Resource allocation method and device, electronic equipment and storage medium

Abstract

The invention provides a resource allocation method and device, electronic equipment and a storage medium, relates to the technical field of artificial intelligence, and is suitable for the field of financial science and technology. The method comprises the steps of performing feature extraction on multi-source heterogeneous data corresponding to a target insurance product type to obtain multi-field features; splicing the disease event impact factor features with the multi-field features to obtain comprehensive impact features; carrying out cost prediction on the comprehensive impact characteristics to obtain a medical cost prediction data sequence; performing matrix construction according to the medical expense prediction data sequence to obtain a target probability matrix; carrying out weight calculation on the candidate resource configuration strategy according to the target probability matrix to obtain a strategy weight; screening the candidate resource configuration strategies according to the strategy weight to obtain a target resource configuration strategy; and performing resource allocation according to the target resource allocation strategy. The resource allocation strategy can be individually determined for different insurance product types, the resource allocation accuracy is improved, and the applicability is high.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and is applicable to the field of financial technology, particularly to a resource allocation method and apparatus, electronic device, and storage medium. Background Technology

[0002] Resource allocation refers to the process of rationally allocating various resources (such as funds, manpower, equipment, time, etc.) under certain goals, strategies, or constraints to achieve expected results or specific objectives. In the insurance scenario within the fintech field, financial companies can formulate resource allocation strategies (such as investment cycles, risk limits, etc.) based on data such as medical expenses, thereby achieving resource allocation.

[0003] In related technologies, resource allocation strategies are often determined based on the total medical costs predicted at a certain point in time. This makes it difficult to reflect the changes in medical costs for different types of insurance products (such as life insurance and health insurance) over different time periods in the future, and the resulting resource allocation strategies cannot be used effectively.

[0004] Therefore, the relevant technologies suffer from poor accuracy in resource allocation. Summary of the Invention

[0005] The main objective of this application is to propose a resource allocation method and apparatus, electronic device, and storage medium that can determine resource allocation strategies for different types of insurance products in a personalized manner, thereby improving the accuracy of resource allocation and having high applicability.

[0006] To achieve the above objectives, a first aspect of this application proposes a resource allocation method, the method comprising:

[0007] Multi-source data collection is performed on the target insurance product type to obtain multi-source heterogeneous data; wherein, the multi-source heterogeneous data includes medical data, insurance data and macroeconomic data;

[0008] Feature extraction is performed on the multi-source heterogeneous data to obtain multi-domain features;

[0009] Obtain the impact factor features of disease events, and concatenate the impact factor features of disease events with the multi-domain features to obtain comprehensive impact features;

[0010] Cost prediction is performed on the comprehensive impact characteristics to obtain a medical cost prediction data sequence;

[0011] A target probability matrix is ​​obtained by constructing a matrix based on the medical cost prediction data sequence; wherein, the target probability matrix is ​​used to represent the probability distribution of resource gaps over multiple consecutive time periods;

[0012] The preset candidate resource allocation strategies are weighted according to the target probability matrix to obtain the strategy weights;

[0013] The candidate resource allocation strategies are filtered according to the strategy weights to obtain the target resource allocation strategy;

[0014] Resources are allocated to the target insurance product type according to the target resource allocation strategy.

[0015] Optionally, the step of performing cost prediction on the comprehensive impact characteristics to obtain a medical cost prediction data sequence includes:

[0016] Obtain the types of diseases for which the target insurance product type supports compensation, thus obtaining the target supported diseases;

[0017] Construct a medical cost prediction model for the target supported diseases;

[0018] The medical cost prediction model for the target disease is used to predict the comprehensive impact characteristics by disease, thereby obtaining a subsequence of medical cost prediction data for the target supported disease; wherein, the subsequence of medical cost prediction data includes medical cost prediction data for the target supported disease in at least two time periods;

[0019] The medical cost prediction data subsequences of each of the target supported diseases are integrated to obtain the medical cost prediction data sequence.

[0020] Optionally, constructing a medical cost prediction model for the target supported disease includes:

[0021] Obtain historical medical expense data for the target supported disease; wherein, the historical medical expense data includes the number of cases, treatment costs, and reimbursement costs for the sample during the historical time period;

[0022] Feature extraction is performed on the number of patients in the sample and the treatment cost of the sample to obtain sample cost features;

[0023] Obtain the impact factor of sample disease events within the historical time period, and concatenate the sample cost feature and the sample disease event impact factor to obtain the sample impact feature;

[0024] The sample impact characteristics are predicted using a preset initial cost prediction model to obtain the predicted compensation cost.

[0025] Loss calculations are performed based on the predicted compensation costs and the sample compensation costs to obtain target loss data;

[0026] The parameters of the initial cost prediction model are adjusted based on the target loss data to obtain the medical cost prediction model for the target disease.

[0027] Optionally, the step of constructing a matrix based on the medical cost prediction data sequence to obtain the target probability matrix includes:

[0028] The medical expense prediction data sequence is organized by time period to obtain the predicted reimbursement amount for each time period;

[0029] Based on the predicted compensation amount, a matrix transformation is performed to construct a resource flow stress test matrix; wherein, each element in the resource flow stress test matrix is ​​used to represent the resource gap corresponding to a time period;

[0030] Probability analysis is performed on each element in the resource flow stress test matrix to obtain the target probability matrix.

