Data processing method and device based on artificial intelligence, computer equipment and medium

Abstract

The invention belongs to the technical field of artificial intelligence, and relates to a data processing method and device based on artificial intelligence, computer equipment and a storage medium, and the method comprises the steps: receiving a business processing request corresponding to a target client; analyzing the service processing request to extract model identification information; determining a candidate containerization unit corresponding to the model identification information based on the model routing component; screening a target containerization unit from the candidate containerization units based on a load balancing algorithm; preprocessing the customer information of the target customer to obtain feature data; reading model parameters of the target prediction model based on the target containerization unit, and performing evaluation processing on the feature data based on the target prediction model to obtain a data evaluation result; and outputting a data evaluation result. In addition, the invention also relates to a block chain technology, and the data evaluation result can be stored in a block chain. The method can be applied to business data evaluation scenes in the field of financial science and technology, and the processing efficiency of data evaluation is effectively improved.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology and can be applied to the financial technology field, particularly to data processing methods, devices, computer equipment and storage media based on artificial intelligence. Background Technology

[0002] In traditional financial data assessment scenarios, such as those related to motor vehicle insurance premium rate determination and pricing service models, customer data assessment primarily relies on embedding the predictive model's file (e.g., the model file itself) into application components for loading and reading, consuming application memory and CPU for predictive calculations. Specifically, this method requires embedding the model file into the application package, necessitating repackaging and application restarts with each model update iteration. This results in long model update cycles, low deployment efficiency, and consequently, inefficient data assessment processing, failing to meet the demands of rapid business growth. This static, embedded model assessment method lacks flexibility, failing to respond promptly to market changes and dynamic adjustments to customer needs, thus limiting the accuracy and timeliness of data assessment. For example, in credit scoring card model updates within the financial sector, traditional methods may require weeks to complete the update and deployment due to model embedding, preventing new risk characteristics from being incorporated into the assessment in a timely manner, increasing the credit risk for financial institutions.

[0003] Therefore, there is an urgent need to provide a dynamic data evaluation method to enhance the flexibility and accuracy of data evaluation, thereby improving the overall effectiveness of financial services. Summary of the Invention

[0004] The purpose of this application is to propose a data processing method, apparatus, computer device, and storage medium based on artificial intelligence, in order to solve the technical problem of low processing efficiency in existing customer data evaluation methods that rely on embedding model files of prediction models into application components for loading and reading, and occupying application memory and CPU for prediction calculations.

[0005] Firstly, an artificial intelligence-based data processing method is provided, including:

[0006] Receive a business processing request corresponding to a target customer; wherein, the business processing request carries model identification information;

[0007] The business processing request is parsed to extract the corresponding model identification information;

[0008] Based on the preset model routing component, candidate containerization units corresponding to the model identification information are determined;

[0009] The target containerized unit is selected from the candidate containerized units based on a preset load balancing algorithm.

[0010] Obtain the customer information of the target customer, and preprocess the customer information to obtain corresponding feature data;

[0011] Based on the target containerization unit, the model parameters of the target prediction model corresponding to the model identification information are read, and the feature data is evaluated based on the target prediction model to obtain the corresponding data evaluation result;

[0012] The data evaluation results are then processed for output.

[0013] Secondly, an artificial intelligence-based data processing device is provided, comprising:

[0014] A receiving module is used to receive a business processing request corresponding to a target customer; wherein the business processing request carries model identification information;

[0015] The extraction module is used to parse the business processing request to extract the corresponding model identification information;

[0016] The determination module is used to determine the candidate containerization unit corresponding to the model identification information based on a preset model routing component;

[0017] The filtering module is used to filter out target containerized units from the candidate containerized units based on a preset load balancing algorithm.

[0018] The processing module is used to acquire customer information of the target customer and preprocess the customer information to obtain corresponding feature data;

[0019] The evaluation module is used to read the model parameters of the target prediction model corresponding to the model identification information based on the target containerization unit, and to evaluate the feature data based on the target prediction model to obtain the corresponding data evaluation results.

[0020] The output module is used to process the data evaluation results.

[0021] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described artificial intelligence-based data processing method.

[0022] Fourthly, a computer-readable storage medium is provided, which stores a computer program that, when executed by a processor, implements the steps of the aforementioned artificial intelligence-based data processing method.

[0023] In the above-mentioned solution implemented by the data processing method, apparatus, computer equipment, and storage medium based on artificial intelligence, the following steps are taken: First, a business processing request corresponding to a target customer is received; wherein, the business processing request carries model identification information; then, the business processing request is parsed to extract the corresponding model identification information; next, a candidate containerization unit corresponding to the model identification information is determined based on a preset model routing component; and a target containerization unit is selected from the candidate containerization units based on a preset load balancing algorithm; subsequently, the customer information of the target customer is obtained, and the customer information is preprocessed to obtain corresponding feature data; further, the model parameters of the target prediction model corresponding to the model identification information are read based on the target containerization unit, and the feature data is evaluated based on the target prediction model to obtain the corresponding data evaluation result; finally, the data evaluation result is output. Based on the above automated processing flow, after receiving a business processing request corresponding to a target customer, this application parses the request to extract the corresponding model identification information. Then, based on the use of the model routing component, candidate containerized units corresponding to the model identification information are determined. Subsequently, based on the use of a load balancing algorithm, a target containerized unit is selected from the candidate containerized units. Next, customer information of the target customer is obtained and preprocessed to obtain feature data. Then, based on the target containerized unit, the model parameters of the target prediction model corresponding to the model identification information are read, and the feature data is evaluated based on the use of the target prediction model to obtain data evaluation results. Finally, the data evaluation results are output. Thus, unlike existing customer data evaluation methods that rely on embedding the prediction model file into application components for loading and reading, consuming application memory and CPU for prediction calculations, this application, through the combined use of the model routing component, load balancing algorithm, and containerized units, can achieve dynamic management and allocation of the prediction model, improving the stability and scalability of the prediction model. It realizes automatic and accurate data evaluation processing of customer feature data based on the accurate use of the target prediction model, effectively improving the processing efficiency of data evaluation. Attached Figure Description

