Risk assessment method and device based on machine learning, computer equipment and medium

Abstract

The invention provides a risk assessment method and device based on machine learning, computer equipment and a medium, and belongs to the technical field of artificial intelligence. The method comprises the steps of obtaining historical claim settlement data and user financial data of a to-be-evaluated agricultural insurance, and obtaining growth data, geographic data and meteorological data of an insurance object crop of the to-be-evaluated agricultural insurance; generating agricultural insurance feature data of the to-be-evaluated agricultural insurance based on the historical claim settlement data, the user financial data, the growth data, the geographic data and the meteorological data; inputting the agricultural insurance feature data into a pre-trained risk assessment model; the risk assessment model is obtained by performing model training on a preset agricultural insurance feature sample data set based on a machine learning algorithm; and calculating the agricultural insurance feature data through a risk assessment model to obtain a risk level assessment result of the to-be-assessed agricultural insurance. According to the invention, the cost of risk assessment of agricultural insurance can be reduced and the accuracy of the risk assessment result can be improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a risk assessment method, apparatus, computer equipment and medium based on machine learning. Background Art

[0002] As an important tool for dispersing and transferring agricultural production risks, agricultural insurance plays an important role in protecting farmers' interests and promoting agricultural development.

[0003] At present, the risk rating of agricultural insurance mainly relies on manual on-site inspection and loss assessment, and then makes a comprehensive judgment based on the real-time dynamic situation in agricultural production. However, due to the characteristics of agricultural production, which has the characteristics of vast crop planting areas, and the workload of fully capturing the complex factors of real-time changes in agricultural production, the risk rating of agricultural insurance requires a very high time and human resource cost.

[0004] Therefore, how to reduce the cost of risk assessment for agricultural insurance has become a technical problem that needs to be solved urgently. Summary of the invention

[0005] The main purpose of the embodiments of the present application is to propose a risk assessment method, device, computer equipment and medium based on machine learning, aiming to reduce the cost of risk assessment for agricultural insurance.

[0006] To achieve the above objectives, a first aspect of an embodiment of the present application proposes a risk assessment method based on machine learning, the method comprising:

[0007] Acquire historical claims data and user financial data of the agricultural insurance to be evaluated, and acquire growth data, geographic data, and meteorological data of the insured crops of the agricultural insurance to be evaluated;

[0008] Generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data;

[0009] The agricultural insurance characteristic data is input into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm;

[0010] The agricultural insurance characteristic data is calculated by the risk assessment model to obtain a risk level assessment result of the agricultural insurance to be assessed.

[0011] In some embodiments, the step of obtaining the historical claims data and user financial data of the agricultural insurance to be evaluated, and obtaining the growth data, geographic data, and meteorological data of the insured crops of the agricultural insurance to be evaluated, includes:

[0012] Obtaining risk assessment indicators from a preset agricultural insurance assessment indicator system; the agricultural insurance assessment indicator system represents a reference standard for risk assessment of agricultural insurance;

[0013] Based on the risk assessment indicators, historical claims data and user financial data of the agricultural insurance to be assessed are collected; and based on the risk assessment indicators, growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be assessed are collected.

[0014] In some embodiments, before obtaining the risk assessment index from the preset agricultural risk assessment index system, the method further includes:

[0015] Obtain historical agricultural insurance business data;

[0016] Based on the business characteristics analysis of the historical business data of agricultural insurance, multiple consideration factors for risk assessment of agricultural insurance are obtained;

[0017] Based on the above-mentioned multiple considerations, an agricultural insurance assessment index system is constructed;

[0018] Among them, one of the multiple considerations corresponds to at least one risk assessment indicator in the agricultural risk assessment indicator system.

[0019] In some embodiments, the historical claim data includes: historical claim time and historical claim amount;

[0020] The generating of the agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data comprises:

[0021] Numerical processing is performed on the growth data, the geographic data and the meteorological data to obtain feature codes corresponding to the growth data, the geographic data and the meteorological data respectively;

[0022] The characteristic codes corresponding to the growth data, the geographic data and the meteorological data, and the historical claim settlement time, the historical claim settlement amount and the user financial data are respectively normalized to obtain the agricultural insurance characteristic data of the agricultural insurance to be evaluated.

[0023] In some embodiments, the method further comprises:

[0024] Obtain agricultural insurance feature sample dataset;

[0025] The agricultural insurance characteristic sample data set is divided into multiple sub-window sample data by time series; the time series of two adjacent sub-window sample data in the multiple sub-window sample data are different;

[0026] Using the plurality of sub-window sample data as a model input of a preset machine learning algorithm;

[0027] The sliding window model is trained on the multiple sub-window sample data through the machine learning algorithm to obtain a risk assessment model.

[0028] In some embodiments, after obtaining the agricultural insurance feature sample data set, the method further includes:

[0029] Dividing the agricultural insurance characteristic sample data set by region to obtain a plurality of sub-region sample data; the plurality of sub-region sample data each contain a different geographical location;

[0030] Using the plurality of sub-region sample data as a model input of a preset machine learning algorithm;

[0031] The machine learning algorithm is used to perform regional model training on the multiple sub-region sample data to obtain a risk assessment model.

[0032] In some embodiments, the machine learning algorithm includes a first machine learning algorithm and a second machine learning algorithm;

[0033] The method of performing regional model training on the plurality of sub-region sample data by using the machine learning algorithm to obtain a risk assessment model includes:

[0034] Performing model training on first sample data among the plurality of sub-region sample data by using the first machine learning algorithm to obtain a first evaluation model; the first sample data is sub-region sample data including the first geographical location among the plurality of sub-region sample data;

[0035] Performing model training on second sample data among the multiple sub-region sample data by the second machine learning algorithm to obtain a second evaluation model; the second sample data is sub-region sample data including a second geographical location among the multiple sub-region sample data, and the second geographical location is different from the first geographical location;

[0036] The first assessment model and the second assessment model are integrated to obtain a risk assessment model.

[0037] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application proposes a risk assessment device based on machine learning, the device comprising:

[0038] An acquisition module, used to acquire historical claims data and user financial data of the agricultural insurance to be evaluated, and to acquire growth data, geographic data and meteorological data of the insured crops of the agricultural insurance to be evaluated;

[0039] A data processing module, used for generating agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data;

[0040] A model input module, used to input the agricultural insurance characteristic data into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm;

[0041] The model calculation module is used to calculate the agricultural insurance characteristic data through the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be assessed.

[0042] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes a computer device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the method described in the first aspect when executing the computer program.

[0043] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0044] To achieve the above-mentioned purpose, a fourth aspect of an embodiment of the present application proposes a computer program product, wherein the computer program product stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0045] The machine learning-based risk assessment method, apparatus, computer equipment, computer-readable storage medium and computer program product proposed in the embodiments of the present application obtain historical claims data and user financial data of the agricultural insurance to be assessed, as well as growth data, geographic data and meteorological data of the insured crops of the agricultural insurance to be assessed; generate agricultural insurance feature data of the agricultural insurance to be assessed based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data; input the agricultural insurance feature data into a pre-trained risk assessment model; the risk assessment model is obtained by model training a preset agricultural insurance feature sample data set based on a machine learning algorithm; the agricultural insurance feature data is calculated by the risk assessment model to obtain a risk level assessment result of the agricultural insurance to be assessed.