[0031] Optionally, after allocating resources for the target insurance product type according to the target resource allocation strategy, the method further includes:

[0032] Get the total remaining resources;

[0033] For each of the target insurance product types, resource repayment estimates are performed to obtain the estimated resource repayment amount;

[0034] Based on the total remaining resources and the estimated resource repayment amount, a risk prediction is made to obtain the resource allocation risk probability.

[0035] The target resource allocation strategy is updated based on the resource allocation risk probability.

[0036] Optionally, updating the target resource allocation strategy based on the resource allocation risk probability includes:

[0037] If the resource allocation risk probability meets the predetermined risk conditions, then the flow type of the target resource allocation strategy is determined to obtain the strategy flow type; wherein, the strategy flow type is used to indicate whether the liquidity of the target resource allocation strategy is high or low;

[0038] If the policy flow type used to indicate that the liquidity of the target resource allocation policy is high, then the target resource allocation policy is deleted.

[0039] In response to the strategy flow type indicating low liquidity of the target resource allocation strategy, an auxiliary resource allocation strategy is selected from the candidate resource allocation strategies, and the target resource allocation strategy is replaced by the auxiliary resource allocation strategy; wherein, the strategy flow type of the auxiliary resource allocation strategy indicates high liquidity.

[0040] Optionally, the acquisition of disease event impact factor characteristics includes:

[0041] Obtain the types of diseases for which the target insurance product type supports compensation, thus obtaining the target supported diseases;

[0042] To develop target disease transmission models for target support diseases;

[0043] The historical number of cases is obtained, and the target disease transmission model is used to predict the number of cases based on the historical number of cases to obtain the target predicted number of cases.

[0044] Based on the predicted number of cases, the features of the disease event impact factor are represented to obtain the features.

[0045] To achieve the above objectives, a second aspect of this application provides a resource allocation apparatus, the apparatus comprising:

[0046] The data acquisition module is used to collect multi-source data on the target insurance product type to obtain multi-source heterogeneous data; wherein, the multi-source heterogeneous data includes medical data, insurance data and macroeconomic data;

[0047] The feature extraction module is used to extract features from the multi-source heterogeneous data to obtain multi-domain features;

[0048] The feature splicing module is used to acquire the impact factor features of disease events, and splice the impact factor features of disease events with the multi-domain features to obtain comprehensive impact features;

[0049] The cost prediction module is used to predict the costs based on the comprehensive impact characteristics, and obtain a medical cost prediction data sequence.

[0050] A matrix construction module is used to construct a matrix based on the medical cost prediction data sequence to obtain a target probability matrix; wherein, the target probability matrix is ​​used to represent the probability distribution of resource gaps over multiple consecutive time periods;

[0051] The weight calculation module is used to calculate the weights of the preset candidate resource configuration strategies based on the target probability matrix to obtain the strategy weights.

[0052] The strategy filtering module is used to filter the candidate resource configuration strategies according to the strategy weights to obtain the target resource configuration strategy.

[0053] The strategy execution module is used to allocate resources to the target insurance product type according to the target resource configuration strategy.

[0054] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the resource allocation method described in the first aspect.

[0055] To achieve the above objectives, a fourth aspect of the present application provides a storage medium, which is a computer-readable storage medium storing a computer program that, when executed by a processor, implements the resource allocation method described in the first aspect.

[0056] The resource allocation method, device, electronic equipment, and computer-readable storage medium proposed in this application first collect multi-source data on the target insurance product type to obtain multi-source heterogeneous data. This multi-source heterogeneous data includes medical data, insurance data, and macroeconomic data, providing rich information for subsequent prediction of the probability distribution of resource gaps. Further, feature extraction is performed on the multi-source heterogeneous data to obtain multi-domain features. Then, the impact factor features of disease events are obtained, and these features are concatenated with the multi-domain features to obtain comprehensive impact features. Next, cost prediction is performed on the comprehensive impact features to obtain a medical cost prediction data sequence. This allows for cost prediction by integrating multi-dimensional information, improving prediction accuracy. Further, a matrix is ​​constructed based on the medical cost prediction data sequence to obtain a target probability matrix. This target probability matrix represents the probability distribution of resource gaps over multiple consecutive time periods, fully expressing the resource gap situation of the target insurance product type in different time periods. Further, weights are calculated on preset candidate resource allocation strategies based on the target probability matrix to obtain strategy weights. Then, candidate resource allocation strategies are screened based on the strategy weights to obtain the target resource allocation strategy. This allows for obtaining a reasonable strategy under resource gap constraints. Finally, resources are allocated to the target insurance product type based on the target resource allocation strategy. In summary, this application can personalize resource allocation strategies for different insurance product types, improving the accuracy of resource allocation and demonstrating high applicability.

[0057] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0058] Figure 1 This is a flowchart of the resource allocation method provided in the embodiments of this application;

[0059] Figure 2 yes Figure 1The flowchart for step 104 in the document;

[0060] Figure 3 yes Figure 2 The flowchart for step 202 in the document;

[0061] Figure 4 yes Figure 1 The flowchart for step 105 in the document;

[0062] Figure 5 This is a flowchart of a resource allocation method provided in another embodiment of this application;

[0063] Figure 6 yes Figure 5 The flowchart for step 504 in the document;

[0064] Figure 7 This is a block diagram of the module structure of the resource allocation device provided in the embodiments of this application;

[0065] Figure 8 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0067] It should be noted that although functional modules are divided in the device schematic diagram and a 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 aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0068] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0069] First, let's analyze some of the terms used in this application:

[0070] Artificial intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0071] Natural Language Processing (NLP): NLP uses computers to process, understand, and utilize human language (such as Chinese and English). It is a branch of artificial intelligence and an interdisciplinary field of computer science and linguistics, often referred to as computational linguistics. NLP includes syntactic analysis, semantic analysis, and discourse understanding. It is commonly used in machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information and image processing, information extraction and filtering, text classification and clustering, sentiment analysis, and opinion mining. It involves data mining, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research, and linguistic research related to language computation.