[0024] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is an exemplary system architecture diagram to which this application can be applied;

[0026] Figure 2 This is a flowchart of an embodiment of the artificial intelligence-based data processing method according to this application;

[0027] Figure 3 This is a schematic diagram of a structure of an embodiment of the artificial intelligence-based data processing apparatus according to this application;

[0028] Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0029] 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 pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0030] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0031] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0032] like Figure 1 As shown, system architecture 100 may include terminal device 101, network 102, and server 103. Terminal device 101 may be a laptop 1011, tablet 1012, or mobile phone 1013. Network 102 is used as a medium to provide a communication link between terminal device 101 and server 103. Network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0033] Users can use terminal device 101 to interact with server 103 via network 102 to receive or send messages, etc. Various communication client applications can be installed on terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0034] Terminal device 101 can be various electronic devices with a display screen and support web browsing. In addition to laptops 1011, tablets 1012, or mobile phones 1013, terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 player (Moving Picture Experts Group Audio Layer IV), a laptop computer, and a desktop computer, etc.

[0035] Server 103 can be a server that provides various services, such as a backend server that provides support for the pages displayed on terminal device 101.

[0036] It should be noted that the data processing method based on artificial intelligence provided in the embodiments of this application is generally executed by a server / terminal device, and correspondingly, the data processing device based on artificial intelligence is generally set in the server / terminal device.

[0037] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0038] Continue to refer to Figure 2 The flowchart illustrates an embodiment of the AI-based data processing method according to this application. The order of steps in the flowchart can be changed, and some steps can be omitted, depending on different needs. The AI-based data processing method provided in this application can be applied to any scenario requiring data evaluation, and thus can be applied to products in these scenarios, such as data evaluation products in the financial and insurance fields. The AI-based data processing method includes the following steps:

[0039] Step S201: Receive a business processing request corresponding to the target customer; wherein the business processing request carries model identification information.

[0040] In this embodiment, the data processing method based on artificial intelligence runs on an electronic device (e.g., Figure 1The server / terminal device shown can obtain business processing requests corresponding to the target customer through wired or wireless connections. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra-Width Band) connections, and other currently known or future-developed wireless connection methods. The executing entity of this application is specifically a data processing system, also known as a car insurance business system, which can be simply referred to as the system. This application can be applied to data evaluation scenarios in the financial insurance field, such as car insurance risk assessment scenarios and car insurance pricing scenarios, etc. The demand triggering scenarios for the aforementioned business processing requests include: when a customer submits a car insurance application through the online application page of the car insurance business system, the system will automatically trigger the corresponding business processing request. For example, after the customer fills in the vehicle information and owner information and clicks the "Get Premium" button, the business system will prepare to send the corresponding business processing request to the model routing component according to preset rules and processes. Alternatively, it could include situations where sales personnel use the sales terminal of the auto insurance business system to calculate premiums for customers, which also triggers corresponding business processing requests. After the sales personnel enter the customer's relevant information, the auto insurance business system will also begin preparing for the business processing request.

[0041] Step S202: The business processing request is parsed to extract the corresponding model identification information.

[0042] In this embodiment, the model routing component can be used to parse the business processing request to extract the model identification information, or model ID, carried in the request (such as the X-Model-ID field in the HTTP header or the modelID field in the JSON body). The model ID is a unique identifier used to distinguish different prediction models, which are models built based on the XGBoost model.

[0043] The model building process for the prediction model includes: 1. Collecting car insurance-related data, including vehicle information (such as model, age, and price), owner information (such as age, gender, driving experience, and driving records), and historical claims information. This data will serve as input features for model training. 2. Building the model training unit: By providing complete tools and environment for model training, the model can be effectively trained. Specific operations include: building a model training unit consisting of a data processing module, a model training engine, a parallel computing framework, parameter tuning tools, model evaluation and validation, model saving and loading, and visualization tools. The data processing module cleans, preprocesses, and performs feature engineering on the collected data, transforming the raw data into a format suitable for model training. The model training engine uses the XGBoost algorithm to train the model based on the processed data. A parallel computing framework accelerates the training process. Parameter tuning tools optimize model parameters to improve model performance. Model evaluation and validation use appropriate evaluation metrics (such as accuracy, recall, and mean squared error) to evaluate and validate the trained model. The model saving and loading tool saves the trained model as a bin file for later use. A visualization tool visually displays various metrics and data during the model training process. 3. Model Training and Saving: By generating a model file that can be used for auto insurance risk prediction / assessment, the specific operation is as follows: The processed data is input into the model training engine, and the XGBoost algorithm is used for training. During training, the model parameters are continuously adjusted using the parameter tuning tool to obtain better model performance. After training, the model saving and loading tool saves the trained model as a bin file and stores it in a unified model storage system (such as a distributed file system or cloud storage). Among them, multiple different versions of the prediction model can be built according to actual needs by utilizing their different training data, and each prediction model corresponds to a model ID.

