵活的任务分配模式 来分配进行风险件核保。在很大程度上避免了处理人员由于工作能力不同或者消极怠工等原因造成的处理缓慢的情况，提高了对风险件的核保评估的反馈时效。
若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。例如，可自动跳转或等待预设时间(5秒)无操作后直接进入预先设定的正常理赔操作界面，以完成正常理赔流程。
本实施例通过提取出业务员上传的保险保单中的预设属性信息，利用预先设定好的风险件判定规则判断该保险保单是否为风险件，若判断该保险保单是风险件，则按预设分配方式对该保险保单进行分配以进一步核保，即进入预先设定好的风险件工作流：若判断该保险保单不是风险件即是正常件，则直接使该保单保险进入正常理赔流程。由于能将所有保险保单中风险件和非风险件的流程统一化，将风险件的审核流程自动化，可将大量保险保单中的风险件和非风险件进行初步的区分判定，无需人工判定大量保险保单中的疑似风险件再将疑似风险件人工发送至核保中心判定，节约了人工和时间成本，而且判断为非风险件的保险保单可直接自动进入正常理赔流程，提高了理赔效率，提升客户体验。
在一可选的实施例中，在上述实施例的基础上，所述步骤S20具体包括：
提取所述保险保单中的预设属性信息，并将提取的各个预设属性信息分别按预设分段转换方式转换成对应的属性数值点；并将转换得到的各个属性数值点代入如下公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))
其中，a1，a2至an为所述保险保单中第1，2至n个预设属性信息转换后的保单属性数值点，A1，A2至An为预设归档数据库中所有正常理赔件对应的第1，2至n个预设属性信息转换后的参考属性数值点的集合；
Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1，2至n个保单属性数值点的出现概率值的叠加值；
若P大于预设概率阈值，则判断所述保险保单不是风险件；
若P小于或等于预设概率阈值，则判断所述保险保单是风险件。
本实施例中，由于理赔风险件和正常理赔件比起来一定会具有较大差异，而正常理赔件的情况都存在相似性。因此，若将各个保险保单即理赔件中的各个参考属性(如被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额等等)转换成各个数值点，则在一定区域内，在历史数据记录的若干正常理赔件中每个出现的点空间内周围存在着一个高斯分布的概率空间。基于此原理，本实施例中，保险保单集中核单系统在接收到业务员上传的保单相关信息后，首先，提取出上传的保单相关信息中的预设属性信息，如可将被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间，报销金额等信息从出险人员上报资料以及保单系统中提取出来，然后 将提取的各个预设属性信息分别按预设分段转换方式转换成对应的属性数值点，例如，以分段的方式来转换成数值点，在此以年龄属性为例说明，如年龄为0-16岁的标识为1；17-22岁的标识为2；23-35岁的标识为3；36-50岁的标识为4；51-65岁的标识为5；66岁及以上的标识为7；其他属性以此类推，在此不再赘述。
在过往历史数据的预设归档数据库中保存有曾经进行过风险件判断的历史案件信息，找出历史案件中的所有正常理赔件。在根据上传的保险保单相关信息判断当前保险保单是否为风险件时，有如下公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))
其中，a1，a2至an为本次核单的当前保险保单中第1，2至n个预设属性信息(例如，选择几个较为重要的属性如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)转换后的保单属性数值点，A1，A2至An为预设归档数据库中所有正常理赔件对应的第1，2至n个预设属性信息(与提取的当前保险保单中相同的属性，如年龄、文化程度、收入水平、治疗医院等级、疾病等级、报销金额)转换后的参考属性数值点的集合。公式中，mean(A1)为求A1的均值，决定了高斯分布(即正态分布)概率图的位置，mean((a1-mean(A1))^2)为求a1与mean(A1)的标准差，也是高斯分布(即正态分布)的分布的幅度。Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值， P为所述保险保单中第1，2至n个保单属性数值点的出现概率值的叠加值；最终计算得到的P值即为本次核单的当前保险保单为正常理赔件的概率。