[0072] The current resource allocation methods in the fintech field suffer from the following technological fragmentation and shortcomings:

[0073] 1. Data isolation: Most medical cost prediction models are built based on historical medical data (such as the number of cases and treatment costs), without integrating macroeconomic data (such as inflation rate and interest rate fluctuations) and insurance data (such as the situation of insurance fund pools, including premium income and loss ratio). Therefore, it is difficult to obtain accurate prediction results and it is impossible to serve the long-term resource allocation requirements.

[0074] 2. Lack of dynamic matching: Resource allocation strategies (such as fixed income and equity asset allocation) mostly use traditional financial models (mean-variance model), without combining their own medical cost forecast data, so they cannot achieve the goal of dynamically adjusting resource allocation strategies based on future payout risk exposure.

[0075] 3. Inadequate risk control: Current technologies mostly rely on single risk indicators (such as duration matching) or static stress testing, which cannot take into account the synergistic effects of non-linear growth in medical costs (such as major public health emergencies) and extreme fluctuations in financial markets from the perspective of resource gap risk. This poses a risk of underestimating resource gap risk.

[0076] 4. Insufficient personalization: The nature of claims differs for various types of insurance products (such as critical illness insurance, life insurance, and health insurance), and existing algorithms have not established a resource allocation strategy for differentiated claims, resulting in low applicability and efficiency of resource allocation.

[0077] 5. Limitations of long-term modeling: Most current medical cost prediction models are based on a medium-term (3 to 5 years) perspective and do not fully consider the impact of long-term factors such as aging and advancements in medical technology. They are unable to adapt to the requirements of long-term resource allocation over a period of more than ten years.

[0078] Based on this, embodiments of this application propose a resource allocation method, resource allocation device, electronic device, and computer-readable storage medium. By establishing a dynamic coupling framework between medical cost prediction and resource allocation strategy, and through multi-source heterogeneous data fusion, resource gap risk prediction, and dynamic rebalancing, the security and profitability of long-term resource allocation are optimized.

[0079] The resource allocation method provided in this application can be applied to terminals and servers, or it can be software running on the server. The server can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or it can be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the resource allocation method, but it is not limited to the above forms.

[0080] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include server computers, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics devices, network PCs, and distributed computing environments including any of the above systems or devices. This 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. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0081] This application provides a resource allocation method, a resource allocation device, an electronic device, and a computer-readable storage medium, which are specifically described through the following embodiments. First, the resource allocation method in the embodiments of this application is described.

[0082] It should be noted that in each specific implementation of this application, when it is necessary to process data related to the user's identity or characteristics, such as the user's medical data, the user's permission or consent will be obtained first. Moreover, the collection, use and processing of this data will comply with relevant laws, regulations and standards.

[0083] Reference Figure 1 , Figure 1 This is an optional flowchart of the resource allocation method provided in the embodiments of this application, which may include, but is not limited to, steps 101 to 108.

[0084] Step 101: Collect multi-source data for the target insurance product type to obtain multi-source heterogeneous data;

[0085] Step 102: Extract features from multi-source heterogeneous data to obtain multi-domain features;

[0086] Step 103: Obtain the impact factor features of the disease event, and concatenate the impact factor features of the disease event with multi-domain features to obtain the comprehensive impact features;

[0087] Step 104: Perform cost prediction on the comprehensive impact characteristics to obtain a medical cost prediction data sequence;

[0088] Step 105: Construct a matrix based on the medical cost prediction data sequence to obtain the target probability matrix;

[0089] Step 106: Calculate the weights of the preset candidate resource allocation strategies based on the target probability matrix to obtain the strategy weights;

[0090] Step 107: Filter the candidate resource allocation strategies according to the strategy weights to obtain the target resource allocation strategy;

[0091] Step 108: Allocate resources for the target insurance product type according to the target resource allocation strategy.

[0092] Steps 101 to 108 as illustrated in this embodiment first involve collecting multi-source data on the target insurance product type to obtain multi-source heterogeneous data. This multi-source heterogeneous data includes medical data, insurance data, and macroeconomic data, providing rich information for subsequent prediction of the probability distribution of resource gaps. Further, feature extraction is performed on the multi-source heterogeneous data to obtain multi-domain features. Then, disease event impact factor features are acquired and concatenated with multi-domain features to obtain comprehensive impact features. Next, cost prediction is performed on the comprehensive impact features to obtain a medical cost prediction data sequence. This allows for comprehensive cost prediction based on multi-dimensional information, improving prediction accuracy. Further, a matrix is ​​constructed based on the medical cost prediction data sequence to obtain a target probability matrix. This target probability matrix represents the probability distribution of resource gaps over multiple consecutive time periods, fully expressing the resource gap situation of the target insurance product type in different time periods. Further, weights are calculated for preset candidate resource allocation strategies based on the target probability matrix to obtain strategy weights. Then, candidate resource allocation strategies are screened based on the strategy weights to obtain the target resource allocation strategy. This allows for obtaining a reasonable strategy under resource gap constraints. Finally, resources are allocated to the target insurance product type based on the target resource allocation strategy. In summary, this application can personalize resource allocation strategies for different insurance product types, improving the accuracy of resource allocation and demonstrating high applicability.

