Insurance calling rule mining method and device, computer equipment and storage medium

Abstract

The invention belongs to the field of artificial intelligence, and relates to an insurance calling rule mining method which comprises the following steps: acquiring original data and processing the original data; constructing an insurance policy fraud factor library by the processed data through a factor combination model; screening a plurality of insurance calling rules from the insurance policy fraud factor library based on genetic programming; and based on a stepwise regression method, obtaining an optimal scheduling rule combination from the multiple insurance scheduling rules. The invention further provides an insurance calling rule mining device, computer equipment and a storage medium. The genetic programming algorithm adopted by the invention can simply and automatically extract the effective calling rules, the useful calling rules can be extracted from a large amount of factor information, the rules are further combined step by step, and the generated optimal rule combination represents the optimal effect of combining all the rules, so that the efficiency of calling is improved. And the data is simplified and the calling effect is ensured as much as possible.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to an insurance adjustment rule mining method, device, computer equipment and storage medium. Background Art

[0002] Insurance fraud identification can help insurance companies take certain measures when underwriting and settling claims, such as imposing certain constraints on the insured, paying special attention to claims or further investigating them, etc., which can effectively reduce insurance fraud. Insurance fraud identification faces great challenges. Traditional insurance companies mainly rely on internal qualitative judgment methods and expert experience to identify insurance fraud. However, due to the diversity of insurance fraud methods and the large and complex amount of data, if traditional insurance fraud identification methods are used alone, it will not only be difficult to accurately identify insurance fraud, but also incur a lot of labor costs. Today, in the issue of accurately identifying insurance fraud, many insurance companies collect and organize the insured's personal information and claims accident information, and use machine learning and artificial intelligence methods to form a set of risk identification rule systems belonging to their own companies. However, there are still some problems in the construction of the risk identification rule base:

[0003] First, there are many risk factors for insurance fraud, and it is difficult to mine rules for fraud detection. Usually, the data related to the fraud risk model covers the entire process of the policy from underwriting to issuance, underwriting, claims, and detection, with many data fields and rich types. At the same time, it is also necessary to construct factors for these data, which include a large amount of text information. After keyword extraction, the dimension will be further increased. Therefore, it is very difficult for both humans and machines to extract useful detection rules from a large amount of relevant information.

[0004] Second, the automation level of rule mining is low and the proportion of manual work is high. Due to the above-mentioned defects of machine learning, most companies currently still rely on manual summary and adjustment of rules, which has a low degree of automation and is difficult to summarize more complex and in-depth insurance risk adjustment rules. Summary of the invention

[0005] The purpose of the embodiments of the present application is to propose a method, device, computer equipment and storage medium for mining insurance adjustment rules, the main purpose of which is to improve the efficiency and accuracy of mining insurance adjustment rules.

[0006] In order to solve the above technical problems, the embodiment of the present application provides an insurance adjustment rule mining method, which adopts the following technical solution:

[0007] A method for mining insurance claim rules comprises the following steps:

[0008] Acquiring raw data and processing the raw data;

[0009] The processed data is used to construct a policy fraud factor library through a factor combination model;

[0010] Screening multiple insurance claim adjustment rules from the insurance policy fraud factor library based on genetic programming;

[0011] Based on the stepwise regression method, the optimal combination of insurance adjustment rules is obtained from multiple insurance adjustment rules.

[0012] Furthermore, the step of obtaining the original data and processing the original data specifically includes:

[0013] Obtain continuous variables including at least amount and time, categorical variables including at least gender, location, and policy type, and text variables including at least the course of the accident and diagnosis results;

[0014] The continuous variables are subjected to factor combination processing, the categorical variables are subjected to one-hot processing, and the text variables are subjected to keyword extraction processing.

[0015] Furthermore, the step of selecting a plurality of insurance adjustment rules from the insurance policy fraud factor library based on genetic programming specifically includes:

[0016] Construct multiple tree formula structures based on the policy fraud factor library;

[0017] Randomly generate a set of first-generation formulas for each tree-shaped formula structure;

[0018] The negative cross entropy between the classification result of the insurance adjustment rule and the real result is used as the fitness function, the value of the fitness function of each first-generation formula is calculated, and the one with the highest value is selected as the parent generation of the next generation evolution;

[0019] The parent generation forms different offspring formulas through crossover, subtree mutation, point mutation, and boosting mutation;

[0020] The offspring formulas cycle through the next round of evolution and continue to multiply and mutate until the preset number of iterations is reached or the population converges to a unique solution, and the insurance adjustment rule corresponding to the final formula of each tree-shaped formula structure is selected;

[0021] The insurance adjustment rules corresponding to the final formula of each tree-shaped formula structure are combined to form a plurality of insurance adjustment rules.