由于上文中提到，在一定区域内,在历史数据记录的若干正常理赔件中每个出现的点空间内周围存在着一个高斯分布的概率空间，即正常理赔件间存在相似性，可利用当前要判断的保险保单是否与历史数据记录的正常理赔件相似来判断当前保险保单是否为风险件。也即当前要判断的保险保单中各个数值点出现在历史数据记录的若干正常理赔件中每个点附近空间位置的概率越高，则当前要判断的保险保单与正常理赔件的相似度越高。具体在公式中，各个点的概率值叠加得到的参数P越高，则当前要判断的保险保单为正常理赔件的可能性越大。
因此，本实施例中可预先设定一个合理的预设概率阈值，若计算得到的当前保单为正常理赔件的概率即P值小于该预设概率阈值，则判断当前保险保单与正常理赔件不相似，即判断当前保险保单为疑似理赔风险件，则交由核保中心人工处理；若P值大于该预设概率阈值，则判断当前保险保单与正常理赔件相似，即判断当前保险保单为正常理赔件，则返回可理赔标记，从而允许其走正常理赔流程。
在预先设定概率阈值时，可将历史数据记录的若干理赔件(包括理赔风险件和正常理赔件)利用公式：
P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*…Gauss(an，mean(An)，mean((an-mean(An))^2))来不断训练及验证概率阈值的合理性，在达到一定准确性后，即可设定好一合理的 预设概率阈值。当然，在后续实际应用过程中，该预设概率阈值也可以被用户根据不同应用场景的需要进行调整，如在要求风险件判断较严格的场景中，可适当调高该阈值；在对风险件判断不太严格的场景中，可适当调低该阈值。更加灵活、实用。
在一可选的实施例中，该方法还包括：
将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
本实施例中，还可引入归档系统，将每一个上传的保险保单与其最终的判定结果进行关联保存并归档，如将当前的所述保险保单与反馈的核保结果(是风险件或正常件)进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。即将保险保单等理赔资料与最终的审批意见(是风险件或正常件，以及最终对风险件的人工核保结果)归档，可对审核意见留证，而且在后续过程中对相同用户以及类似理赔情况可给予参考。
此外，本申请还提供一种计算机可读存储介质，所述计算机可读存储介质存储有保险保单集中核单系统，所述保险保单集中核单系统可被至少一个处理器执行，以使所述至少一个处理器执行如上述实施例中的保险保单集中核单方法的步骤，该保险保单集中核单方法的步骤S10、S20、S30等具体实施过程如上文所述，在此不再赘述。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过 程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件来实现，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端设备(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
以上参照附图说明了本申请的优选实施例，并非因此局限本申请的权利范围。上述本申请实施例序号仅仅为了描述，不代表实施例的优劣。另外，虽然在流程图中示出了逻辑顺序，但是在某些情况下，可以以不同于此处的顺序执行所示出或描述的步骤。
本领域技术人员不脱离本申请的范围和实质，可以有多种变型方案实现本申请，比如作为一个实施例的特征可用于另一实施例而得到又一实施例。凡在运用本申请的技术构思之内所作的任何修改、等同替换和改进，均应在本申请的权利范围之内。