[0093] In step 101 of some embodiments, multi-source data collection is performed on the target insurance product type to obtain multi-source heterogeneous data. Insurance product type refers to the category of insurance contracts based on different coverage scopes, insured objects, coverage periods, and liability characteristics of the product design. Insurance product types can include critical illness insurance, life insurance, and health insurance, etc., and the target insurance product type is one of several insurance product types. Critical illness insurance (also known as major illness insurance): An insurance product that provides a one-time financial compensation or payment when the insured is diagnosed with a critical illness specified in the contract during the insurance period. Its core purpose is to address the high medical expenses and income interruption risks caused by critical illnesses. Life insurance (life insurance): An insurance product that pays insurance money to the beneficiary upon the death of the insured or the end of the insurance period. Life insurance mainly protects against life risks during the insured's lifetime, providing financial security for the family. Health insurance: An insurance product that primarily covers the insured's health-related medical needs, such as hospitalization, outpatient services, surgery, and disease treatment.

[0094] Multi-source data acquisition specifically refers to collecting data from multiple data sources. A data source is the origin or platform used to collect, store, and provide data, providing raw information for data analysis, modeling, and decision-making. Data sources for medical data can include: hospital electronic health record systems, health insurance companies, and medical statistical databases. Data sources for insurance data can include: insurance company claims systems, customer information management systems, and third-party data providers. Data sources for macroeconomic data can include: financial market data providers, news and financial information websites, and academic research databases.

[0095] Multi-source heterogeneous data includes medical data, insurance data, and macroeconomic data. Medical data refers to various data related to health, disease, diagnosis, treatment, and healthcare services.

[0096] Medical data includes, but is not limited to: disease statistics (disease spectrum change data, incidence rate, treatment costs, etc.), patient electronic health records (such as medical records, drug use, test results, imaging data, etc.), medical institution operational data, and medical device and drug usage data.

[0097] Insurance data refers to data and information related to insurance business, claims, and risk assessment. Insurance data includes, but is not limited to, total insurance funds, premium payments, premium income, and loss ratios. Macroeconomic data refers to statistical data reflecting the overall economic situation of a region or area. Macroeconomic data includes GDP, inflation rate, unemployment rate, and interest rate fluctuations.

[0098] In step 102 of some embodiments, feature extraction is performed on the multi-source heterogeneous data to obtain multi-domain features. Specifically, a temporal neural network (LSTM+Attention) can be used to extract features from the multi-source heterogeneous data to obtain multi-domain features.

[0099] In one example, the process of extracting multi-domain features using a temporal neural network (LSTM+Attention) includes: 1. Data preprocessing: Cleaning, standardizing, and formatting medical data, insurance data, and macroeconomic data respectively. 2. Feature extraction: Extracting features from medical data, insurance data, and macroeconomic data using an LSTM+Attention model. For example, extracting payout trend features from historical payout data in insurance data, extracting disease type change features from disease spectrum change data in medical data, and extracting economic indicator features from macroeconomic data. 3. Feature concatenation: Concatenating the extracted features from each domain to form multi-domain features.

[0100] In step 103 of some embodiments, disease event impact factor features are obtained, and these features are concatenated with multi-domain features to obtain comprehensive impact features. Disease event impact factor features are features extracted from disease event impact factors. Disease event impact factors (which may also be called "disease impact factors" or "disease influence factors") are indicators used to describe the impact or influence of a disease event on individuals, families, societies, or economic systems. Disease event impact factors are indicators or coefficients that measure the intensity or degree of the impact of a disease event (such as infection, major disease, or epidemic outbreak) on relevant subjects (individuals, families, society, economy, etc.).

[0101] In one embodiment, step 103 may include: obtaining the diseases supported by the target insurance product type to obtain the target supported diseases; constructing a target disease transmission model for the target supported diseases; obtaining the historical number of cases, and using the target disease transmission model to predict the number of cases based on the historical number of cases to obtain the target predicted number of cases; and representing the target predicted number of cases using features to obtain the disease event impact factor features. The target supported diseases can be referred to the description of step 202 below. The target disease transmission model is a time-series prediction model, such as LSTM. Different disease transmission models can be set for different diseases. Specifically, the disease event impact factor can be the disease transmission model. The disease transmission model can be used to predict the number of cases to obtain the target predicted number of cases; feature extraction can be performed on the target predicted number of cases to obtain the number of cases features, and the number of cases features can be determined as the disease event impact factor features. The disease event impact factor features may also include payout rate growth trend features, disease cure rate growth trend features, etc.

[0102] The advantage of the above embodiments is that they can individually obtain disease event impact factor characteristics based on multiple diseases supported by the target insurance product type, thereby improving the fine granularity of factor feature collection and helping to improve the accuracy of resource allocation.

[0103] In step 104 of some embodiments, cost prediction is performed on the comprehensive impact characteristics to obtain a medical cost prediction data sequence. The comprehensive impact characteristics can be predicted using a pre-trained medical cost prediction model to obtain the medical cost prediction data sequence. The medical cost prediction model can employ LSTM, Transformer, or other models suitable for time series prediction. The medical cost prediction data sequence includes medical cost prediction data for at least two consecutive time periods.

[0104] In one embodiment, reference is made to Figure 2 Step 104 may include:

[0105] Step 201: Obtain the types of diseases for which the target insurance product type supports compensation, thus obtaining the target supported diseases;

[0106] Step 202: Construct a medical cost prediction model for the target disease;

[0107] Step 203: The medical cost prediction model for the target disease is used to predict the cost of the comprehensive impact characteristics by disease, and the medical cost prediction data subsequence for the target supported disease is obtained.