[0044] In addition, the system includes a model management component for effective management and dynamic allocation of model runtime resources, ensuring stable model operation. Container orchestration tools (such as Kubernetes) can be used to dynamically create and destroy Pods (containerized units). Automated management of model Pods is achieved through the Kubernetes API, including Pod lifecycle management, health checks, and automatic recovery. Simultaneously, based on the model bin files in the model storage system, a corresponding Pod is created for each model, and the model files are loaded into the Pod. Furthermore, the system deploys model runtime units to provide independent runtime environments for models, achieving isolation between models and dynamic resource allocation. Specifically, based on the configuration of the model management component, different prediction models are deployed separately in different Pods, achieving isolation between model execution and invocation. The number of Pods and resource allocation (such as CPU and memory) are dynamically adjusted based on the actual invocation volume of the models. For example, for models with high invocation volume, the number of Pods and memory are automatically increased to improve processing capacity; for models with low invocation volume, the number of Pods and memory are reduced to save resources. At the same time, a configuration interface is provided, allowing administrators or the system to manually adjust the number of Pods and resource allocation (such as CPU, memory, etc.) according to business needs, for example, allocating more resources to high-priority models to ensure their performance.

[0045] Step S203: Based on the preset model routing component, determine the candidate containerization unit corresponding to the model identification information.

[0046] In this embodiment, the aforementioned model routing component is a pre-configured route for flexible model invocation, ensuring that business applications can accurately invoke the functions of the required prediction models. The model routing component provides flexible routing rule configuration, supporting precise routing based on model identifiers and dynamic routing based on business logic. It can invoke the corresponding prediction model on demand based on the model ID passed in by the business application. The specific implementation process of determining the candidate containerization unit corresponding to the model identifier information based on the pre-configured model routing component will be described in further detail in subsequent embodiments of this application, and will not be elaborated upon here.

[0047] Step S204: Select the target containerized unit from the candidate containerized units based on the preset load balancing algorithm.

[0048] In this embodiment, the specific implementation process of selecting the target containerized unit from the candidate containerized units based on the preset load balancing algorithm will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0049] Step S205: Obtain the customer information of the target customer and preprocess the customer information to obtain the corresponding feature data.

[0050] In this embodiment, the process of collecting customer information from the aforementioned target customers includes: Auto insurance sales personnel gather detailed information about customers through face-to-face communication, telephone conversations, etc. Regarding vehicle information, sales personnel will inquire about the vehicle's model, age, price, and intended use, recording this information in a dedicated customer information form. Alternatively, a customer self-service platform can be used to guide customers to fill in relevant information themselves through an online form. The platform will set required and optional fields to ensure the collection of key information while improving the convenience for customers. Then, the sales personnel or platform administrator will promptly enter the collected customer information into the auto insurance business system. During the entry process, the system will perform some basic checks, such as checking whether the vehicle model is in the preset list and whether the vehicle age is a reasonable value. For any ambiguous or missing information, the system will generate a prompt message, requiring the sales personnel or administrator to communicate with the customer for confirmation. For example, if the customer's occupation information is unclear, the system will prompt for further verification.

[0051] Furthermore, the system can encapsulate the organized customer information (i.e., feature data) according to the data format agreed upon with the model routing component. If JSON format is used, the system will encapsulate vehicle information, owner information, driving records, etc., in their respective fields to form a complete JSON object. Additionally, the specific implementation process of preprocessing the customer information to obtain the corresponding feature data will be further described in detail in subsequent embodiments of this application, and will not be elaborated upon here.

[0052] Step S206: Based on the target containerization unit, read the model parameters of the target prediction model corresponding to the model identification information, and evaluate the feature data based on the target prediction model to obtain the corresponding data evaluation result.

[0053] In this embodiment, the model parameters of the trained target prediction model, which match the model identification information, are read using the aforementioned target containerization unit. This feature data is then input into the target prediction model for prediction calculation. The target prediction model assesses the target customer's auto insurance risk based on learned data patterns and feature importance, and outputs corresponding data assessment results. For example, the model analyzes data from multiple dimensions, such as vehicle information, owner information, and driving records, to comprehensively determine the probability of a customer making a claim and generate corresponding prediction results, such as risk score data.

[0054] Step S207: Output the data evaluation results.

[0055] In this embodiment, the specific implementation process of outputting the data evaluation results described above will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0056] This application first receives a business processing request corresponding to a target customer; wherein the business processing request carries model identification information; then, the business processing request is parsed to extract the corresponding model identification information; then, a candidate containerization unit corresponding to the model identification information is determined based on a preset model routing component; and a target containerization unit is selected from the candidate containerization units based on a preset load balancing algorithm; subsequently, the customer information of the target customer is obtained, and the customer information is preprocessed to obtain corresponding feature data; further, the model parameters of the target prediction model corresponding to the model identification information are read based on the target containerization unit, and the feature data is evaluated based on the target prediction model to obtain the corresponding data evaluation result; finally, the data evaluation result is output. Based on the above automated processing flow, after receiving a business processing request corresponding to a target customer, this application parses the request to extract the corresponding model identification information. Then, based on the use of the model routing component, candidate containerized units corresponding to the model identification information are determined. Subsequently, based on the use of a load balancing algorithm, a target containerized unit is selected from the candidate containerized units. Next, customer information of the target customer is obtained and preprocessed to obtain feature data. Then, based on the target containerized unit, the model parameters of the target prediction model corresponding to the model identification information are read, and the feature data is evaluated based on the use of the target prediction model to obtain data evaluation results. Finally, the data evaluation results are output. Thus, unlike existing customer data evaluation methods that rely on embedding the prediction model file into application components for loading and reading, consuming application memory and CPU for prediction calculations, this application, through the combined use of the model routing component, load balancing algorithm, and containerized units, can achieve dynamic management and allocation of the prediction model, improving the stability and scalability of the prediction model. It realizes automatic and accurate data evaluation processing of customer feature data based on the accurate use of the target prediction model, effectively improving the processing efficiency of data evaluation.