Claims (20)

  1. 一种电子装置，其特征在于，所述电子装置包括存储器、处理器，所述存储器上存储有可在所述处理器上运行的保险保单集中核单系统，所述保险保单集中核单系统被所述处理器执行时实现如下步骤：
    接收待处理的保险保单；
    提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；
    若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。
  2. 如权利要求1所述的电子装置，其特征在于，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件的步骤包括：
    提取所述保险保单中的各个预设属性信息，并将提取的预设属性信息按预设分段转换方式转换成对应的属性数值点；并将转换得到的各个属性数值点代入如下公式：
    P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*Gauss(an，mean(An)，mean((an-mean(An))^2))
    其中，a1至an为所述保险保单中第1至n个预设属性信息转换后的保单属性数值点，A1至An为预设归档数据库中所有正常理赔件对应的第1，至n个预设属性信息转换后的参考属性数值点的集合；
    Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1至n个保单属性数值点的出现概率值的叠加值；
    若P大于预设概率阈值，则判断所述保险保单不是风险件；
    若P小于或等于预设概率阈值，则判断所述保险保单是风险件。
  3. 如权利要求1所述的电子装置，其特征在于，所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配 至对应的用户进行风险件核保。
  4. 如权利要求2所述的电子装置，其特征在于，所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配至对应的用户进行风险件核保。
  5. 如权利要求1所述的电子装置，其特征在于，所述保险保单是医疗保险保单，所述预设属性信息包括：
    被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间、报销金额中的至少一个。
  6. 如权利要求2所述的电子装置，其特征在于，所述保险保单是医疗保险保单，所述预设属性信息包括：
    被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间、报销金额中的至少一个。
  7. 如权利要求1所述的电子装置，其特征在于，所述保险保单集中核单系统被所述处理器执行时还包括：
    将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
  8. 如权利要求2所述的电子装置，其特征在于，所述保险保单集中核单系统被所述处理器执行时还包括：
    将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
  9. 一种保险保单集中核单方法，其特征在于，所述保险保单集中核单方法包括：
    接收待处理的保险保单；
    提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；
    若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。
  10. 如权利要求9所述的保险保单集中核单方法，其特征在于，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件的步骤包括：
    提取所述保险保单中的各个预设属性信息，并将提取的预设属性 信息按预设分段转换方式转换成对应的属性数值点；并将转换得到的各个属性数值点代入如下公式：
    P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*Gauss(an，mean(An)，mean((an-mean(An))^2))
    其中，a1至an为所述保险保单中第1至n个预设属性信息转换后的保单属性数值点，A1至An为预设归档数据库中所有正常理赔件对应的第1至n个预设属性信息转换后的参考属性数值点的集合；
    Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1至n个保单属性数值点的出现概率值的叠加值；
    若P大于预设概率阈值，则判断所述保险保单不是风险件；
    若P小于或等于预设概率阈值，则判断所述保险保单是风险件。
  11. 如权利要求9所述的保险保单集中核单方法，其特征在于，所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配至对应的用户进行风险件核保。
  12. 如权利要求10所述的保险保单集中核单方法，其特征在于，所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配至对应的用户进行风险件核保。
  13. 如权利要求9所述的保险保单集中核单方法，其特征在于，所述保险保单是医疗保险保单，所述预设属性信息包括：
    被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间、报销金额中的至少一个。
  14. 如权利要求10所述的保险保单集中核单方法，其特征在于，所述保险保单是医疗保险保单，所述预设属性信息包括：
    被保人的出险地，性别，年龄，文化程度，工作行业，收入水平，治疗医院等级，疾病等级，持续时间、报销金额中的至少一个。
  15. 如权利要求9所述的保险保单集中核单方法，其特征在于，该方法还包括：
    将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
  16. 如权利要求10所述的保险保单集中核单方法，其特征在于，该方法还包括：
    将所述保险保单与反馈的核保结果进行关联保存，并将所述保险保单与反馈的核保结果存储至预设归档数据库中。
  17. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质上存储有保险保单集中核单系统，所述保险保单集中核单系统被处理器执行时实现如下步骤：
    接收待处理的保险保单；
    提取所述保险保单中的预设属性信息，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件；
    若判断所述保险保单是风险件，则按预设分配方式分配所述保险保单进入风险件核保流程；若判断所述保险保单不是风险件，则直接进入预先设定的正常理赔流程。
  18. 如权利要求17所述的计算机可读存储介质，其特征在于，基于所述预设属性信息并利用预先设定的风险件判定规则判断所述保险保单是否为风险件的步骤包括：
    提取所述保险保单中的各个预设属性信息，并将提取的预设属性信息按预设分段转换方式转换成对应的属性数值点；并将转换得到的各个属性数值点代入如下公式：
    P＝Gauss(a1，mean(A1)，mean((a1-mean(A1))^2))*Gauss(a2，mean(A2)，mean((a2-mean(A2))^2))*Gauss(an，mean(An)，mean((an-mean(An))^2))
    其中，a1至an为所述保险保单中第1至n个预设属性信息转换后的保单属性数值点，A1至An为预设归档数据库中所有正常理赔件对应的第1至n个预设属性信息转换后的参考属性数值点的集合；
    Gauss(an，mean(An)，mean((an-mean(An))^2))为所述保险保单中第n个保单属性数值点在对应的第n个参考属性数值点的集合的高斯分布概率空间中的出现概率值，P为所述保险保单中第1至n个保单属性数值点的出现概率值的叠加值；
    若P大于预设概率阈值，则判断所述保险保单不是风险件；
    若P小于或等于预设概率阈值，则判断所述保险保单是风险件。
  19. 如权利要求17所述的计算机可读存储介质，其特征在于， 所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配至对应的用户进行风险件核保。
  20. 如权利要求18所述的计算机可读存储介质，其特征在于，所述预设分配方式包括：
    通过预先设置的手动获取核保任务触发按钮接收用户发出的手动获取核保任务指令，并将所述保险保单发送至该用户，以供该用户对所述保险保单进行风险件核保；或者，
    按预设任务处理比例将所述保险保单自动分配至对应的用户进行风险件核保；或者，
    根据所述保险保单的保险上报类型或者保险上报地域自动分配至对应的用户进行风险件核保。
PCT/CN2018/089719 2018-03-06 2018-06-03 保险保单集中核单方法、电子装置及可读存储介质 WO2019169768A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810182943.9A CN108346036A (zh) 2018-03-06 2018-03-06 保险保单集中核单方法、电子装置及可读存储介质
CN201810182943.9 2018-03-06