[0108] Step 204: Integrate the medical cost prediction data subsequences of each target supported disease to obtain the medical cost prediction data sequence.

[0109] In step 201, taking critical illness insurance as an example, the types of diseases (disease categories) supported by critical illness insurance include: (1) Cancer, various malignant tumors, including lung cancer, breast cancer, stomach cancer, liver cancer, lymphoma, leukemia, etc. (2) Major organ transplants or surgeries, such as kidney transplants, heart transplants, liver transplants, etc. (3) Brain diseases, including stroke (cerebrovascular accident), brain tumors, severe brain injury, etc. (4) Cardiovascular diseases, including high-risk heart disease, coronary artery bypass grafting, aortic surgery, etc. (5) Severe burns or scalds, including skin damage or functional impairment caused by severe burns. Therefore, the target supported diseases may include cancer, major organ transplants or surgeries, brain diseases, cardiovascular diseases, and severe burns or scalds.

[0110] In step 202, the target disease medical cost prediction model is a time series model specifically designed to predict the medical costs of the target supported disease. The target disease medical cost prediction model can employ Long Short-Term Memory Networks (LSTM), Transformer, or other models suitable for time series prediction.

[0111] In one embodiment, reference is made to Figure 3 Step 202 may include:

[0112] Step 301: Obtain historical medical expense data for the target supported disease; wherein, the historical medical expense data includes the number of cases, treatment costs, and reimbursement costs for the sample during the historical period.

[0113] Step 302: Extract features from the number of patients and treatment costs in the sample to obtain sample cost features;

[0114] Step 303: Obtain the impact factors of sample disease events within the historical time period, and concatenate the sample cost features and the sample disease event impact factors to obtain the sample impact features.

[0115] Step 304: Predict the cost of the sample impact characteristics using a preset initial cost prediction model to obtain the predicted compensation cost.

[0116] Step 305: Calculate the loss based on the predicted compensation cost and the sample compensation cost to obtain the target loss data;

[0117] Step 306: Adjust the parameters of the initial cost prediction model based on the target loss data to obtain the target disease medical cost prediction model.

[0118] Specifically, the process of establishing the target disease medical cost prediction sub-model is as follows: 1. Data preparation: Collect historical medical cost data for the target supported disease, including the number of cases, treatment costs, and compensation amounts. 2. Feature engineering: Extract data features for the target supported disease, including incidence trends, changes in treatment costs, and the impact of economic indicators. 3. Model design: Design an independent target disease medical cost prediction sub-model for the target supported disease, which can use LSTM, Transformer, or other models suitable for time series prediction. 4. Model training: Train the target disease medical cost prediction sub-model using historical data, and adjust the model parameters to optimize prediction accuracy. 5. Model validation: Evaluate the model performance through cross-validation or holding a validation set to ensure the prediction accuracy of the target disease medical cost prediction sub-model.

[0119] The advantage of the embodiments of steps 301 to 306 above is that it is possible to independently establish a disease-specific cost prediction model based on the disease, improve the accuracy of medical cost prediction, and adapt to the requirements of long-term resource allocation.

[0120] In step 203, the medical cost prediction model for the target disease is used to predict the costs of the comprehensive impact characteristics by disease, resulting in a subsequence of medical cost prediction data for the target supported disease. This subsequence of medical cost prediction data includes medical cost prediction data for the target supported disease over at least two time periods.

[0121] In step 204, the medical cost prediction data subsequences of each target supported disease are integrated. Specifically, the medical cost prediction data of each target supported disease corresponding to each time period are added together to obtain the overall medical cost prediction data corresponding to that time period, thereby obtaining the medical cost prediction data sequence.

[0122] The advantage of the embodiments of steps 201 to 204 above is that by first predicting medical expenses by disease type and then integrating them to obtain a medical expense prediction data sequence, the accuracy of medical expense prediction can be improved and the difficulty of prediction can be reduced.

[0123] In step 105 of some embodiments, a target probability matrix is ​​constructed based on the medical cost prediction data sequence to represent the probability distribution of resource gaps over multiple consecutive time periods.

[0124] In one embodiment, reference is made to Figure 4 Step 105 may include:

[0125] Step 401: Organize the medical expense prediction data series by time period to obtain the predicted reimbursement amount for each time period;

[0126] Step 402: Perform matrix transformation based on the predicted compensation amount to construct a resource flow stress test matrix; where each element in the resource flow stress test matrix represents the resource gap corresponding to a time period.

[0127] Step 403: Perform probability analysis on each element in the resource flow stress test matrix to obtain the target probability matrix.

[0128] Specifically, the transformation process of converting the medical expense prediction data sequence into a target probability matrix is ​​as follows: 1. Prediction result organization: Organize the medical expense prediction data sequence by time period to obtain the predicted reimbursement amount for each time period. 2. Resource flow stress test matrix construction: Based on the predicted reimbursement amount, construct a resource flow stress test matrix, where each element represents the resource gap within the corresponding time period. 3. Probability distribution calculation: Calculate the probability distribution of the resource gap within each time period using Monte Carlo simulation or other probability analysis methods. The aforementioned resource gap may include funding gaps, manpower gaps, equipment gaps, etc.

[0129] The advantage of the embodiments of steps 401 to 403 above is that the probability distribution of resource gaps in each time period can be quantified, which can be used to guide the subsequent determination of resource allocation strategies and help improve the accuracy of resource allocation.