[0057] In some alternative implementations, step S203 includes the following steps:

[0058] Retrieve preset routing rules.

[0059] In this embodiment, routing rules are constructed based on actual business needs. These routing rules typically exist in the form of structured text (such as YAML / JSON) or native Kubernetes resources (ConfigMap / Secret). Each routing rule specifies the correspondence between a model ID range (or model identification information) and a model service group, and may include metadata such as protocol and timeout.

[0060] The model identification information is mapped based on the routing rules to obtain the corresponding model service group.

[0061] In this embodiment, the model identification information, i.e. the model ID, is mapped to a specific model service group (Dep oy oy t / Stateful Set), or target service group, through pre-configured routing rules.

[0062] The model service group location process includes: Kubernetes resource query: Based on the matched `target_service` field value, the routing component locates the service group in the following ways: Dependency: Suitable for stateless model services, supporting horizontal scaling. Stateful Set: Suitable for stateful models (such as scenarios requiring persistent caching), each containerized unit (Pod) has a unique identifier. If a service discovery mechanism is used: Directly query the Kubernetes API: via `kubectl get deployment`. <name>Equivalent API for retrieving service group details. Caching acceleration: The routing component can cache service group information (such as the number of Pods and tags) to reduce real-time query overhead.

[0063] Get the preset filter criteria.

[0064] In this embodiment, the above-mentioned screening criteria refer to the criteria of being healthy and matching the tags.

[0065] Based on the model service group, all specified containerized units that meet the filtering criteria are selected.

[0066] In this embodiment, the selection process for the specified containerization unit includes: Pod list acquisition: Pods that meet the criteria are selected using the Selector tag of the model service group (e.g., app=risk-model). For example, if the Selector for Dep loyment risk-mode l-dep loyment is app=risk-model, version=v1, then all Pods with these two tags are selected. Health status filtering: Pods that are not in the Ready state (e.g., starting, CrashLoopBackOff) are excluded to ensure that candidate Pods can receive requests. Furthermore, all healthy Pods that match the tags constitute a candidate set, which serves as candidate containerization units. A specific instance is then selected from this set by the load balancing strategy.

[0067] All the specified containerization units are integrated to obtain the corresponding candidate containerization units.

[0068] In this embodiment, all specified containerization units are integrated, and the resulting set of containerization units is used as the required candidate containerization units.

[0069] This application obtains preset routing rules; maps the model identification information based on the routing rules to obtain corresponding model service groups; then obtains preset filtering conditions; subsequently, based on the model service groups, filters out all specified containerized units that meet the filtering conditions; and then integrates all the specified containerized units to obtain corresponding candidate containerized units. Based on the above processing flow, this application obtains model service groups by mapping model identification information based on routing rules, then filters out all specified containerized units that meet preset filtering conditions based on the model service groups, and then integrates all the specified containerized units. This enables the automatic and accurate determination of candidate containerized units corresponding to model identification information, achieving deterministic and low-latency mapping from model identification information to candidate containerized units, improving the efficiency of candidate containerized unit determination, ensuring the accuracy of the obtained candidate containerized units, and laying a foundation for subsequent load balancing and model inference.

[0070] In some optional implementations of this embodiment, step S204 includes the following steps:

[0071] Obtain the order information of all the candidate containerization units.

[0072] In this embodiment, when the routing component starts up and obtains candidate containerized units (such as Pod-A, Pod-B, and Pod-C) from the target service group, it assigns a fixed order to each Pod (usually sorted by creation time or name), forming an ordered queue: [Pod-A, Pod-B, Pod-C]. Simultaneously, it maintains an atomic counter (initially 0) to record the current polling position.

[0073] Get the preset sequential calling algorithm.

[0074] In this embodiment, the core logic of the above-mentioned sequential calling algorithm is: to allocate requests in sequence according to the preset Pod order, and repeat the process to ensure that the requests are evenly distributed to all candidate Pods.

[0075] The sequential information is analyzed based on the sequential invocation algorithm to select the first containerized unit corresponding to the business processing request from the candidate containerized units.

[0076] In this embodiment, the containerized unit invocation process based on the sequential invocation algorithm includes the following steps: Step 1: Read the counter. Each time a request arrives, the current counter value is read atomically (e.g., current_index = 1). Step 2: Select the target Pod. The corresponding Pod is selected from the sequential list based on the counter value (e.g., index = 1 corresponds to Pod-B). Step 3: Update the counter. The counter is incremented by 1; if it exceeds the list length, it is reset to zero (e.g., 1→2→3→0). Step 4: Handle exceptions. If the selected Pod is unavailable (e.g., health check fails), the Pod is skipped, the counter is incremented, and the next Pod is selected.

[0077] Specifically, the corresponding containerized unit call processing can be executed based on the sequential call algorithm to select the first containerized unit corresponding to the above-mentioned business processing request from the above-mentioned candidate containerized units.

[0078] The example process includes: Initial state: [Pod-A(0),Pod-B(1),Pod-C(2)],current_index=0; Request 1: Select Pod-A, and update the counter to 1; Request 2: Select Pod-B, and update the counter to 2; Request 3: Select Pod-C, and update the counter to 0; Request 4: Select Pod-A, and update the counter to 1, and so on.

[0079] The first containerization unit is used as the target containerization unit.

[0080] In this embodiment, the use of the sequential call algorithm enables deterministic allocation: the request order strictly follows the list polling, which is suitable for stateless services, and ensures thread safety: concurrent contention is avoided through atomic counters.