[0046] Compared with the method of conducting risk assessment of agricultural insurance through manual on-site investigation, the embodiment of the present application obtains a risk assessment model that can be used for agricultural planting insurance by training with a machine learning algorithm, so that it is only necessary to obtain the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, and then generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, user financial data, growth data, geographical data and meteorological data of the insured crops, and then input the agricultural insurance characteristic data into the risk assessment model, and then the agricultural insurance characteristic data can be calculated by the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be evaluated.

[0047] In this way, the embodiment of the present application uses machine learning algorithms to intelligently conduct risk assessments on agricultural insurance, without relying on a large number of manpower to conduct on-site surveys and capture real-time changes and complex factors in agricultural production, thereby effectively reducing the time and labor costs of risk assessments on agricultural insurance. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A schematic diagram of the steps of the risk assessment method based on machine learning provided in some embodiments of the present application;

[0049] Figure 2 for Figure 1 Schematic diagram of detailed process flow of step S102;

[0050] Figure 3 for Figure 1 Schematic diagram of detailed process flow of step S101;

[0051] Figure 4 A schematic diagram of the steps of the risk assessment method based on machine learning provided in the embodiment of the present application in other embodiments;

[0052] Figure 5 A schematic diagram of a model training step flow involved in a risk assessment method based on machine learning provided in an embodiment of the present application;

[0053] Figure 6 A schematic diagram of another model training step flow involved in the machine learning-based risk assessment method provided in an embodiment of the present application;

[0054] Figure 7 is a schematic diagram of the structure of a risk assessment device based on machine learning provided in an embodiment of the present application;

[0055] Figure 8 It is a schematic diagram of the hardware structure of the computer device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0057] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0059] First, some nouns involved in this application are analyzed:

[0060] Artificial intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. AI is a branch of computer science that attempts to understand the essence of intelligence and produce a new type of intelligent machine that can respond in a similar way to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems. AI can simulate the information process of human consciousness and thinking. AI is also a 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.

[0061] Natural language processing (NLP): NLP uses computers to process, understand and apply human languages ​​(such as Chinese, English, etc.). NLP is a branch of artificial intelligence and an interdisciplinary subject between computer science and linguistics. It is often referred to as computational linguistics. Natural language processing includes grammatical analysis, semantic analysis, and text understanding. Natural language processing is often used in technical fields such as machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intent recognition, information extraction and filtering, text classification and clustering, public opinion analysis and opinion mining. It involves data mining related to language processing, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computing.

[0062] Information Extraction: A text processing technology that extracts specified types of entity, relationship, event and other factual information from natural language text and forms structured data output. Information extraction is a technology that extracts specific information from text data. Text data is composed of some specific units, such as sentences, paragraphs, and chapters. Text information is composed of some small specific units, such as characters, words, phrases, sentences, paragraphs, or a combination of these specific units. Extracting noun phrases, names, place names, etc. from text data are all text information extraction. Of course, the information extracted by text information extraction technology can be various types of information.

[0063] Next, the overall concept of the embodiments of the present application is described.

[0064] As a basic industry of the national economy, the stable development of agriculture is of great significance to the national economic security and social stability. Agricultural production is affected by many factors such as the natural environment and market fluctuations, and the risk is relatively high. As an important tool for dispersing and transferring agricultural risks, agricultural insurance plays an important role in protecting the interests of farmers and promoting agricultural development. However, there are many shortcomings in the current risk rating of agricultural insurance. For example, relevant technologies often rely on historical data and the experience and judgment of business personnel, and because it is difficult to fully capture the real-time changes and complex factors in agricultural production, the risk assessment results are not accurate enough; at the same time, due to the lack of flexibility and dynamism due to the reliance on historical data and the experience of business personnel, the risk rating of agricultural insurance cannot respond to changes in the agricultural production environment in a timely manner.

[0065] Based on the above-mentioned deficiencies in agricultural insurance risk rating, insurance companies have encountered many obstacles in operating agricultural insurance business. On the one hand, due to the complex market situation of agricultural insurance business, it is difficult for insurance companies to price agricultural insurance business; on the other hand, due to the many constraints on agricultural insurance claims, insurance companies often rely on manual investigation and loss assessment. Therefore, exploring an intelligent risk rating method for agricultural insurance customers has important theoretical significance and practical value.

[0066] Based on this, the embodiments of the present application provide a risk assessment method, apparatus, computer equipment, computer-readable storage medium and computer program product based on machine learning, which obtains historical claims data and user financial data of the agricultural insurance to be evaluated, and obtains growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated; based on the historical claims data, the user financial data, the growth data, the geographical data and the meteorological data, generates agricultural insurance feature data of the agricultural insurance to be evaluated; inputs the agricultural insurance feature data into a pre-trained risk assessment model; the risk assessment model is obtained by model training a preset agricultural insurance feature sample data set based on a machine learning algorithm; the agricultural insurance feature data is calculated by the risk assessment model to obtain a risk level assessment result of the agricultural insurance to be evaluated.

[0067] Compared with the method of conducting risk assessment of agricultural insurance through manual on-site investigation, the embodiment of the present application obtains a risk assessment model that can be used for agricultural insurance by training with a machine learning algorithm, so that it is only necessary to obtain the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, and then generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, user financial data, growth data, geographical data and meteorological data of the insured crops, and then input the agricultural insurance characteristic data into the risk assessment model, and then the agricultural insurance characteristic data can be calculated by the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be evaluated.

[0068] In this way, the embodiment of the present application uses machine learning algorithms to intelligently conduct risk assessments on agricultural insurance, which eliminates the need to rely on a large number of manual on-site inspections and capture real-time changes and complex factors in agricultural production, thereby effectively reducing the time and labor costs of risk assessments on agricultural insurance. In addition, the embodiment of the present application uses machine learning algorithms to conduct risk assessments on agricultural insurance, which can also improve the insurance company's intelligent operation level in agricultural insurance, thereby reducing the labor costs and claims risks of agricultural insurance business, and effectively contributing to the healthy development of agricultural insurance.

[0069] Based on the overall concept of the present application, the specific embodiments of the risk assessment method, device, computer equipment, computer readable storage medium and computer program product based on machine learning provided in the embodiments of the present application are proposed. First, the specific embodiments of the risk assessment method based on machine learning in the embodiments of the present application are described in detail.

[0070] The risk assessment method based on machine learning provided in the embodiment of the present application relates to the field of artificial intelligence technology. The embodiment 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.

[0071] 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.

[0072] It should be noted that in each specific implementation of the embodiments of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain sensitive personal information of the user, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0073] It should be noted that the risk assessment method based on machine learning provided in the embodiments of the present application can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements a risk assessment method based on machine learning, etc., but is not limited to the above forms.