[0130] In step 106 of some embodiments, the preset candidate resource allocation strategies are weighted according to the target probability matrix to obtain the strategy weights. The strategy weights refer to the probability that the target insurance product type will select a candidate resource allocation strategy. In one example, a risk budget-driven investment model is established, with the minimization of funding gap risk in the resource gap as the optimization objective function. Solvency adequacy ratio constraints, regulatory requirement constraints, and return volatility constraints are set, and a stochastic programming algorithm is used to obtain the strategy weights of the candidate resource allocation strategies.

[0131] In step 107 of some embodiments, candidate resource allocation strategies are screened based on strategy weights to obtain target resource allocation strategies. A higher strategy weight indicates that the candidate resource allocation strategy is more suitable for the target insurance product type, and therefore, the more likely the candidate resource allocation strategy is to become the target resource allocation strategy. Conversely, a lower strategy weight indicates that the candidate resource allocation strategy is less suitable for the target insurance product type, and therefore, the less likely the candidate resource allocation strategy is to become the target resource allocation strategy.

[0132] In step 108 of some embodiments, resources are allocated to the target insurance product type according to the target resource allocation strategy. Specifically, a target resource allocation strategy is selected according to the target insurance product type (e.g., life insurance and health insurance), which specifies suitable investment duration, risk limits, etc. For example, critical illness insurance tends to choose short- to medium-term highly liquid assets; while pension insurance can choose some long-term inflation-resistant assets.

[0133] In one embodiment, reference is made to Figure 5 After step 108, the resource allocation method may further include:

[0134] Step 501: Obtain the total remaining resources;

[0135] Step 502: Estimate the resource repayment amount for each target insurance product type.

[0136] Step 503: Based on the total remaining resources and the estimated resource repayment amount, risk prediction is performed to obtain the resource allocation risk probability;

[0137] Step 504: Update the target resource allocation strategy based on the resource allocation risk probability.

[0138] In step 501, considering that resources were retrieved from the preset resource pool in step 108 according to the target resource allocation strategy for the target insurance product type, there are still remaining resources in the preset resource pool. The total amount of remaining resources can be obtained by statistically analyzing these remaining resources. The preset resource pool refers to a collection of similar or identical resources that are centrally stored, managed, and scheduled. The preset resource pool may include an insurance fund pool. An insurance fund pool refers to a collection of funds used for insurance investment.

[0139] In step 502, considering that different types of insurance products have different natures and risk characteristics, resource repayment forecasting is needed in advance to avoid future defaults. For example, critical illness insurance has a shorter repayment period but requires a higher amount of repayment resources. On the other hand, pension insurance has a longer repayment period and requires a lower amount of repayment resources. Resource repayment forecasting can be performed using a pre-trained resource prediction model, such as LSTM.

[0140] In step 503, the ratio of the remaining total resources to the estimated resource repayment amount can be calculated to obtain the resource coverage ratio. The resource allocation risk probability is then determined based on the resource coverage ratio. A higher resource coverage ratio indicates a lower resource allocation risk probability, while a lower resource coverage ratio indicates a higher resource allocation risk probability.

[0141] In step 504, the target resource allocation strategy is updated based on the resource allocation risk probability. The higher the resource allocation risk probability, the more necessary it is to update the target resource allocation strategy.

[0142] In one embodiment, reference is made to Figure 6 Step 504 may include:

[0143] Step 601: If the resource allocation risk probability meets the predetermined risk conditions, then determine the flow type of the target resource allocation strategy and obtain the strategy flow type.

[0144] Step 602: In response to the strategy flow type indicating that the liquidity of the target resource allocation strategy is high, the target resource allocation strategy is deleted.

[0145] Step 603: In response to the low liquidity of the strategy flow type used to indicate that the target resource allocation strategy has low liquidity, an auxiliary resource allocation strategy is selected from the candidate resource allocation strategies, and the target resource allocation strategy is replaced by the auxiliary resource allocation strategy.

[0146] In step 601, the predetermined risk condition can be that the resource allocation risk probability is less than or equal to a risk probability threshold. The liquidity type of the target resource allocation strategy can be determined through a predetermined classification algorithm / rule. The strategy liquidity type indicates whether the liquidity of the target resource allocation strategy is high or low. For example, if the target resource allocation strategy specifies that the target insurance product type should select short- to medium-term highly liquid assets, then the liquidity of the strategy liquidity type is high. Conversely, if the target resource allocation strategy specifies that the target insurance product type should select long-term, illiquid assets, then the liquidity of the strategy liquidity type is low.

[0147] In step 602, if the strategy flow type is used to indicate that the liquidity of the target resource allocation strategy is high, it means that the period for recovering the resources of the target resource allocation strategy cannot be shortened further. However, the probability of resource allocation risk is already too high at this time, so the target resource allocation strategy is deleted.

[0148] In step 603, if the strategy flow type indicates low liquidity for the target resource allocation strategy, it means the timeframe for recovering resources under the target resource allocation strategy can be further shortened. Therefore, an auxiliary resource allocation strategy is selected from the candidate resource allocation strategies, and the target resource allocation strategy is replaced by the auxiliary resource allocation strategy. The strategy flow type of the auxiliary resource allocation strategy indicates high liquidity.

[0149] The advantage of the embodiments of steps 601 to 603 described above is that the resource allocation strategy can be dynamically adjusted in a timely manner according to the probability of resource allocation risk, reducing the possibility of default and improving the accuracy and adaptability of resource allocation.