[0081] This application obtains the sequence information of all candidate containerized units; then obtains a preset sequential invocation algorithm; subsequently, it analyzes the sequence information based on the sequential invocation algorithm to select the first containerized unit corresponding to the business processing request from the candidate containerized units; and then uses the first containerized unit as the target containerized unit. Based on the above processing flow, this application obtains the sequence information of all candidate containerized units and then analyzes the sequence information based on the use of the obtained sequential invocation algorithm, thereby automatically and intelligently selecting the target containerized unit corresponding to the business processing request from the candidate containerized units. This enables efficient containerized unit selection in different scenarios, effectively balancing uniformity, resource utilization, and implementation complexity.

[0082] In some alternative implementations, step S204 includes the following steps:

[0083] Obtain the weight data for all the candidate containerized units.

[0084] In this embodiment, the aforementioned weight data may refer to static weights or dynamic weights.

[0085] Obtain the preset weights and apply the algorithm.

[0086] In this embodiment, the core logic of the weighted invocation algorithm is to dynamically allocate requests based on the preset weight of the Pod or the real-time load, ensuring that high-weight / low-load Pods receive more requests.

[0087] The weight data is analyzed based on the weight invocation algorithm to select a second containerization unit corresponding to the business processing request from the candidate containerization units.

[0088] In this embodiment, the containerized unit invocation process based on the weight invocation algorithm includes a first invocation process corresponding to static weights and a second invocation process corresponding to dynamic weights.

[0089] Specifically, the first call process mentioned above includes: static weight allocation (based on configuration). 1. Weight configuration. Configure static weights for each Pod in the routing table (e.g., Pod-A weight = 5, Pod-B weight = 3, Pod-C weight = 2). Weights usually reflect the processing capabilities of the Pod (e.g., CPU / memory resources). 2. Weight polling table construction. Generate an extended list based on the weights, repeating the Pod names to match the weight values: original list: [Pod-A(5), Pod-B(3), Pod-C(2)], extended list: [Pod-A, Pod-A, Pod-A, Pod-A, Pod-A, Pod-B, Pod-B, Pod-C, Pod-C]. And maintain a circular pointer, initially pointing to the head of the list. 3. Request allocation process. Step 1: Select Pod. Each request selects a Pod from the extended list based on the pointer position (e.g., selecting Pod-A for the first time). Step 2: Move pointer. Move the pointer one position to the right, and reset it to zero if it exceeds the list length. Step 3: Dynamic adjustment. If runtime weight updates are supported, regenerate the expanded list and reset the pointers.

[0090] The second invocation process described above includes: Dynamic weight allocation (based on load). 1. Load metric collection. Regularly collect real-time metrics for Pods (e.g., CPU utilization, active connections, request queue length). Calculate dynamic weights: Reciprocal weighting method: Weight = 1 / current load (higher load, lower weight). Threshold limiting method: If a Pod's load exceeds a threshold (e.g., 80% CPU), temporarily set its weight to 0. 2. Weight normalization. Normalize the dynamic weights of all Pods to integer ratios (e.g., Pod-A = 70%, Pod-B = 30% → weight 7:3). Generate an expanded list (e.g., 7 Pod-A and 3 Pod-B). 3. Request allocation process. Same as the static weighting scheme, but the expanded list is dynamically updated as the load changes. Smooth transition: Avoid sudden weight changes that could cause request skew (e.g., adjust only 10% of the list items each time).

[0091] Example flow (static weights): Configuration: [Pod-A(5),Pod-B(3),Pod-C(2)]; Extended list: [A,A,A,A,A,B,B,B,C,C]; Request assignment: Request 1 → Pod-A (pointer moves); Request 2 → Pod-A...; Request 6 → Pod-B; Request 9 → Pod-C; Request 10 → Pod-A (pointer returns to zero).

[0092] The second containerization unit is used as the target containerization unit.

[0093] In this embodiment, the use of a weight-based call algorithm can improve flexibility: support static configuration or dynamic load awareness, and ensure fairness: high-weight Pods get more requests, but low-weight Pods still have a chance to process them.

[0094] This application obtains the weight data of all candidate containerized units; then obtains a preset weight invocation algorithm; subsequently, it analyzes the weight data based on the weight invocation algorithm to select a second containerized unit corresponding to the business processing request from the candidate containerized units; and finally, it uses the second containerized unit as the target containerized unit. Based on the above processing flow, this application obtains the weight data of all candidate containerized units and then analyzes the weight data based on the use of the obtained weight invocation algorithm, thereby automatically and intelligently selecting the target containerized unit corresponding to the business processing request from the candidate containerized units. This enables efficient containerized unit selection in different scenarios, effectively balancing uniformity, resource utilization, and implementation complexity.

[0095] In some optional implementations, the preprocessing of the customer information to obtain corresponding feature data in step S205 includes the following steps:

[0096] The customer information is standardized to obtain the corresponding first processed information.

[0097] In this embodiment, the above-mentioned format standardization process includes: for vehicle model information, the system will encode it according to a unified classification standard, for example, encoding sedans as "01" and SUVs as "02". For numerical data such as age and vehicle age, the system will perform range validation and format standardization. For example, the system specifies that the vehicle age range is between 0 and 30 years. If the vehicle age entered by the customer exceeds this range, the system will prompt for modification; at the same time, the format of all numerical data will be standardized to integer or two decimal places.

[0098] The first processing information is encoded and converted to obtain the corresponding second processing information.