[0022] Furthermore, the step of constructing a tree formula structure based on the policy fraud factor library specifically includes:

[0023] Obtain the number of individuals in the population based on the insurance policy fraud factor library and the genetic planning formula table;

[0024] A tree formula structure is constructed according to the number of individuals and the genetic planning formula table.

[0025] Furthermore, the step of forming different offspring formulas through crossover, subtree mutation, point mutation, and boosting mutation of the parent generation specifically includes:

[0026] Randomly select a cotyledon from the parent generation and replace it with a cotyledon in the formula tree with the next highest fitness;

[0027] Randomly select a cotyledon from the parent and replace it with any random subtree;

[0028] Randomly select a cotyledon from the parent and replace it with a single point at random;

[0029] A subtree is randomly selected from the mutated formula tree for pruning to obtain different offspring formulas.

[0030] Furthermore, the genetic planning formula table includes a logic formula table and a judgment table: the logic formula table includes a logic formula name and a corresponding number of parameters, and the judgment table includes a judgment formula name and a number of parameters; the tree formula structure is constructed according to the number of individuals and the genetic planning formula table, including:

[0031] Select the name of the logical formula of the individual;

[0032] The tree-shaped formula structure is formed by taking the name of each individual logic formula as an internal node and the name of the judgment formula as a leaf node.

[0033] Furthermore, the step of deriving the optimal combination of insurance adjustment rules from multiple insurance adjustment rules based on the stepwise regression method specifically includes:

[0034] Step 1: Get each unused rule and add it to the rule combination to form a new combination;

[0035] Step 2: Calculate the classification result of the new combination;

[0036] Step 3, compare the fitness of the classification results of the new combination and the old combination;

[0037] Step 4: If the fitness is improved, the new combination is retained as the original combination for adding rules next time;

[0038] Step 5: If the fitness remains unchanged or decreases, the original combination is retained;

[0039] Step 6, repeat steps 1 to 5 until there are no more individual rules that meet the conditions;

[0040] Step seven, output the combination of rules obtained by genetic programming as the optimal adjustment rule combination.

[0041] In order to solve the above technical problems, the embodiment of the present application also provides an insurance adjustment rule mining device, which adopts the following technical solution:

[0042] An insurance adjustment rule mining device, comprising:

[0043] An acquisition module, used for acquiring original data and processing the original data;

[0044] A construction module is used to construct a policy fraud factor library using the processed data through a factor combination model;

[0045] A genetic planning module, used for selecting a plurality of insurance adjustment rules from the insurance policy fraud factor library based on genetic planning;

[0046] The stepwise regression module is used to obtain the optimal combination of insurance adjustment rules from multiple insurance adjustment rules based on the stepwise regression method.

[0047] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0048] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the above-mentioned insurance adjustment rule mining method when executing the computer-readable instructions.

[0049] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0050] A computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the insurance adjustment rule mining method as described above.

[0051] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects: the present application adopts the method of obtaining and processing the original data, constructing a policy fraud factor library through a factor combination model of the processed data, and then screening multiple insurance adjustment rules from the policy fraud factor library based on genetic programming, and deriving the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The genetic programming algorithm can simply and automatically extract effective adjustment rules, and finally derive the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The present application can extract useful adjustment rules from a large amount of factor information, and further gradually combine the rules. The generated optimal rule combination represents the optimal effect of all rules combined, which simplifies the data while ensuring the adjustment effect as much as possible. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0054] Figure 2 A flowchart of an embodiment of the insurance adjustment rule mining method according to the present application;

[0055] Figure 3 This Figure 2 A flowchart of an embodiment of step S201;

[0056] Figure 4 It is a common formula structure diagram in the prior art;

[0057] Figure 5 It is a tree-shaped formula structure designed by the instinctive application technology;

[0058] Figure 6 This Figure 2 A flowchart of an embodiment of step S203;