[0074] In addition, the present application embodiment can also be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer computer devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0075] For ease of understanding and explanation, the following text will take the risk assessment method based on machine learning provided by the terminal device application embodiment of the present application as an example for detailed description. The implementation of the risk assessment method based on machine learning provided by the embodiment of the present application in any of the above-mentioned forms of subject application can refer to the process of the risk assessment method based on machine learning applied by the terminal device described later.

[0076] Please refer to Figure 1 , Figure 1 A schematic diagram of the steps of the risk assessment method based on machine learning in some embodiments of the present application. It should be understood that although Figure 1 The execution order of some method steps is shown in the figure, but based on different design requirements of actual applications, the risk assessment method based on machine learning provided in the embodiment of the present application can certainly adopt an execution order different from the method steps shown in the figure. That is, Figure 1The order of the steps in the method shown does not constitute a limitation on the logical order of execution of the risk assessment method based on machine learning provided in the embodiment of the present application. Figure 1 Reasonable changes to the order of the method steps shown should be included in the scope of protection of the machine learning-based risk assessment method provided in the embodiments of the present application.

[0077] like Figure 1 As shown, in some embodiments, the risk assessment method based on machine learning provided in the embodiments of the present application may include but is not limited to steps S101 to S104.

[0078] Step S101, obtaining historical claims data and user financial data of the agricultural insurance to be evaluated, and obtaining growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated.

[0079] It should be noted that the agricultural insurance to be evaluated may include agricultural insurance purchased for agricultural crops or agricultural animals. Alternatively, the agricultural insurance to be evaluated may also include users who have purchased agricultural insurance for agricultural crops or agricultural animals. In addition, in the case where the agricultural insurance to be evaluated is agricultural insurance purchased for agricultural crops, and, in the case where the agricultural insurance to be evaluated is a user who purchases agricultural insurance for agricultural crops, the insured crops are agricultural crops, and in the case where the agricultural insurance to be evaluated is agricultural insurance purchased for agricultural animals, and, in the case where the agricultural insurance to be evaluated is a user who purchases agricultural insurance for agricultural animals, the insured crops are agricultural animals.

[0080] When the terminal device performs risk rating on the agricultural insurance to be evaluated, it first collects the relevant business data of the agricultural insurance to be evaluated, thereby obtaining the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated.

[0081] Step S102, generating agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data.

[0082] The terminal device cleans and preprocesses the acquired historical claims data of the agricultural insurance to be evaluated, the user financial data, the growth data of the insured crops of the agricultural insurance to be evaluated, the geographical data and the meteorological data, so as to generate the agricultural insurance characteristic data of the agricultural insurance to be evaluated.

[0083] If any of the historical claims data, user financial data, growth data, geographic data and meteorological data obtained by the terminal device is missing or abnormal, the terminal device will first process the missing values ​​and abnormal values ​​during data cleaning and preprocessing, and then perform necessary feature engineering on each data to generate agricultural insurance feature data for the agricultural insurance to be evaluated.

[0084] In some embodiments, the terminal device performs missing value processing on one or more of the acquired historical claims data, user financial data, growth data, geographic data, and meteorological data, including: the terminal device fills in items with missing characteristic values ​​in the historical claims data, user financial data, growth data, geographic data, and meteorological data, usually with the median of the item.

[0085] In some embodiments, the terminal device performs outlier processing on one or more of the acquired historical claims data, user financial data, growth data, geographic data, and meteorological data, including: the terminal device directly removes the feature of the item with obviously unreasonable values ​​in the historical claims data, user financial data, growth data, geographic data, and meteorological data.

[0086] Step S103, inputting the agricultural insurance characteristic data into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm.

[0087] Step S104, calculating the agricultural insurance characteristic data through the risk assessment model to obtain a risk level assessment result of the agricultural insurance to be assessed.

[0088] It should be noted that the preset agricultural insurance feature sample data set is the model training sample data generated by the terminal device after collecting relevant business data of agricultural insurance in advance, and then cleaning and preprocessing the historical claims data and user financial data of agricultural insurance in the relevant business data, as well as the growth data, geographical data and meteorological data of the insured crops of agricultural insurance.

[0089] The terminal device pre-trains the model of the agricultural insurance characteristic sample data set based on the machine learning algorithm to obtain a risk assessment model that can be used for agricultural insurance. In this way, after the terminal device generates the agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insurance target crops of the agricultural insurance to be evaluated, the terminal device further inputs the agricultural insurance characteristic data into the pre-trained risk assessment model, so as to calculate the agricultural insurance characteristic data through the risk assessment model. Afterwards, when the risk assessment model completes the calculation and outputs the model calculation result, the terminal device can use the model calculation result as the risk level assessment result for the intelligent risk rating of the agricultural insurance to be evaluated.

[0090] In some embodiments, when the terminal device performs model training on the agricultural insurance feature sample data set based on the machine learning algorithm, the output of the model training can be set as the risk level for risk assessment prediction of the agricultural insurance. For example, the terminal device specifically sets the output of the model training as three predicted risk levels of low risk, medium risk and high risk. In this way, after the terminal device inputs the agricultural insurance feature data of the agricultural insurance to be evaluated into the trained risk assessment model, the model calculation result output by the risk assessment model for the agricultural insurance feature data is low risk, medium risk or high risk, and the terminal device can directly use the low risk, medium risk or high risk output by the risk assessment model as the risk level assessment result of the agricultural insurance to be evaluated.

[0091] In an embodiment of the present application, when the terminal device performs risk rating for the agricultural insurance to be evaluated, it first collects the relevant business data of the agricultural insurance to be evaluated, thereby obtaining the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insurance target crops of the agricultural insurance to be evaluated. For the historical claims data, user financial data of the agricultural insurance to be evaluated, the growth data, geographical data and meteorological data of the insurance target crops of the agricultural insurance to be evaluated, the terminal device performs data cleaning and preprocessing respectively, thereby generating agricultural insurance characteristic data of the agricultural insurance to be evaluated. Afterwards, the terminal device further inputs the agricultural insurance characteristic data into the pre-trained risk assessment model, so as to calculate the agricultural insurance characteristic data through the risk assessment model. Finally, when the risk assessment model completes the calculation and outputs the model calculation result, the terminal device can use the model calculation result as the risk level assessment result for the intelligent risk rating of the agricultural insurance to be evaluated.

[0092] Compared with the method of conducting risk assessment of agricultural insurance through manual on-site investigation, the embodiment of the present application obtains a risk assessment model that can be used for agricultural insurance by training with a machine learning algorithm, so that it is only necessary to obtain the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, and then generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, user financial data, growth data, geographical data and meteorological data of the insured crops, and then input the agricultural insurance characteristic data into the risk assessment model, and then the agricultural insurance characteristic data can be calculated by the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be evaluated.