Publications (1)

Publication Number Publication Date
WO2019169768A1 true WO2019169768A1 (zh) 2019-09-12

Family

ID=62957848

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/089719 WO2019169768A1 (zh) 2018-03-06 2018-06-03 保险保单集中核单方法、电子装置及可读存储介质

Country Status (2)

Country Link
CN (1) CN108346036A (zh)
WO (1) WO2019169768A1 (zh)

Families Citing this family (12)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109165849A (zh) * 2018-08-27 2019-01-08 众安信息技术服务有限公司 风险评估方法和装置
CN109492095A (zh) * 2018-10-16 2019-03-19 平安健康保险股份有限公司 理赔数据处理方法、装置、计算机设备及存储介质
CN109360112B (zh) * 2018-10-27 2024-06-25 平安医疗健康管理股份有限公司 一种基于数据处理的精神性疾病认证方法及相关设备
CN109636625A (zh) * 2018-10-31 2019-04-16 深圳壹账通智能科技有限公司 保险电子合同的处理方法及装置、存储介质、计算机设备
CN109636510A (zh) * 2018-11-28 2019-04-16 阿里巴巴集团控股有限公司 一种确定用户风险偏好、理财产品的推荐方法和装置
CN109885736A (zh) * 2019-01-24 2019-06-14 平安科技（深圳）有限公司 异常保单的识别方法、装置、计算机设备及存储介质
CN110009515A (zh) * 2019-03-12 2019-07-12 中国平安财产保险股份有限公司 基于人脸识别的单据校验方法、装置、服务器及介质
CN111008902B (zh) * 2019-11-25 2023-07-18 泰康保险集团股份有限公司 核保数据处理方法、装置、设备和介质
CN111105318A (zh) * 2019-12-29 2020-05-05 北京车与车科技有限公司 一种车险线上智能核保与影像处理方法
CN111652746B (zh) * 2020-05-29 2023-08-29 泰康保险集团股份有限公司 信息生成方法、装置、电子设备及存储介质
CN112419072B (zh) * 2020-10-30 2022-10-14 中汽研（天津）汽车工程研究院有限公司 基于汽车物理属性的汽车保险反渗漏方法
CN113469826B (zh) * 2021-07-22 2022-12-09 阳光人寿保险股份有限公司 一种信息处理的方法、装置、设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106651588A (zh) * 2016-11-09 2017-05-10 前海企保科技（深圳）有限公司 一种物流保险保单的核保方法和装置
CN106682987A (zh) * 2016-11-21 2017-05-17 中国平安财产保险股份有限公司 一种财产险理赔实现方法及实现系统
US20170293990A1 (en) * 2016-04-01 2017-10-12 OneTrust, LLC Data processing systems and methods for implementing audit schedules for privacy campaigns