[0059] Figure 7 This Figure 4 A flowchart of an embodiment of step S2031;

[0060] Figure 8 is a rule tree of age and insurance coverage interval in the embodiment of the present application;

[0061] Fig. 9 This Figure 4 A flowchart of an embodiment of step S2033;

[0062] Fig.10 This Figure 2 A flowchart of an embodiment of step S204;

[0063] Fig.11 It is a structural schematic diagram of an embodiment of an insurance adjustment rule mining device according to the present application;

[0064] Fig.12 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0065] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0066] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0067] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0068] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0069] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0070] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0071] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0072] It should be noted that the insurance adjustment rule mining method provided in the embodiment of the present application is generally executed by a server, and accordingly, the insurance adjustment rule mining device is generally set in a terminal device.

[0073] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0074] Continue to refer Figure 2 , shows a flow chart of an embodiment of the insurance adjustment rule mining method according to the present application. The insurance adjustment rule mining method comprises the following steps:

[0075] Step S201, obtaining original data and processing the original data.

[0076] Specific, combined Figure 3 As shown, step S201 specifically includes:

[0077] Step S2011, obtaining continuous variables including at least amount and time, categorical variables including at least gender, location, and policy type, and text variables including at least the course of the accident and the diagnosis result.

[0078] Step S2012, performing factor combination processing on the continuous variables, performing one-hot processing on the categorical variables, and performing keyword extraction processing on the text variables.

[0079] Step S202: The processed data is used to construct a policy fraud factor library through a factor combination model.

[0080] In this embodiment, the factor combination model may select an existing model, and a policy fraud factor library may be constructed based on the existing factor combination model.

[0081] Step S203, screening multiple insurance adjustment rules from the insurance policy fraud factor library based on genetic programming.

[0082] In this embodiment, genetic programming is one of the important technologies in the field of evolutionary computing and can be used in various industries. It is currently often used in factor mining. Genetic programming gradually generates a group of formulas that meet specific goals by simulating the process of biological genetic evolution in nature to generate excellent offspring.

[0083] Genetic programming starts from a randomly generated group of formulas and gradually generates a group of formulas that meet specific goals through genetic evolution. The formulas in the algorithm are mainly expressed in the form of a tree, which is composed of internal nodes and leaf nodes. The internal nodes are called function nodes, which contain functions and algorithm operations. This type of node constitutes a function set. The leaf nodes are called terminals, which contain variables, constants, and functions without parameters. This type of node constitutes a terminal set. Common formula tree structures are as follows: Figure 4 As shown, the example formula is X1+X2.

[0084] Among them, Plus is an internal node, and X1 and X2 are leaf terminal nodes. This application uses genetic programming for rule mining and designs different tree formula structures. For example (X1>0 and X2==1) Figure 5 shown.

[0085] Combination Figure 6 As shown, the specific steps of screening multiple insurance adjustment rules from the insurance policy fraud factor library based on genetic programming in this application include:

[0086] Step S2031, constructing a tree-like multiple formula structure based on the insurance policy fraud factor library.

[0087] Step S2032, randomly generate a set of first-generation formulas for each tree-shaped formula structure.

[0088] Step S2033, using the negative cross entropy of the classification result of the insurance adjustment rule and the true result as the fitness function, calculate the value of the fitness function of each first-generation formula, and select the one with the highest value as the parent generation of the next generation evolution.

[0089] In this embodiment, cross entropy is an important concept in information theory, which is used to measure the difference information between two probability distributions. The calculation process of cross entropy is as follows:

[0090]

[0091] Since the higher the cross entropy, the greater the difference between the two results, this embodiment uses negative cross entropy as the fitness function in the genetic programming process. The higher the value of this function, the better the classification result of the rule.

[0092] Step S2034, the parent generation forms different offspring formulas through crossover, subtree mutation, point mutation, and boosting mutation.

[0093] Step S2035, the offspring formula cycles through the next round of evolution and continues to multiply and mutate until a preset number of iterations is reached or the population converges to a unique solution, and the insurance adjustment rule corresponding to the final formula of each tree-shaped formula structure is selected.

[0094] Step S2036, combining the insurance adjustment rules corresponding to the final formula of each tree formula structure to form a plurality of insurance adjustment rules.