[0093] In this way, the embodiment of the present application uses machine learning algorithms to intelligently conduct risk assessments on agricultural insurance, which eliminates the need to rely on a large number of manual on-site inspections and capture real-time changes and complex factors in agricultural production, thereby effectively reducing the time and labor costs of risk assessments on agricultural insurance. In addition, the embodiment of the present application uses machine learning algorithms to conduct risk assessments on agricultural insurance, which can also improve the insurance company's intelligent operation level in agricultural insurance, thereby reducing the labor costs and claims risks of agricultural insurance business, and effectively contributing to the healthy development of agricultural insurance.

[0094] In some embodiments, the historical claims data acquired by the terminal device includes: historical claims time and historical claims amount.

[0095] Please refer to Figure 2 , Figure 2 for Figure 1 Schematic diagram of the detailed step flow of step S102. Figure 2 As shown, in the case where the historical claim data includes historical claim time and historical claim amount, the above-mentioned step S102 may include steps S201 to S202 as shown below.

[0096] Step S201 , numerically processing the growth data, the geographic data and the meteorological data to obtain feature codes corresponding to the growth data, the geographic data and the meteorological data respectively.

[0097] When the terminal device performs necessary feature engineering on the acquired historical claims data, user financial data, growth data, geographic data and meteorological data, it first performs numerical processing on the non-numerical growth data, geographic data and meteorological data to obtain the feature codes corresponding to the growth data, geographic data and meteorological data respectively.

[0098] For example, it is assumed that the growth data of the insurance subject crop obtained by the terminal device is the crop planting situation. In this way, the terminal device can encode the features according to the crop growth period to digitize the growth data, such as:

[0099] Growth cycle budding stage Seedling stage Growth period Maturity Feature encoding value 1 2 3 4

[0100] In addition, it is assumed that the geographical data of the insured crop obtained by the terminal device is the geographical location where the crop is planted. In this way, the terminal device can encode the geographical location according to the geographical area division to digitize the non-numerical geographical data, such as:

[0101] Location northeast North China Central China South China East China southwest northwest Feature Encoding 1 2 3 4 5 6 7

[0102] Furthermore, it is assumed that the meteorological data of the insured crops obtained by the terminal device is the climatic conditions of the area where the crops are planted. In this way, the terminal device can classify the meteorological disaster warning signals into different levels to digitally encode the meteorological values, such as:

[0103] Weather warning signal blue yellow orange color red Level coding 0 1 2 3

[0104] It should be noted that the terminal device can also reduce the complexity of the historical claim time feature by numerical processing for the historical claim time in the historical claim data obtained. For example, the terminal device can divide the historical claim time by month and define the month as 1-12.

[0105] In addition, since the historical claim amounts and user financial data in the historical claim data acquired by the terminal device are both numerical data, the terminal device does not need to encode the two data.

[0106] Step S202, respectively performing feature value normalization processing on the feature codes corresponding to the growth data, the geographic data and the meteorological data, and the historical claim time, the historical claim amount and the user financial data, to obtain the agricultural insurance feature data of the agricultural insurance to be evaluated.

[0107] After the terminal device digitizes the non-numerical growth data, geographic data and meteorological data, it further normalizes the characteristic values ​​corresponding to the numerical growth data, geographic data and meteorological data, and the historical claim time, historical claim amount and user financial data, which are also numerical, to obtain the agricultural insurance characteristic data of the agricultural insurance to be evaluated.

[0108] Exemplarily, after completing the feature coding of the growth data, geographic data, and meteorological data, the terminal device uses the following formula to normalize the feature values ​​of each of the feature codes corresponding to the growth data, geographic data, and meteorological data, and the historical claim time, historical claim amount, and user financial data:

[0109]

[0110] Among them, x ′ is the normalized eigenvalue, x max is the maximum value in a data item, x mix is the minimum value in the data, and x is the original eigenvalue in the data.

[0111] In this embodiment, for non-numerical growth data, geographic data and meteorological data with relatively complex data formats, the terminal device obtains the feature codes corresponding to the growth data, geographic data and meteorological data respectively by digitizing them, thereby reducing the complexity of these data features. In addition, the terminal device further normalizes the feature values ​​of the feature codes after digitization with other numerical historical claim times, historical claim amounts and user financial data, and then uses the normalized feature data for model calculation. In this way, the complexity of the subsequent risk assessment model in calculating the agricultural insurance feature data is effectively reduced, thereby improving the efficiency of intelligent risk assessment of agricultural insurance based on the risk assessment model.

[0112] The terminal device can build a unified reference standard for risk assessment of agricultural insurance, and then collect corresponding data according to the unified reference standard for subsequent intelligent risk level assessment based on the risk assessment model.

[0113] Please refer to Figure 3 , Figure 3 for Figure 1 Schematic diagram of the detailed process flow of step S101.

[0114] like Figure 3 As shown, in some embodiments, the above-mentioned step S101, obtaining the historical claims data and user financial data of the agricultural insurance to be evaluated, and obtaining the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, may include steps S301 to S302 as shown below.

[0115] Step S301, obtaining risk assessment indicators from a preset agricultural insurance assessment indicator system; the agricultural insurance assessment indicator system represents a reference standard for risk assessment of agricultural insurance.

[0116] After the terminal device has pre-built the agricultural insurance assessment index system as a unified reference standard for risk assessment of agricultural insurance, when obtaining relevant business data of the agricultural insurance to be assessed, it can collect targeted data according to the reference standard. For example, the terminal device can obtain the risk assessment index from the agricultural insurance assessment index system. The risk assessment index is the characteristic of the reference standard used to characterize the risk assessment of agricultural insurance in the agricultural insurance assessment index system. In this way, the terminal device collects the relevant data of the agricultural insurance to be assessed under the risk assessment index, and can obtain the historical claims data and user financial data of the agricultural insurance to be assessed, as well as the growth data, geographic data and meteorological data of the insured crops of the agricultural insurance to be assessed.

[0117] Step S302, collecting historical claims data and user financial data of the agricultural insurance to be evaluated based on the risk assessment indicators, and collecting growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated based on the risk assessment indicators.

[0118] After obtaining the agricultural insurance assessment index, the terminal device collects business data related to the agricultural insurance to be assessed based on the risk assessment index, thereby collecting historical claims data and user financial data of the agricultural insurance to be assessed based on the guidance of the risk assessment index, and collecting growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be assessed based on the guidance of the risk assessment index.

[0119] It should be noted that the risk assessment indicators in the agricultural insurance assessment indicator system include one or more. In some embodiments, when there is one risk assessment indicator, the one risk assessment indicator can simultaneously characterize multiple features of the agricultural insurance business, so that the terminal device can collect the aforementioned multiple feature data according to the risk assessment indicator. In addition, in other embodiments, when there are multiple risk assessment indicators, each of the multiple risk assessment indicators can simultaneously characterize the characteristics of at least one agricultural insurance business, so that the terminal device can also collect the aforementioned multiple feature data according to the multiple risk assessment indicators at the same time.