[0099] In this embodiment, the above-mentioned encoding conversion process includes: for text-type information, such as occupation, marital status, etc., the system will perform classification encoding conversion. For example, occupations are divided into categories such as "corporate employee", "civil servant" and "freelancer", and each is assigned a corresponding code.

[0100] The second processing information is quantized to obtain the corresponding third processing information.

[0101] In this embodiment, the aforementioned quantification process refers to quantifying the traffic violations and claims information in the driving record. Specifically, the system assigns different scores based on the severity of the violation, such as 1 point for speeding and 2 points for running a red light; the number of claims and the amount of claims are normalized and converted into values ​​between 0 and 1 so that the model can better understand and process this data.

[0102] The third processing information is used as the feature data.

[0103] This application standardizes the customer information to obtain corresponding first processed information; then, it encodes and converts the first processed information to obtain corresponding second processed information; subsequently, it quantizes the second processed information to obtain corresponding third processed information; and finally, it uses the third processed information as the feature data. Based on the above processing flow, this application achieves efficient and accurate preprocessing of customer information by standardizing the format, converting the encoding, and quantizing the customer information. This improves the data quality of the obtained feature data, which is beneficial for providing accurate and standardized data input for subsequent model prediction, thereby helping to improve the accuracy of the generated data evaluation results.

[0104] In some optional implementations of this embodiment, step S207 includes the following steps:

[0105] Obtain the preset target format.

[0106] In this embodiment, the target format refers to a specific format corresponding to the requested data format, such as JSON or XML.

[0107] The data evaluation results are converted based on the target format to obtain the corresponding first data evaluation result.

[0108] In this embodiment, the first data evaluation result can be obtained by encapsulating the above data evaluation results into the target format. For example, if the requested data is in JSON format, the model prediction result, i.e., the first data evaluation result, will also be encapsulated in JSON format, including fields such as risk score and expected claim amount.

[0109] The first data evaluation result is validated based on a preset data validation strategy.

[0110] In this embodiment, the data verification strategy includes the following: 1. Result format verification. Structure verification: Check whether the data conforms to a predefined JSON Schema or Protobuf format (e.g., it must contain the prediction_format and confidence_level fields). Value range verification: Verify whether the numerical fields are within a reasonable range (e.g., probability values ​​must be between 0 and 1). Business rule verification: Check the validity of the data according to business logic (e.g., car insurance risk scores cannot be negative). 2. Integrity verification. Hash verification: Append a hash value (e.g., SHA-256) to the returned data. The receiver (model routing component) recalculates the hash and compares it to ensure the data has not been tampered with. Signature verification: If asymmetric encryption is enabled, sign the data using the model service private key. The routing component verifies the authenticity of the signature using the public key. 3. Exception handling. If verification fails, record the error log and return a standardized error response (e.g., HTTP 400 Bad Request + error code NVAL ID_MODEL_OUTPUT). Additionally, forward the failed request to the backup model service group or the degradation processing logic.

[0111] Specifically, data verification of the first data evaluation result can be performed based on the strategy content of the aforementioned data verification strategy, and a corresponding data verification result can be generated. The data verification result includes whether the first data evaluation result passes the data verification or whether the first data evaluation result fails the data verification.

[0112] If the first data evaluation result passes the data verification, the first data evaluation result is encrypted to obtain the corresponding second data evaluation result.

[0113] In this embodiment, the encryption process includes: converting the first data evaluation result into byte stream data (e.g., converting a JSON string to UTF-8 encoding), then encrypting the byte stream data according to a selected encryption protocol, and encapsulating the resulting encrypted data into a standard response body as the second data evaluation result. The encryption protocol can employ transport layer encryption or application layer encryption. Transport layer encryption includes encrypting the entire communication channel using TLS 1.2 / 1.3 (e.g., HTTPS) to ensure automatic encryption of data during transmission between the routing component and the business system. Application layer encryption includes encrypting sensitive fields (e.g., user ID, precise risk score) individually using symmetric encryption algorithms (e.g., AES-256) or asymmetric encryption (e.g., RSA-OAEP). Encryption effectively prevents data from being eavesdropped on or tampered with during transmission, ensuring compliance with security and compliance requirements.

[0114] The evaluation results of the second data are then processed for output.

[0115] In this embodiment, the generated second data evaluation result can be sent to the target customer and the target business personnel corresponding to the target customer by calling the above-mentioned model routing component, thereby completing the output processing of the above-mentioned second data evaluation result.

[0116] This application obtains a preset target format; then, based on the target format, it converts the data evaluation result to obtain a corresponding first data evaluation result; subsequently, it performs data verification on the first data evaluation result based on a preset data verification strategy; if the first data evaluation result passes the data verification, it encrypts the first data evaluation result to obtain a corresponding second data evaluation result; and finally, it outputs the second data evaluation result. Based on the above processing flow, this application obtains a first data evaluation result by converting the generated data evaluation result to a target format, then performs data verification on the first data evaluation result to ensure data integrity and accuracy, and then encrypts the first data evaluation result before outputting the encrypted second data evaluation result. This effectively prevents data leakage and tampering, and effectively improves the accuracy, standardization, intelligence, and security of the output data evaluation result.

[0117] In some optional implementations of this embodiment, after step S207, the electronic device may further perform the following steps:

[0118] Obtain the preset decision rules.

[0119] In this embodiment, the aforementioned decision rule is a pre-built auto insurance pricing strategy based on actual business needs. The strategy includes: setting a risk level and pre-building a correspondence table between risk scores and premium rates; if the data assessment results show that the customer's risk score is 0.7, the system will determine the corresponding premium rate based on the correspondence table.

[0120] Based on the decision rules, the data evaluation results are processed to obtain the corresponding decision data.