[0095] Combination Figure 7 As shown, the steps of constructing a tree formula structure based on the policy fraud factor library specifically include:

[0096] Step S20311, obtain the number of individuals in the population based on the insurance policy fraud factor library and the genetic planning formula table.

[0097] Step S20312, constructing a tree formula structure according to the number of individuals and the genetic planning formula table.

[0098] In this embodiment, the genetic planning formula table includes a logic formula table and a judgment table: the logic formula table includes the logic formula name and the corresponding number of parameters, as shown in Table 1, and the judgment table includes the judgment formula name and the number of parameters, as shown in Table 2.

[0099] Table 1

[0100] Formula Name Number of parameters explain And 2 and Or 2 or Not 1 No

[0101] Table 2

[0102] Formula Name Number of parameters explain > 2 Greater than >＝ 2 Greater than or equal to < 2 Less than <＝ 2 Less than or equal to ＝＝ 2 equal

[0103] In this embodiment, for example, the rule tree of age and insurance coverage interval is as follows: Figure 8 shown.

[0104] The constructing of a tree-shaped formula structure according to the number of individuals and the genetic planning formula table comprises:

[0105] First, the name of the individual logic formula is selected.

[0106] Then, the tree-shaped formula structure is formed with the name of each individual logic formula as an internal node and the name of the judgment formula as a leaf node.

[0107] In this embodiment, the core operation of the genetic programming algorithm is the reproduction operation, which mainly includes three operations: selection, crossover and mutation. These three operations are combined to generate a new generation of population. Specifically, firstly, an excellent individual is selected through the selection operation, and then the crossover and mutation operations are performed on the individual to generate a new individual, and finally a new generation of population is obtained.

[0108] The selection operation is to select the excellent individuals in the current population as the basis of the next generation population. There are various ways of selection operation, such as roulette selection, competitive selection, ranking selection, etc. In the mining of insurance adjustment rules, the ranking selection method can be used for selection operation, and the formula tree with the highest fitness is used as the parent, that is, the one with the highest selection value in step S2033 is used as the parent of the next generation evolution.

[0109] See also Fig. 9 As shown, in this embodiment, the steps of forming different offspring formulas through crossover, subtree mutation, point mutation, and lifting mutation of the parent generation specifically include:

[0110] Step S20341, crossover operation, i.e. randomly selecting cotyledons from the parent generation and replacing them with cotyledons in the formula tree with the second highest fitness; this operation inherits the characteristics of the parent generation on the one hand, and retains the excellent attribute of higher fitness on the other hand. It is one of the most important operations in the genetic algorithm and can effectively increase the diversity and optimization effect of the population.

[0111] Step S20342, subtree mutation randomly selects a cotyledon from the parent generation and replaces it with any random subtree. This has a higher uncertainty, but has a certain effect on the diversity of the population.

[0112] Step S20343, point mutation randomly selects a cotyledon from the parent generation and replaces it with a random single point. Compared with subtree mutation, point mutation replaces it with a random single point, which is more beneficial to the diversity of the population. However, due to the smaller range of point mutation, the improvement effect may also be relatively small.

[0113] Step S20344 is an operation to combat excessive variation, which randomly selects a subtree from the mutated formula tree for pruning to obtain different offspring formulas. Under the premise of ensuring diversity, this variation is more conducive to simplifying the formula.

[0114] Step S204, deriving the optimal combination of insurance adjustment rules from multiple insurance adjustment rules based on a stepwise regression method.

[0115] Combination Fig.10 As shown, step S204 specifically includes:

[0116] Step S2041, obtain each unused rule, and add the rule to the rule combination to form a new combination.

[0117] Step S2042, calculate the classification result of the new combination, and the calculation principle is: if any rule in the combination determines that it is a raise, then it is raised.

[0118] Step S2043, comparing the fitness of the classification results of the new combination and the old combination.

[0119] Step S2044: If the fitness is improved, the new combination is retained as the original combination for adding rules next time.

[0120] Step S2045: If the fitness remains unchanged or decreases, the original combination is retained without change.

[0121] Step S2046, repeat steps S2401-S2405 until there are no separate rules that meet the conditions.

[0122] Step S2047, output the combination of rules obtained by genetic programming as the optimal adjustment rule combination.