[0120] Exemplarily, when the terminal device obtains multiple risk assessment indicators, it first directly collects the historical claim time and historical claim amount of the agricultural insurance to be evaluated as historical claim data based on indicator 1 that characterizes the claim characteristics of agricultural insurance among the multiple risk assessment indicators, and collects the annual income level of the user of the agricultural insurance to be evaluated as user financial data based on indicator 2 that characterizes the user characteristics of agricultural insurance among the multiple risk assessment indicators. In addition, the terminal device also collects the planting conditions, geographical location, and meteorological disaster warnings of the agricultural insurance to be evaluated in turn based on indicator 3 that characterizes the characteristics of the subject matter of agricultural insurance among the multiple risk assessment indicators, and then uses the collected planting conditions as the growth data of the crop of the insurance subject of the agricultural insurance to be evaluated, uses the collected geographical location as the geographical data of the crop of the insurance subject of the agricultural insurance to be evaluated, and uses the collected meteorological disaster warnings as the meteorological data of the crop of the insurance subject of the agricultural insurance to be evaluated.

[0121] The terminal device can construct a unified agricultural insurance assessment indicator system by performing feature analysis on agricultural insurance-related business data.

[0122] Please refer to Figure 4 , Figure 4 A schematic flowchart of the steps of the machine learning-based risk assessment method provided in the embodiment of the present application in other embodiments.

[0123] like Figure 4 As shown, in some embodiments, in the aforementioned step S301, before obtaining the risk assessment index from the preset agricultural insurance assessment index system, the risk assessment method based on machine learning provided by the application embodiment may also include steps S401 to S403 as shown below.

[0124] Step S401, obtaining historical agricultural insurance business data.

[0125] When constructing a unified agricultural insurance assessment index system, the terminal device first obtains all agricultural insurance historical business data recorded by the business system from the system.

[0126] It should be noted that the business system can be a data platform used by insurance companies for agricultural insurance pricing, sales, contracting and claims, and the business system can record the relevant business data of agricultural insurance after each operation of agricultural insurance pricing, sales, contracting and claims. In this way, the terminal device can flexibly obtain the relevant business data of any agricultural insurance as agricultural insurance historical business data by connecting with the business system.

[0127] Step S402, based on the business characteristics analysis of the historical business data of agricultural insurance, multiple consideration factors for risk assessment of agricultural insurance are obtained.

[0128] After acquiring the historical business data of agricultural insurance, the terminal device further performs business characteristic analysis on the historical business data of agricultural insurance to obtain multiple considerations for risk assessment of agricultural insurance.

[0129] In some embodiments, the terminal device can extract information from historical agricultural insurance business data through natural language processing, thereby obtaining multiple considerations for risk assessment of agricultural insurance.

[0130] For example, agricultural insurance has a series of special characteristics, and these characteristics are particularly important when evaluating and analyzing agricultural insurance. In addition, the type, cycle, frequency and intensity of agricultural risks are difficult to determine, which makes it difficult for agricultural insurance operators to effectively disperse similar risks when facing disasters. The larger the insured amount, the greater the risk, and the higher the compensation rate. Therefore, by performing feature analysis on historical agricultural insurance business data, the terminal device can conclude that at least the following factors need to be considered when rating agricultural insurance risks:

[0131] (1) Market value of the subject matter of insurance and extent of loss:

[0132] Assess the market value of different types of crops or farmed animals at different stages of growth, as well as the extent of losses they may suffer from possible natural disasters.

[0133] (2) Frequency and intensity of natural disasters:

[0134] Analyze historical data on natural disasters in different regions, predict the types and intensity of disasters that may occur in the future, and formulate reasonable insurance premiums.

[0135] (3) Regional differences:

[0136] Differentiated insurance products and rates are formulated based on the climate, soil, crop types and other conditions in different regions to meet the needs of local farmers.

[0137] (4) Seasonal and cyclical factors:

[0138] Taking into account the seasonal characteristics of agricultural production, we will rationally arrange business processes such as insurance development, underwriting, and claims settlement to ensure the timeliness and effectiveness of insurance services.

[0139] (5) Claims ratio and risk diversification capability:

[0140] Analyze agricultural insurance claims ratio data, evaluate insurance companies’ risk tolerance and diversification capabilities, and develop reasonable reinsurance strategies.

[0141] Step S403: construct an agricultural insurance assessment index system based on the multiple consideration factors.

[0142] It should be noted that one of the multiple considerations corresponds to at least one risk assessment indicator in the constructed agricultural insurance assessment indicator system.

[0143] After analyzing the above-mentioned multiple considerations, the terminal device can construct a unified agricultural insurance assessment indicator system by summarizing each consideration factor into at least one quantifiable characteristic variable and then using the characteristic variable as a unified reference risk assessment indicator when assessing the risk level of agricultural insurance.

[0144] For example, the terminal device analyzes the above five indicators in combination with business data to summarize the following quantifiable characteristic variables: historical compensation time, crop planting conditions, geographical location, climate conditions, annual income level, and compensation amount. Afterwards, the terminal device can construct a unified agricultural insurance assessment indicator system by taking each characteristic variable as a risk assessment indicator.

[0145] In this embodiment, the terminal device pre-constructs an agricultural insurance assessment index system as a unified reference standard for risk assessment of agricultural insurance, and then when obtaining relevant business data of the agricultural insurance to be assessed, the terminal device can obtain the risk assessment index from the agricultural insurance assessment index system. The risk assessment index is the characteristic of the reference standard for risk assessment of agricultural insurance in the agricultural insurance assessment index system. In this way, the terminal device collects the relevant data of the agricultural insurance to be assessed under the risk assessment index, and can obtain the historical claims data and user financial data of the agricultural insurance to be assessed, as well as the growth data, geographical data and meteorological data of the insurance target crops of the agricultural insurance to be assessed. In this way, this embodiment can not only efficiently collect various characteristic data of the agricultural insurance to be assessed for subsequent intelligent risk level assessment, but also overcome the problem of inaccurate assessment results caused by the difficulty in fully capturing the real-time changes and complex factors in agricultural production, that is, this embodiment improves the efficiency of risk rating for agricultural insurance to be assessed, and also improves the accuracy of risk rating for agricultural insurance to be assessed.

[0146] The terminal device can use the extreme gradient boosting (EXtreme Gradient Boosting, XGBoost) algorithm in the machine learning algorithm for model training. That is, by setting appropriate parameters of the XGBoost algorithm (such as learning rate, number of trees, depth of trees, etc.), and then using the pre-constructed agricultural insurance feature sample data set as the model input, and setting the output as the predicted risk level (such as low risk, medium risk, and high risk), the risk assessment model for assessing the risk level of agricultural insurance can be trained. In addition, when training the model, the terminal device can also consider the temporal and regional characteristics of the agricultural insurance feature data, and use a sliding window or regional method to train the model input to improve the generalization ability of the risk assessment model finally trained.

[0147] Please refer to Figure 5 , Figure 5 A flow chart of a model training step involved in a risk assessment method based on machine learning provided in an embodiment of the present application.