[0121] In this embodiment, the decision processing of the above data evaluation results can be performed according to the strategy content of the auto insurance pricing strategy corresponding to the above decision rules, so as to obtain the target premium rate corresponding to the data evaluation results and use it as the above decision data.

[0122] Calls the preset display page.

[0123] In this embodiment, the aforementioned display page is a pre-built page in the system for displaying business data, such as an online insurance application page.

[0124] The decision data is displayed and processed on the display page.

[0125] In this embodiment, the determined decision premium rate, i.e., the target premium rate, can be displayed on the aforementioned display page to show customers or sales personnel, along with detailed rate explanations and descriptions. For example, on the online insurance application page, the premium rate and its composition are clearly displayed so that customers understand how the premium is calculated. Simultaneously, relevant data from the auto insurance pricing process (such as feature data, data evaluation results, and final premium rates) are stored in a database for subsequent auditing and data analysis.

[0126] This application obtains preset decision rules; then processes the data evaluation results based on these rules to obtain corresponding decision data; subsequently, it calls a preset display page; and finally, it displays the decision data on the display page. Based on this process, this application automatically and accurately completes the decision processing of data evaluation results by processing the data evaluation results based on decision rules, thereby improving the efficiency of decision data generation. Furthermore, by displaying the decision data on the display page, it provides accurate business data support for customer business processes, and customers can clearly view the relevant decision data through the display page, which helps improve customer experience and satisfaction.

[0127] In some alternative implementations, the user information obtained is subject to user consent and complies with relevant laws and policies.

[0128] Furthermore, any software tools or components not belonging to our company that appear in the embodiments of this application are merely illustrative examples and do not represent actual use.

[0129] Furthermore, this application enables configurable online management and orchestration of models, and dynamic load balancing routing of models at runtime. Specifically, this is reflected in the following points: 1. Models can be dynamically started directly after training, without requiring application modifications, dynamic loading, and restarts; 2. Different models run on different instances, achieving isolation between model execution and invocation, reducing coupling between models, and improving model stability; 3. A routing mechanism is added to model execution invocation, providing a unified model invocation platform, and routing to the specified model instance to be invoked based on different invocation parameters or model parameters; 4. Dynamically configurable load balancing strategies allow for different invocation strategies to be configured according to actual scenarios, such as random, weighted, or sequential invocation, achieving load balancing configuration for model instance invocation and greatly increasing the scalability and stability of model execution.

[0130] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0131] It should be emphasized that, to further ensure the privacy and security of the above data evaluation results, the data evaluation results can also be stored in a blockchain node.

[0132] The blockchain referred to in this application is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0133] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. 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 that knowledge to obtain optimal results.

[0134] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0135] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware through computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the methods described above. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0136] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0137] Further reference Figure 3 As a response to the above Figure 2 To implement the method shown, this application provides an embodiment of a data processing device based on artificial intelligence, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0138] like Figure 3 As shown, the artificial intelligence-based data processing device 300 described in this embodiment includes: a receiving module 301, an extraction module 302, a determining module 303, a filtering module 304, a processing module 305, an evaluation module 306, and an output module 307. Wherein:

[0139] The receiving module 301 is used to receive a business processing request corresponding to a target customer; wherein the business processing request carries model identification information;

[0140] Extraction module 302 is used to parse the business processing request to extract the corresponding model identification information;

[0141] The determination module 303 is used to determine the candidate containerization unit corresponding to the model identification information based on the preset model routing component;

[0142] The filtering module 304 is used to filter out target containerized units from the candidate containerized units based on a preset load balancing algorithm.

[0143] The processing module 305 is used to acquire the customer information of the target customer and preprocess the customer information to obtain corresponding feature data;

[0144] Evaluation module 306 is used to read the model parameters of the target prediction model corresponding to the model identification information based on the target containerization unit, and to evaluate the feature data based on the target prediction model to obtain the corresponding data evaluation result;

[0145] The output module 307 is used to output the data evaluation results.

[0146] In some optional implementations of this embodiment, the determining module 303 includes:

[0147] The first acquisition submodule is used to acquire preset routing rules;

[0148] The mapping submodule is used to map the model identification information based on the routing rules to obtain the corresponding model service group;

[0149] The second acquisition submodule is used to acquire preset filtering conditions;

[0150] The filtering submodule is used to filter out all specified containerized units that meet the filtering conditions based on the model service group.

[0151] The integration submodule is used to integrate all the specified containerization units to obtain the corresponding candidate containerization units.

[0152] In some optional implementations of this embodiment, the filtering module 304 includes:

[0153] The third acquisition submodule is used to acquire the order information of all the candidate containerized units;

[0154] The fourth submodule is used to obtain the preset sequential calling algorithm;

[0155] The first filtering submodule is used to analyze the sequence information based on the sequential calling algorithm, so as to filter out the first containerized unit corresponding to the business processing request from the candidate containerized units;

[0156] The first determining submodule is used to use the first containerization unit as the target containerization unit.

[0157] In some optional implementations of this embodiment, the filtering module 304 includes:

[0158] The fifth acquisition submodule is used to acquire the weight data of all the candidate containerized units;

[0159] The sixth submodule is used to obtain the preset weight call algorithm;

[0160] The second filtering submodule is used to analyze the weight data based on the weight calling algorithm, so as to filter out the second containerized unit corresponding to the business processing request from the candidate containerized units.

[0161] The second determining submodule is used to use the second containerization unit as the target containerization unit.