[0123] The present application obtains and processes the original data, constructs a policy fraud factor library through a factor combination model for the processed data, and then screens multiple insurance adjustment rules from the policy fraud factor library based on genetic programming, and obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The genetic programming algorithm can simply and automatically extract effective adjustment rules, and finally obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The present application can extract useful adjustment rules from a large amount of factor information, and further gradually combine the rules. The generated optimal rule combination represents the optimal effect of all rules combined, which simplifies the data while ensuring the adjustment effect as much as possible.

[0124] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0125] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0126] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0127] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0128] Further references Fig.11 , as a response to the above Figure 2 The present application provides an embodiment of an insurance adjustment rule mining device, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0129] like Fig.11 As shown, the insurance adjustment rule mining device 1100 described in this embodiment includes: an acquisition module 110, a construction module 120, a genetic programming module 130 and a stepwise regression module 140. Among them:

[0130] The acquisition module 110 is used to acquire original data and process the original data.

[0131] The construction module 120 is used to construct a policy fraud factor library using the processed data through a factor combination model.

[0132] The genetic programming module 130 is used to select multiple insurance adjustment rules from the policy fraud factor library based on genetic programming.

[0133] The stepwise regression module 140 is used to obtain an optimal combination of adjustment rules from a plurality of insurance adjustment rules based on a stepwise regression method.

[0134] The present application adopts the acquisition module 110 to acquire the original data and process the original data, and constructs a policy fraud factor library with the processed data through a factor combination model, and then screens multiple insurance adjustment rules from the policy fraud factor library based on genetic programming according to the construction module 120. The genetic programming module 130 obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The genetic programming algorithm can simply and automatically extract effective adjustment rules. Finally, the stepwise regression module 140 obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The present application can extract useful adjustment rules from a large amount of factor information, and further gradually combine the rules. The generated optimal rule combination represents the optimal effect of all rules combined, which simplifies the data while ensuring the adjustment effect as much as possible.

[0135] To solve the above technical problems, the present application also provides a computer device. Fig.12 , Fig.12 This is a basic structural block diagram of the computer device in this embodiment.

[0136] The computer device 12 includes a memory 121, a processor 122, and a network interface 123 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 12 with components 71-73, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0137] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0138] The memory 121 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 121 can be an internal storage unit of the computer device 12, such as a hard disk or memory of the computer device 12. In other embodiments, the memory 121 can also be an external storage device of the computer device 12, such as a plug-in hard disk equipped on the computer device 12, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 121 can also include both the internal storage unit of the computer device 12 and its external storage device. In this embodiment, the memory 121 is generally used to store the operating system and various application software installed on the computer device 12, such as computer-readable instructions of the insurance adjustment rule mining method, etc. In addition, the memory 121 can also be used to temporarily store various types of data that have been output or are to be output.

[0139] The processor 122 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 122 is generally used to control the overall operation of the computer device 12. In this embodiment, the processor 122 is used to run computer-readable instructions or process data stored in the memory 121, such as computer-readable instructions for running the insurance adjustment rule mining method.

[0140] The network interface 123 may include a wireless network interface or a wired network interface. The network interface 123 is generally used to establish a communication connection between the computer device 12 and other electronic devices.

[0141] The present application obtains and processes the original data, constructs a policy fraud factor library through a factor combination model for the processed data, and then screens multiple insurance adjustment rules from the policy fraud factor library based on genetic programming, and obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The genetic programming algorithm can simply and automatically extract effective adjustment rules, and finally obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The present application can extract useful adjustment rules from a large amount of factor information, and further gradually combine the rules. The generated optimal rule combination represents the optimal effect of all rules combined, which simplifies the data while ensuring the adjustment effect as much as possible.

[0142] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the insurance adjustment rule mining method as described above.

[0143] The present application obtains and processes the original data, constructs a policy fraud factor library through a factor combination model for the processed data, and then screens multiple insurance adjustment rules from the policy fraud factor library based on genetic programming, and obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The genetic programming algorithm can simply and automatically extract effective adjustment rules, and finally obtains the optimal adjustment rule combination from multiple insurance adjustment rules based on the stepwise regression method. The present application can extract useful adjustment rules from a large amount of factor information, and further gradually combine the rules. The generated optimal rule combination represents the optimal effect of all rules combined, which simplifies the data while ensuring the adjustment effect as much as possible.