[0148] like Figure 5 As shown, in some embodiments, the risk assessment method based on machine learning provided in the embodiments of the present application may execute steps S501 to S504 as shown below when performing model training.

[0149] Step S501, obtaining an agricultural insurance feature sample data set.

[0150] When the terminal device performs model training based on the machine learning algorithm, it obtains the agricultural insurance feature sample data set used for model training.

[0151] In some embodiments, the terminal device may perform data cleaning and preprocessing on the aforementioned agricultural insurance historical business data, thereby generating an agricultural insurance feature sample data set for model training.

[0152] It should be noted that the process in which the terminal device performs data and preprocessing on the historical business data of agricultural insurance to generate an agricultural insurance characteristic sample data set can refer to the aforementioned terminal device based on the historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, to generate the agricultural insurance characteristic data of the agricultural insurance to be evaluated. The same content will not be repeated here.

[0153] In some embodiments, the terminal device may set the learning rate, number of trees, tree depth and other parameters of the machine learning algorithm (such as the XGBoost algorithm) before acquiring the agricultural insurance feature sample data set. Alternatively, the terminal device may first acquire the agricultural insurance feature sample data set and then set the parameters of the machine learning algorithm.

[0154] Step S502, dividing the agricultural insurance characteristic sample data set into a time series to obtain a plurality of sub-window sample data; the time series of two adjacent sub-window sample data in the plurality of sub-window sample data are different.

[0155] Step S503: using the plurality of sub-window sample data as a model input of a preset machine learning algorithm.

[0156] It should be noted that the preset machine learning algorithm can be an XGBoost algorithm. It should be understood that based on different design requirements of actual applications, in other feasible implementations, the terminal device can of course also use other machine learning algorithms other than the XGBoost algorithm for model training. The machine learning-based risk assessment method provided in the embodiment of the present application is not limited to the specific type of machine learning algorithm.

[0157] After the terminal device obtains the agricultural insurance characteristic sample data set, based on the consideration of the time series of the agricultural insurance characteristic sample data, the terminal device further divides the agricultural insurance characteristic sample data set into time series, thereby obtaining a plurality of sub-window sample data. Afterwards, the terminal device uses the plurality of sub-window sample data as the model input of the machine learning algorithm.

[0158] It should be noted that, among the multiple sub-window sample data obtained by dividing the agricultural insurance feature sample data by the terminal device, the time series of any two adjacent sub-window sample data are different.

[0159] Step S504: Perform sliding window model training on the plurality of sub-window sample data using the machine learning algorithm to obtain a risk assessment model.

[0160] After the terminal device inputs multiple sub-window sample data into the machine learning algorithm and the parameters of the machine learning algorithm are set, the terminal device can perform sliding window model training on the multiple sub-window sample data based on the set parameters, so that after the model training converges, a risk assessment model for risk level assessment of agricultural insurance is obtained.

[0161] Please refer to Figure 6 , Figure 6 A flowchart of another model training step involved in the machine learning-based risk assessment method provided in an embodiment of the present application.

[0162] like Figure 6 As shown, in some embodiments, after obtaining the agricultural insurance feature sample data set in the above-mentioned step S501, the risk assessment method based on machine learning provided in the embodiment of the present application can also perform model training by executing steps S601 to S603 as shown below to obtain a risk assessment model.

[0163] Step S601, dividing the agricultural insurance characteristic sample data set into regions to obtain a plurality of sub-region sample data; the plurality of sub-region sample data each contain a different geographical location.

[0164] Step S602: Using the plurality of sub-region sample data as a model input of a preset machine learning algorithm.

[0165] After the terminal device obtains the agricultural insurance characteristic sample data set, based on the regionality of the agricultural insurance characteristic sample data, the terminal device further divides the agricultural insurance characteristic sample data set into regions, thereby obtaining multiple sub-region sample data. Afterwards, the terminal device uses the multiple sub-region sample data as model inputs of the machine learning algorithm.

[0166] It should be noted that among the multiple sub-region sample data obtained by dividing the agricultural insurance characteristic sample data by the terminal device, the geographical locations contained in each of the multiple sub-region sample data are different. For example, when the terminal device divides the agricultural insurance characteristic sample data set into regions, it can be divided into one sub-region sample data, the agricultural insurance characteristic sample data with the geographical location of Northeast China is divided into another sub-region sample data, and the agricultural insurance characteristic sample data with the geographical location of North China is divided into another sub-region sample data, and the agricultural insurance characteristic sample data with the geographical location of North China is divided into another sub-region sample data, and the agricultural insurance characteristic sample data with the geographical location of Northwest China is divided into another sub-region sample data, so that it can be ensured that the geographical locations contained in each sub-region sample data are different.

[0167] Step S603: Performing regional model training on the plurality of sub-region sample data by using the machine learning algorithm to obtain a risk assessment model.

[0168] After the terminal device inputs multiple sub-region sample data into the machine learning algorithm and the parameters of the machine learning algorithm are set, the terminal device can perform regional model training on the multiple sub-region sample data based on the set parameters, so that after the model training converges, a risk assessment model for risk level assessment of agricultural insurance can be obtained.

[0169] In some embodiments, after performing sliding window model training on multiple sub-window sample data to obtain a risk assessment model, the terminal device may further input multiple sub-region sample data into the risk assessment model, thereby further performing sub-region model training on the multiple sub-region sample data through the risk assessment model to obtain the final risk assessment model.

[0170] In addition, in some embodiments, for different sub-region sample data, the terminal device can use a machine learning algorithm to simultaneously process these sub-region sample data, thereby obtaining a risk assessment model suitable for assessing the risk level of agricultural insurance in different geographical locations.

[0171] In some embodiments, for different sub-region sample data, the terminal device may also use multiple machine learning algorithms to process the sub-region sample data respectively, thereby obtaining multiple risk assessment models. Among them, each of the multiple risk assessment models is suitable for assessing the risk level of agricultural insurance in a geographical location.

[0172] In some embodiments, for different sub-region sample data, the terminal device can also use multiple machine learning algorithms to respectively process these sub-region sample data to obtain multiple risk assessment models, and then further perform feature fusion on the multiple risk assessment models to obtain a risk assessment model suitable for risk level assessment of agricultural insurance in different geographical locations.

[0173] Based on this, in the case where the machine learning algorithm includes a first machine learning algorithm and a second machine learning algorithm, the above step S603, performing regional model training on the plurality of sub-region sample data by the machine learning algorithm to obtain a risk assessment model, may include:

[0174] Performing model training on first sample data among the plurality of sub-region sample data by using the first machine learning algorithm to obtain a first evaluation model; the first sample data is sub-region sample data including the first geographical location among the plurality of sub-region sample data;

[0175] Performing model training on second sample data among the multiple sub-region sample data by the second machine learning algorithm to obtain a second evaluation model; the second sample data is sub-region sample data including a second geographical location among the multiple sub-region sample data, and the second geographical location is different from the first geographical location;

[0176] The first assessment model and the second assessment model are integrated to obtain a risk assessment model.