[0150] In one example, after step 108, the resource allocation method may further include: updating the resource allocation strategy based on key points output by the medical cost prediction model (e.g., population inflection points); or, conducting real-time monitoring of funding coverage (present value of investment assets / present value of future payouts) and updating the resource allocation strategy when thresholds are exceeded. Methods for updating the resource allocation strategy include: using new derivative investments or physical investments, such as interest rate swaps, longevity risk securitization products, and healthcare real estate investment trusts, etc.

[0151] In one example, extreme scenario stress testing and resilience enhancement can be conducted, including: 1. Designing multi-layered stress scenarios: 1.1: The reference scenario assumes an annual growth rate of 6% in medical costs and a 4% return on investment. 1.2: A sudden infectious disease outbreak increases total costs by 30%, leading to a simultaneous decline in both stocks and bonds. 1.3: The widespread adoption of gene therapy lowers long-term costs. 2. Using Monte Carlo simulations to calculate over 100,000 investment paths to test the robustness of the investment strategy and to specifically enhance resilience-resistant resource allocation strategies.

[0152] In one example, visualized decision support and feedback learning can also be implemented, including: 1. Creating an interactive dashboard and visualizing key indicators across the entire "prediction-investment-risk control" chain (such as dynamic funding gaps, risk budget consumption, etc.). 2. Embedding a reinforcement learning feedback loop: adjusting the compensation amount and investment return difference according to the actual situation, adjusting model parameters and strategy rule base, and then iterating again in a closed loop.

[0153] In summary, the present application achieves at least the following beneficial effects: 1. Cross-domain collaboration: Breaking down data barriers between healthcare and finance, forming a closed-loop cycle from prediction to strategy feedback. 2. Dynamic adaptation: Detecting the probability of resource allocation risks to mitigate market uncertainties and dangers posed by the healthcare environment. 3. Refined risk management: Upgrading from simple duration matching to multi-dimensional risk budget control. 4. Enhanced interpretability: The visualization platform supports regulatory compliance and human intervention.

[0154] Please see Figure 7 This application embodiment also provides a resource allocation device that can implement the above-described resource allocation method. Figure 7The present application provides a module structure block diagram of a resource allocation device, which includes: a data acquisition module 701, a feature extraction module 702, a feature splicing module 703, a cost prediction module 704, a matrix construction module 705, a weight calculation module 706, a strategy filtering module 707, and a strategy execution module 708. The system comprises the following modules: a data acquisition module 701, used to collect multi-source data on the target insurance product type to obtain multi-source heterogeneous data, including medical data, insurance data, and macroeconomic data; a feature extraction module 702, used to extract features from the multi-source heterogeneous data to obtain multi-domain features; a feature concatenation module 703, used to obtain disease event impact factor features and concatenate these features with multi-domain features to obtain comprehensive impact features; a cost prediction module 704, used to predict costs based on the comprehensive impact features to obtain a medical cost prediction data sequence; a matrix construction module 705, used to construct a matrix based on the medical cost prediction data sequence to obtain a target probability matrix, which represents the probability distribution of resource gaps over multiple consecutive time periods; a weight calculation module 706, used to calculate the weights of preset candidate resource allocation strategies based on the target probability matrix to obtain strategy weights; a strategy selection module 707, used to select candidate resource allocation strategies based on the strategy weights to obtain a target resource allocation strategy; and a strategy execution module 708, used to allocate resources to the target insurance product type according to the target resource allocation strategy.

[0155] In one embodiment, the resource allocation device further includes a strategy update module, configured to: obtain the total remaining resources; estimate the resource repayment for each target insurance product type to obtain the estimated resource repayment amount; predict the risk based on the total remaining resources and the estimated resource repayment amount to obtain the resource allocation risk probability; and update the target resource allocation strategy based on the resource allocation risk probability.

[0156] It should be noted that the specific implementation of this resource allocation device is basically the same as the specific implementation of the resource allocation method described above, and will not be repeated here.

[0157] This application also provides an electronic device, which includes: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for communication between the processor and the memory. When the program is executed by the processor, it implements the resource allocation method described above. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0158] Please see Figure 8 , Figure 8 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0159] The processor 801 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, 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 this application.

[0160] The memory 802 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 802 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 802 and is called and executed by the processor 801 using the resource allocation method of the embodiments of this application.

[0161] The 803 input / output interface is used to implement information input and output.

[0162] The communication interface 804 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0163] Bus 805 transmits information between various components of the device (e.g., processor 801, memory 802, input / output interface 803, and communication interface 804);

[0164] The processor 801, memory 802, input / output interface 803, and communication interface 804 are connected to each other within the device via bus 805.

[0165] This application embodiment also provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, which can be executed by one or more processors to implement the above-described resource allocation method.

[0166] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0167] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0168] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0169] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0170] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0171] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0172] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0173] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0174] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0175] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0176] If the integrated unit is implemented as 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 this application, in essence, 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. This computer software product is stored in a storage medium and includes multiple instructions to cause an electronic device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0177] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A resource allocation method, characterized in that, The method includes: Multi-source data collection is performed on the target insurance product type to obtain multi-source heterogeneous data; wherein, the multi-source heterogeneous data includes medical data, insurance data and macroeconomic data; Feature extraction is performed on the multi-source heterogeneous data to obtain multi-domain features; Obtain the impact factor features of disease events, and concatenate the impact factor features of disease events with the multi-domain features to obtain comprehensive impact features; Cost prediction is performed on the comprehensive impact characteristics to obtain a medical cost prediction data sequence; A target probability matrix is ​​obtained by constructing a matrix based on the medical cost prediction data sequence; wherein, the target probability matrix is ​​used to represent the probability distribution of resource gaps over multiple consecutive time periods; The preset candidate resource allocation strategies are weighted according to the target probability matrix to obtain the strategy weights; The candidate resource allocation strategies are filtered according to the strategy weights to obtain the target resource allocation strategy; Resources are allocated to the target insurance product type according to the target resource allocation strategy.