[0144] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0145] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for mining insurance adjustment rules, characterized in that: The steps include: Acquiring raw data and processing the raw data; The processed data is used to construct a policy fraud factor library through a factor combination model; Screening multiple insurance claim adjustment rules from the insurance policy fraud factor library based on genetic programming; Based on the stepwise regression method, the optimal combination of insurance adjustment rules is obtained from multiple insurance adjustment rules.

2. The insurance claim rule mining method according to claim 1, characterized in that: The step of obtaining the original data and processing the original data specifically includes: Obtain continuous variables including at least amount and time, categorical variables including at least gender, location, and policy type, and text variables including at least the course of the accident and diagnosis results; The continuous variables are subjected to factor combination processing, the categorical variables are subjected to one-hot processing, and the text variables are subjected to keyword extraction processing.

3. The insurance claim rule mining method according to claim 1, characterized in that: The step of selecting a plurality of insurance adjustment rules from the insurance policy fraud factor library based on genetic programming specifically includes: Construct multiple tree formula structures based on the policy fraud factor library; Randomly generate a set of first-generation formulas for each tree-shaped formula structure; The negative cross entropy between the classification result of the insurance adjustment rule and the real result is used as the fitness function, the value of the fitness function of each first-generation formula is calculated, and the one with the highest value is selected as the parent generation of the next generation evolution; The parent generation forms different offspring formulas through crossover, subtree mutation, point mutation, and boosting mutation; The offspring formulas cycle through the next round of evolution and continue to multiply and mutate until the preset number of iterations is reached or the population converges to a unique solution, and the insurance adjustment rule corresponding to the final formula of each tree-shaped formula structure is selected; The insurance adjustment rules corresponding to the final formula of each tree-shaped formula structure are combined to form a plurality of insurance adjustment rules.

4. The insurance claim rule mining method according to claim 3 is characterized in that: The step of constructing a tree formula structure based on the policy fraud factor library specifically includes: Obtain the number of individuals in the population based on the insurance policy fraud factor library and the genetic planning formula table; A tree formula structure is constructed according to the number of individuals and the genetic planning formula table.

5. The insurance claim rule mining method according to claim 3 is characterized in that: The step of forming different offspring formulas through crossover, subtree mutation, point mutation, and boosting mutation of the parent generation specifically includes: Randomly select a cotyledon from the parent generation and replace it with a cotyledon in the formula tree with the next highest fitness; Randomly select a cotyledon from the parent and replace it with any random subtree; Randomly select a cotyledon from the parent and replace it with a single point at random; A subtree is randomly selected from the mutated formula tree for pruning to obtain different offspring formulas.

6. The insurance claim rule mining method according to claim 4, characterized in that: The genetic programming formula table includes a logic formula table and a judgment table: the logic formula table includes the logic formula name and the corresponding parameter number, and the judgment table includes the judgment formula name and the parameter number; The constructing of a tree-shaped formula structure according to the number of individuals and the genetic planning formula table comprises: Select the name of the logical formula of the individual; The tree-shaped formula structure is formed by taking the name of each individual logic formula as an internal node and the name of the judgment formula as a leaf node.

7. The insurance claim rule mining method according to claim 1, characterized in that: The step of deriving the optimal combination of insurance adjustment rules from multiple insurance adjustment rules based on the stepwise regression method specifically includes: Step 1: Get each unused rule and add it to the rule combination to form a new combination; Step 2: Calculate the classification result of the new combination; Step 3, compare the fitness of the classification results of the new combination and the old combination; Step 4: If the fitness is improved, the new combination is retained as the original combination for adding rules next time; Step 5: If the fitness remains unchanged or decreases, the original combination is retained; Step 6, repeat steps 1 to 5 until there are no more individual rules that meet the conditions; Step seven, output the combination of rules obtained by genetic programming as the optimal adjustment rule combination.

8. An insurance adjustment rule mining device, characterized in that: include: An acquisition module, used for acquiring original data and processing the original data; A construction module is used to construct a policy fraud factor library using the processed data through a factor combination model; A genetic planning module, used for selecting a plurality of insurance adjustment rules from the insurance policy fraud factor library based on genetic planning; The stepwise regression module is used to obtain the optimal combination of insurance adjustment rules from multiple insurance adjustment rules based on the stepwise regression method.

9. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, it implements the steps of the insurance adjustment rule mining method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the insurance adjustment rule mining method as described in any one of claims 1 to 7.

Images