[0177] When the terminal device performs regional model training on multiple sub-region sample data based on the machine learning algorithm, in the case where the machine learning algorithm includes a first machine learning algorithm and a second machine learning algorithm, the terminal device can set the same parameters for the first machine learning algorithm and the second machine learning algorithm, and then use the first machine learning algorithm to perform model training on the first sample data (the sub-region sample data of the first geographical location included in the multiple sub-region sample data) among the multiple sub-region sample data, so as to obtain a first evaluation model when the model training converges. In addition, the terminal device uses the second machine learning algorithm to perform model training on the second sample data (the sub-region sample data of the second geographical location included in the multiple sub-region sample data, and the second geographical location is different from the first geographical location) among the multiple sub-region sample data, so as to obtain a second evaluation model when the model training converges. Thereafter, the terminal device further performs feature fusion processing on the first evaluation model and the second evaluation model, so as to fuse the first evaluation model and the second evaluation model into a risk assessment model.

[0178] In some embodiments, after the terminal device integrates the first assessment model and the second assessment model to form a risk assessment model, it can also verify the risk assessment model through an agricultural insurance feature sample data set, so that when the risk assessment model passes the verification, it is determined that the model training is completed and finally a risk assessment model for risk level assessment of agricultural insurance is obtained.

[0179] In some embodiments, after the risk assessment model is trained, the terminal device can deploy the trained risk assessment model to the production environment and connect it to the business system. In addition, the terminal device can also set a monitoring mechanism for the risk assessment model to monitor the operation of the model in real time to ensure the stability and accuracy of the intelligent risk level assessment of the agricultural insurance to be assessed based on the risk assessment model. For example, the terminal device can regularly update and maintain the risk assessment model to adapt to changes in data and business needs, thereby ensuring the stability and accuracy of the risk assessment model in assessing the risk level of the agricultural insurance to be assessed.

[0180] In some embodiments, after the terminal device uses the risk assessment model to assess the risk level of the agricultural insurance to be assessed, thereby obtaining the risk level assessment result of the agricultural insurance to be assessed, the terminal device can also provide a clear risk rating report and explanation based on the risk level assessment result, thereby helping agricultural insurance business personnel understand the rating result and the logic behind it. In addition, the terminal device can also improve the risk management level of the agricultural insurance to be assessed by applying the risk level assessment result to various links such as pricing, underwriting, and claims settlement of agricultural insurance products.

[0181] In this embodiment, a risk assessment model that can be used for agricultural insurance is obtained by training a terminal device using a machine learning algorithm (such as the Xgboost algorithm, etc.), and the result of the risk assessment model for agricultural insurance is used to provide a risk assessment report for agricultural insurance product pricing, sales, and claims, thereby filling the gap in the traditional agricultural insurance field where there is no intelligent and unified risk assessment method. In addition, this embodiment uses a machine learning algorithm to train a risk assessment model to rate the risk of agricultural insurance, which also improves the intelligent operation level of property insurance companies in agricultural insurance, reduces labor costs and claims risks, and can effectively contribute to the healthy development of agricultural insurance.

[0182] See also Figure 7 The embodiment of the present application also provides a risk assessment device based on machine learning, which can implement the above-mentioned risk assessment method based on machine learning. The device includes: an acquisition module 701, a data processing module 702, a model input module 703 and a model calculation module 704, wherein:

[0183] The acquisition module 701 is used to acquire the historical claims data and user financial data of the agricultural insurance to be evaluated, and to acquire the growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated;

[0184] A data processing module 702, configured to generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data;

[0185] The model input module 703 is used to input the agricultural insurance characteristic data into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm;

[0186] The model calculation module 704 is used to calculate the agricultural insurance characteristic data through the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be assessed.

[0187] In some embodiments, the acquisition module is also used to obtain risk assessment indicators from a preset agricultural insurance assessment index system; the agricultural insurance assessment index system represents a reference standard for risk assessment of agricultural insurance; based on the risk assessment indicators, historical claims data and user financial data of the agricultural insurance to be assessed are collected, and based on the risk assessment indicators, growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be assessed are collected.

[0188] In some embodiments, the risk assessment device based on machine learning provided in the embodiments of the present application further includes:

[0189] A system construction module is used to obtain historical business data of agricultural insurance; based on business feature analysis of the historical business data of agricultural insurance, multiple considerations are obtained when conducting risk assessment on agricultural insurance; and, based on the multiple considerations, an agricultural insurance assessment index system is constructed; wherein one of the multiple considerations corresponds to at least one risk assessment index in the agricultural insurance assessment index system.

[0190] In some embodiments, the historical claims data includes: historical claims time and historical claims amount; the data processing module 702 is also used to perform numerical processing on the growth data, the geographic data and the meteorological data to obtain the feature codes corresponding to the growth data, the geographic data and the meteorological data respectively; and, perform feature value normalization processing on the feature codes corresponding to the growth data, the geographic data and the meteorological data respectively, and the historical claims time, the historical claims amount and the user financial data, to obtain the agricultural insurance feature data of the agricultural insurance to be evaluated.

[0191] In some embodiments, the risk assessment device based on machine learning provided in the embodiments of the present application further includes:

[0192] Model training module, used to obtain agricultural insurance feature sample data set;

[0193] The agricultural insurance characteristic sample data set is divided into multiple sub-window sample data by time series; the time series of two adjacent sub-window sample data in the multiple sub-window sample data are different;

[0194] Using the plurality of sub-window sample data as a model input of a preset machine learning algorithm;

[0195] The sliding window model is trained on the multiple sub-window sample data through the machine learning algorithm to obtain a risk assessment model.

[0196] In some embodiments, the model training module is also used to divide the agricultural insurance characteristic sample data set into regions to obtain multiple sub-region sample data; the multiple sub-region sample data each contain a different geographical location; the multiple sub-region sample data are used as a model input of a preset machine learning algorithm; and, through the machine learning algorithm, the multiple sub-region sample data are subjected to regional model training to obtain a risk assessment model.

[0197] In some embodiments, the machine learning algorithm includes a first machine learning algorithm and a second machine learning algorithm; the model training module is also used to perform model training on the first sample data among the multiple sub-region sample data through the first machine learning algorithm to obtain a first evaluation model; the first sample data is the sub-region sample data including the first geographic location among the multiple sub-region sample data; the second sample data is the sub-region sample data including the second geographic location among the multiple sub-region sample data, and the second geographic location is different from the first geographic location; and the first evaluation model and the second evaluation model are merged to obtain a risk assessment model.

[0198] It should be noted that the specific implementation of the risk assessment device based on machine learning provided in the embodiment of the present application is basically the same as the specific implementation of the risk assessment method based on machine learning provided in the above-mentioned embodiment of the present application, and will not be repeated here.