[0162] In some optional implementations of this embodiment, the processing module 305 includes:

[0163] The first processing submodule is used to perform format standardization processing on the customer information to obtain corresponding first processing information;

[0164] The second processing submodule is used to perform encoding conversion processing on the first processing information to obtain the corresponding second processing information;

[0165] The third processing submodule is used to quantize the second processing information to obtain the corresponding third processing information;

[0166] The third determining submodule is used to use the third processing information as the feature data.

[0167] In some optional implementations of this embodiment, the output module 307 includes:

[0168] The seventh submodule is used to obtain the preset target format;

[0169] The conversion submodule is used to convert the data evaluation result according to the target format to obtain the corresponding first data evaluation result.

[0170] The verification submodule is used to verify the first data evaluation result based on a preset data verification strategy.

[0171] The encryption submodule is used to encrypt the first data evaluation result if the first data evaluation result passes the data verification, so as to obtain the corresponding second data evaluation result.

[0172] The output submodule is used to process the output of the second data evaluation result.

[0173] In some optional implementations of this embodiment, the artificial intelligence-based data processing device further includes:

[0174] The acquisition module is used to acquire preset decision rules;

[0175] The decision module is used to process the data evaluation results based on the decision rules to obtain corresponding decision data;

[0176] The calling module is used to invoke the preset display page;

[0177] The display module is used to display and process the decision data on the display page.

[0178] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed]. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0179] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected via a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0180] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0181] The memory 41 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as the hard disk or memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Of course, the memory 41 may include both the internal storage unit and its external storage device of the computer device 4. In this embodiment, the memory 41 is typically used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions for data processing methods based on artificial intelligence. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or will be output.

[0182] In some embodiments, the processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions stored in the memory 41 or to process data, for example, to execute computer-readable instructions of the artificial intelligence-based data processing method.

[0183] The network interface 43 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 4 and other electronic devices.

[0184] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based data processing method described above.

[0185] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0186] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.< / name>

Claims

1. A data processing method based on artificial intelligence, characterized in that, Includes the following steps: Receive a business processing request corresponding to a target customer; wherein, the business processing request carries model identification information; The business processing request is parsed to extract the corresponding model identification information; Based on the preset model routing component, candidate containerization units corresponding to the model identification information are determined; The target containerized unit is selected from the candidate containerized units based on a preset load balancing algorithm. Obtain the customer information of the target customer, and preprocess the customer information to obtain corresponding feature data; Based on the target containerization unit, the model parameters of the target prediction model corresponding to the model identification information are read, and the feature data is evaluated based on the target prediction model to obtain the corresponding data evaluation result; The data evaluation results are then processed for output.

2. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of determining the candidate containerization unit corresponding to the model identification information based on the preset model routing component specifically includes: Retrieve preset routing rules; The model identification information is mapped based on the routing rules to obtain the corresponding model service group; Get the preset filter criteria; Based on the model service group, all specified containerized units that meet the filtering conditions are selected. All the specified containerization units are integrated to obtain the corresponding candidate containerization units.

3. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of selecting the target containerized unit from the candidate containerized units based on the preset load balancing algorithm specifically includes: Obtain the order information of all the candidate containerization units; Retrieve the preset sequential calling algorithm; The sequential information is analyzed based on the sequential invocation algorithm to select the first containerized unit corresponding to the business processing request from the candidate containerized units. The first containerization unit is used as the target containerization unit.

4. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of selecting the target containerized unit from the candidate containerized units based on the preset load balancing algorithm specifically includes: Obtain the weight data of all the candidate containerized units; Obtain the preset weights and apply the algorithm; The weight data is analyzed based on the weight invocation algorithm to select a second containerization unit corresponding to the business processing request from the candidate containerization units. The second containerization unit is used as the target containerization unit.

5. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of preprocessing the customer information to obtain corresponding feature data specifically includes: The customer information is standardized to obtain the corresponding first processed information; The first processed information is encoded and converted to obtain the corresponding second processed information; The second processing information is quantized to obtain the corresponding third processing information; The third processing information is used as the feature data.

6. The data processing method based on artificial intelligence according to claim 1, characterized in that, The step of outputting the data evaluation results specifically includes: Obtain the preset target format; Based on the target format, the data evaluation result is converted to obtain the corresponding first data evaluation result; The first data evaluation result is validated based on a preset data validation strategy. If the first data evaluation result passes the data verification, the first data evaluation result is encrypted to obtain the corresponding second data evaluation result. The evaluation results of the second data are then processed for output.

7. The data processing method based on artificial intelligence according to claim 1, characterized in that, Following the step of outputting the data evaluation results, the method further includes: Obtain the preset decision rules; Based on the decision rules, the data evaluation results are processed to obtain corresponding decision data; Call the preset display page; The decision data is displayed and processed on the display page.

8. A data processing device based on artificial intelligence, characterized in that, include: A receiving module is used to receive a business processing request corresponding to a target customer; wherein the business processing request carries model identification information; The extraction module is used to parse the business processing request to extract the corresponding model identification information; The determination module is used to determine the candidate containerization unit corresponding to the model identification information based on a preset model routing component; The filtering module is used to filter out target containerized units from the candidate containerized units based on a preset load balancing algorithm. The processing module is used to acquire customer information of the target customer and preprocess the customer information to obtain corresponding feature data; The evaluation module is used to read the model parameters of the target prediction model corresponding to the model identification information based on the target containerization unit, and to evaluate the feature data based on the target prediction model to obtain the corresponding data evaluation results. The output module is used to process the data evaluation results.

9. A computer device, characterized in that, It includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor executes the computer-readable instructions to implement the steps of the data processing method based on artificial intelligence as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the data processing method based on artificial intelligence as described in any one of claims 1 to 7.

Images