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105159948B (zh) * 2015-08-12 2019-04-02 成都数联易康科技有限公司 一种基于多特征的医疗保险欺诈检测方法
CN107103658B (zh) * 2017-04-25 2019-05-14 哈尔滨工业大学 基于深度学习与高斯分布的航空发动机气路异常检测方法
CN107240024A (zh) * 2017-05-22 2017-10-10 中国平安人寿保险股份有限公司 保险理赔的反欺诈识别方法及装置

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20170293990A1 (en) * 2016-04-01 2017-10-12 OneTrust, LLC Data processing systems and methods for implementing audit schedules for privacy campaigns
CN106651588A (zh) * 2016-11-09 2017-05-10 前海企保科技（深圳）有限公司 一种物流保险保单的核保方法和装置
CN106682987A (zh) * 2016-11-21 2017-05-17 中国平安财产保险股份有限公司 一种财产险理赔实现方法及实现系统

Also Published As

Publication number Publication date
CN108346036A (zh) 2018-07-31

Similar Documents

Publication Publication Date Title
WO2019169768A1 (zh) 保险保单集中核单方法、电子装置及可读存储介质
JP5586373B2 (ja) 支払請求を処理するコンポーネントの機能をコンピュータシステムに実現させるプログラムが記録されているコンピュータ読み取り可能な記憶媒体、およびコンピュータシステムに支払請求を処理させるコンピュータシステムの動作方法
US9195936B1 (en) System and method for updating or modifying an application without manual coding
US10205729B2 (en) Automation tool for providing users with security access to an electronic system
US20020198750A1 (en) Risk management application and method
WO2017118435A1 (zh) 车险查勘任务的发布系统、设备、方法及可读存储介质
US20160306808A1 (en) Computer-implemented method for determining roof age of a structure
JP2015222596A (ja) 将来の損失に関連する頻度を予測し、損失決定ユニットの関連する自動処理のためのシステム及び方法
US10083061B2 (en) Cloud embedded process tenant system for big data processing
US20080066071A1 (en) Capturing effort level by task upon check-in to source control management system
US20020178036A1 (en) Project management method and project management system
US20210407014A1 (en) Hail data evaluation computer system
CN115547466A (zh) 基于大数据的医疗机构登记评审系统及其方法
US20150193881A1 (en) Computer-implemented method for determining roof age of a structure
US8073717B2 (en) Systems for digital management of underwriting insurance contracts
CN115577983B (zh) 基于区块链的企业任务匹配方法、服务器及存储介质
CN116843481A (zh) 知识图谱分析方法、装置、设备及存储介质
CN111242779A (zh) 金融数据特征选择和预测方法、装置、设备及存储介质
CN114840660A (zh) 业务推荐模型训练方法、装置、设备及存储介质
US8073716B2 (en) Methods for digital management of underwriting insurance contracts
KR20030079852A (ko) 감사 관리 시스템 및 감사 관리 방법
US9710774B2 (en) Configuration of embedded intelligence
US20150213563A1 (en) Methods and Systems of Production System Management
TWI705395B (zh) 顧客價值管理系統
US20240169405A1 (en) Peer-based auditing system and method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18908647

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 09/12/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 18908647

Country of ref document: EP

Kind code of ref document: A1