[0199] The embodiment of the present application also provides a computer device, the computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned risk assessment method based on machine learning when executing the computer program. The computer device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0200] See also Figure 8 , Figure 8 The hardware structure of a computer device according to another embodiment is shown, and the computer device includes:

[0201] The processor 801 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0202] The memory 802 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 802 can store an operating system and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 802, and the processor 801 calls and executes the risk assessment method based on machine learning in the embodiment of this application;

[0203] Input / output interface 803, used to implement information input and output;

[0204] The communication interface 804 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0205] A bus 805 that transmits information between the various components of the device (e.g., the processor 801, the memory 802, the input / output interface 803, and the communication interface 804);

[0206] The processor 801 , the memory 802 , the input / output interface 803 and the communication interface 804 are connected to each other in communication within the device via a bus 805 .

[0207] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program, and when the computer program is executed by a processor, the above-mentioned risk assessment method based on machine learning is implemented.

[0208] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0209] An embodiment of the present application also provides a computer program product, which stores a computer program. When a processor executes the computer program, the above-mentioned risk assessment method based on machine learning is implemented.

[0210] The machine learning-based risk assessment method, machine learning-based risk assessment device, computer equipment, computer-readable storage medium, and computer program product provided in the embodiments of the present application obtain a risk assessment model that can be used for agricultural insurance by training with a machine learning algorithm, thereby only needing to obtain the historical claims data and user financial data of the agricultural insurance to be assessed, as well as the growth data, geographic data, and meteorological data of the insured crops of the agricultural insurance to be assessed. Then, based on the historical claims data, user financial data, and the growth data, geographic data, and meteorological data of the insured crops, agricultural insurance characteristic data of the agricultural insurance to be assessed is generated. Then, the agricultural insurance characteristic data is input into the risk assessment model, and the agricultural insurance characteristic data can be calculated through the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be assessed.

[0211] In this way, the embodiment of the present application uses machine learning algorithms to intelligently conduct risk assessments on agricultural insurance, which eliminates the need to rely on a large number of manual on-site inspections and capture real-time changes and complex factors in agricultural production, thereby effectively reducing the time and labor costs of risk assessments on agricultural insurance. In addition, the embodiment of the present application uses machine learning algorithms to conduct risk assessments on agricultural insurance, which can also improve the insurance company's intelligent operation level in agricultural insurance, thereby reducing the labor costs and claims risks of agricultural insurance business, and effectively contributing to the healthy development of agricultural insurance.

[0212] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0213] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0214] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0215] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0216] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0217] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0218] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0219] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0220] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0221] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0222] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A risk assessment method based on machine learning, characterized in that: The method comprises: Acquire historical claims data and user financial data of the agricultural insurance to be evaluated, and acquire growth data, geographic data, and meteorological data of the insured crops of the agricultural insurance to be evaluated; Generate agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data; The agricultural insurance characteristic data is input into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm; The agricultural insurance characteristic data is calculated by the risk assessment model to obtain a risk level assessment result of the agricultural insurance to be assessed.

2. The method according to claim 1, characterized in that The obtaining of historical claims data and user financial data of the agricultural insurance to be evaluated, as well as the obtaining of growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be evaluated, includes: Obtaining risk assessment indicators from a preset agricultural insurance assessment indicator system; the agricultural insurance assessment indicator system represents a reference standard for risk assessment of agricultural insurance; Based on the risk assessment indicators, historical claims data and user financial data of the agricultural insurance to be assessed are collected; and based on the risk assessment indicators, growth data, geographical data and meteorological data of the insured crops of the agricultural insurance to be assessed are collected.

3. The method according to claim 2, characterized in that Before obtaining the risk assessment index from the preset agricultural risk assessment index system, the method further includes: Obtain historical agricultural insurance business data; Based on the business characteristics analysis of the historical business data of agricultural insurance, multiple consideration factors for risk assessment of agricultural insurance are obtained; Based on the above-mentioned multiple considerations, an agricultural insurance assessment index system is constructed; Among them, one of the multiple considerations corresponds to at least one risk assessment indicator in the agricultural risk assessment indicator system.

4. The method according to claim 1, characterized in that: The historical claims data include: historical claims time and historical claims amount; The generating of the agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data comprises: Numerical processing is performed on the growth data, the geographic data and the meteorological data to obtain feature codes corresponding to the growth data, the geographic data and the meteorological data respectively; The characteristic codes corresponding to the growth data, the geographic data and the meteorological data, and the historical claim settlement time, the historical claim settlement amount and the user financial data are respectively normalized to obtain the agricultural insurance characteristic data of the agricultural insurance to be evaluated.

5. The method according to any one of claims 1 to 4, characterized in that The method further comprises: Obtain agricultural insurance feature sample dataset; The agricultural insurance characteristic sample data set is divided into multiple sub-window sample data by time series; the time series of two adjacent sub-window sample data in the multiple sub-window sample data are different; Using the plurality of sub-window sample data as a model input of a preset machine learning algorithm; The sliding window model is trained on the multiple sub-window sample data through the machine learning algorithm to obtain a risk assessment model.

6. The method according to claim 5, characterized in that After obtaining the agricultural insurance feature sample data set, the method further includes: Dividing the agricultural insurance characteristic sample data set by region to obtain a plurality of sub-region sample data; the plurality of sub-region sample data each contain a different geographical location; Using the plurality of sub-region sample data as a model input of a preset machine learning algorithm; The machine learning algorithm is used to perform regional model training on the multiple sub-region sample data to obtain a risk assessment model.

7. The method according to claim 6, characterized in that The machine learning algorithm includes a first machine learning algorithm and a second machine learning algorithm; The method of performing regional model training on the plurality of sub-region sample data by using the machine learning algorithm to obtain a risk assessment model includes: Performing model training on first sample data among the plurality of sub-region sample data by using the first machine learning algorithm to obtain a first evaluation model; the first sample data is sub-region sample data including the first geographical location among the plurality of sub-region sample data; Performing model training on second sample data among the multiple sub-region sample data by the second machine learning algorithm to obtain a second evaluation model; the second sample data is sub-region sample data including a second geographical location among the multiple sub-region sample data, and the second geographical location is different from the first geographical location; The first assessment model and the second assessment model are integrated to obtain a risk assessment model.

8. A risk assessment device based on machine learning, characterized in that: The device comprises: An acquisition module, used to acquire historical claims data and user financial data of the agricultural insurance to be evaluated, and to acquire growth data, geographic data and meteorological data of the insured crops of the agricultural insurance to be evaluated; A data processing module, used for generating agricultural insurance characteristic data of the agricultural insurance to be evaluated based on the historical claims data, the user financial data, the growth data, the geographic data and the meteorological data; A model input module, used to input the agricultural insurance characteristic data into a pre-trained risk assessment model; the risk assessment model is obtained by training a preset agricultural insurance characteristic sample data set based on a machine learning algorithm; The model calculation module is used to calculate the agricultural insurance characteristic data through the risk assessment model to obtain the risk level assessment result of the agricultural insurance to be assessed.

9. A computer device, characterized in that: The computer device includes a memory and a processor, the memory stores a computer program, and the processor implements the risk assessment method based on machine learning as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the risk assessment method based on machine learning described in any one of claims 1 to 7 is implemented.

Images