2. The method according to claim 1, characterized in that, The process of predicting costs based on the comprehensive impact characteristics yields a medical cost prediction data sequence, including: Obtain the types of diseases for which the target insurance product type supports compensation, thus obtaining the target supported diseases; Construct a medical cost prediction model for the target supported diseases; The medical cost prediction model for the target disease is used to predict the comprehensive impact characteristics by disease, thereby obtaining a subsequence of medical cost prediction data for the target supported disease; wherein, the subsequence of medical cost prediction data includes medical cost prediction data for the target supported disease in at least two time periods; The medical cost prediction data subsequences of each of the target supported diseases are integrated to obtain the medical cost prediction data sequence.

3. The method according to claim 2, characterized in that, The construction of a medical cost prediction model for the target supported disease includes: Obtain historical medical expense data for the target supported disease; wherein, the historical medical expense data includes the number of cases, treatment costs, and reimbursement costs for the sample during the historical time period; Feature extraction is performed on the number of patients in the sample and the treatment cost of the sample to obtain sample cost features; Obtain the impact factor of sample disease events within the historical time period, and concatenate the sample cost feature and the sample disease event impact factor to obtain the sample impact feature; The sample impact characteristics are predicted using a preset initial cost prediction model to obtain the predicted compensation cost. Loss calculations are performed based on the predicted compensation costs and the sample compensation costs to obtain target loss data; The parameters of the initial cost prediction model are adjusted based on the target loss data to obtain the medical cost prediction model for the target disease.

4. The method according to any one of claims 1 to 3, characterized in that, The step of constructing a target probability matrix based on the medical cost prediction data sequence includes: The medical expense prediction data sequence is organized by time period to obtain the predicted reimbursement amount for each time period; Based on the predicted compensation amount, a matrix transformation is performed to construct a resource flow stress test matrix; wherein, each element in the resource flow stress test matrix is ​​used to represent the resource gap corresponding to a time period; Probability analysis is performed on each element in the resource flow stress test matrix to obtain the target probability matrix.

5. The method according to any one of claims 1 to 3, characterized in that, After allocating resources to the target insurance product type according to the target resource allocation strategy, the method further includes: Get the total remaining resources; For each of the target insurance product types, resource repayment estimates are performed to obtain the estimated resource repayment amount; Based on the total remaining resources and the estimated resource repayment amount, a risk prediction is made to obtain the resource allocation risk probability. The target resource allocation strategy is updated based on the resource allocation risk probability.

6. The method according to claim 5, characterized in that, The step of updating the target resource allocation strategy based on the resource allocation risk probability includes: If the resource allocation risk probability meets the predetermined risk conditions, then the flow type of the target resource allocation strategy is determined to obtain the strategy flow type; wherein, the strategy flow type is used to indicate whether the liquidity of the target resource allocation strategy is high or low; If the policy flow type used to indicate that the liquidity of the target resource allocation policy is high, then the target resource allocation policy is deleted. In response to the strategy flow type indicating low liquidity of the target resource allocation strategy, an auxiliary resource allocation strategy is selected from the candidate resource allocation strategies, and the target resource allocation strategy is replaced by the auxiliary resource allocation strategy; wherein, the strategy flow type of the auxiliary resource allocation strategy indicates high liquidity.

7. The method according to any one of claims 1 to 3, characterized in that, The acquisition of disease event impact factor characteristics includes: Obtain the types of diseases for which the target insurance product type supports compensation, thus obtaining the target supported diseases; To develop target disease transmission models for target support diseases; The historical number of cases is obtained, and the target disease transmission model is used to predict the number of cases based on the historical number of cases to obtain the target predicted number of cases. Based on the predicted number of cases, the features of the disease event impact factor are represented to obtain the features.

8. A resource allocation device, characterized in that, The device includes: The data acquisition module is used to collect multi-source data on the target insurance product type to obtain multi-source heterogeneous data; wherein, the multi-source heterogeneous data includes medical data, insurance data and macroeconomic data; The feature extraction module is used to extract features from the multi-source heterogeneous data to obtain multi-domain features; The feature splicing module is used to acquire the impact factor features of disease events, and splice the impact factor features of disease events with the multi-domain features to obtain comprehensive impact features; The cost prediction module is used to predict the costs based on the comprehensive impact characteristics, and obtain a medical cost prediction data sequence. A matrix construction module is used to construct a matrix based on the medical cost prediction data sequence to obtain a target probability matrix; wherein, the target probability matrix is ​​used to represent the probability distribution of resource gaps over multiple consecutive time periods; The weight calculation module is used to calculate the weights of the preset candidate resource configuration strategies based on the target probability matrix to obtain the strategy weights. The strategy filtering module is used to filter the candidate resource configuration strategies according to the strategy weights to obtain the target resource configuration strategy. The strategy execution module is used to allocate resources to the target insurance product type according to the target resource configuration strategy.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the resource allocation method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the resource allocation method according to any one of claims 1 to 7.

Images