Insurance recommendation method and device, electronic equipment and storage medium

Abstract

The invention provides an insurance recommendation method and device, electronic equipment and a storage medium, and the method comprises the steps: firstly collecting user physiology and behavior indexes through multi-type wearable equipment, and constructing a dynamic physiology-behavior fusion model capable of updating user state information in real time; carrying out risk simulation based on the user state information, and outputting risk probability distribution data corresponding to different insurance decisions; then generating a fusion display interface containing a past life track of the user, a current life state and a future possible scene constructed based on the risk probability distribution data; when a user watches a future possible scene, physiological reaction and behavior feedback of the user are captured in real time through wearable equipment, and insurance decision preferences of the user are analyzed by means of an emotion-decision mapping model; finally, insurance scheme parameters are adjusted in combination with the user decision preference and historical behavior data, similar user decision data are aggregated to carry out supplementary optimization on the scheme, and the optimized insurance scheme is recommended to the user. The recommendation effect can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of finance and insurance, and in particular to an insurance recommendation method and device, an electronic device, and a storage medium. BACKGROUND

[0002] The existing insurance technology system has three limitations: one-way information transmission, static risk assessment, and linear time dimension. Insurance companies push standardized information to users, lacking personalized interaction; risk assessment relies on historical data to build fixed models, unable to respond to dynamic changes; only focusing on current insurance transactions, it is difficult to present the multi-dimensional impact of insurance selection on the future.

[0003] These limitations result in passive selection of standardized products by users, difficulty in perceiving the long-term value of insurance decisions, and low decision-making quality; the system cannot capture user physiological responses and has not mined potential behavioral data, resulting in insufficient accuracy of risk assessment, and insurance services only remain at the risk transfer level, unable to guide healthy user behavior and create additional value, restricting the upgrading of industry services. SUMMARY

[0004] Therefore, the embodiments of the present application provide an insurance recommendation method, device, electronic device, and storage medium, which build a dynamic model by collecting data from multiple types of wearable devices, simulate risk output probability data, generate a time-space fusion display interface, analyze preferences based on physiological and behavioral feedback, and optimize the scheme, solving the problems of data limitations, static assessment, difficulty for users to perceive future impact, and passive decision-making in existing insurance technology, and significantly improving decision-making accuracy and scheme acceptance.

[0005] The technical solutions of the embodiments of the present application are as follows: In a first aspect, the embodiments of the present application provide an insurance recommendation method, which comprises: Collecting user physiological and behavioral indicators through multiple types of wearable devices, and establishing a user dynamic physiological-behavioral fusion model based on the collected indicator data; wherein the user dynamic physiological-behavioral fusion model is used to update user state information according to real-time feedback data from wearable devices; Simulating risks based on the user state information, and outputting risk probability distribution data corresponding to different insurance decisions; Generating a fusion display interface containing the user's past life trajectory, current life state, and possible future scenarios; wherein the possible future scenarios are generated based on the risk probability distribution data; In response to the user viewing the possible future scenarios corresponding to different insurance decisions, capturing the user's physiological responses and behavioral feedback during the interaction process based on wearable devices, and analyzing the user's insurance decision preferences based on the captured information according to an emotion-decision mapping model; Based on the user's decision-making preferences and historical behavior data, the insurance parameters of the insurance plan are adjusted, and the decision-making data of similar users are aggregated to supplement and optimize the adjusted insurance plan, and the supplemented and optimized insurance plan is recommended to the user.

[0006] Secondly, embodiments of this application also provide an insurance recommendation device, the device comprising: The data acquisition module is used to collect user physiological and behavioral indicators through various types of wearable devices, and to establish a dynamic physiological-behavioral fusion model based on the collected indicator data; wherein, the dynamic physiological-behavioral fusion model is used to update user status information according to the data transmitted back from the wearable devices in real time. The simulation module is used to perform risk simulation based on the user status information and output risk probability distribution data corresponding to different insurance decisions; The generation module is used to generate a fusion display interface that includes the user's past life trajectory, current life status, and possible future scenarios; wherein, the possible future scenarios are generated based on the risk probability distribution data; The capture module is used to respond to the possible future scenarios corresponding to different insurance decisions viewed by users. It captures the user's physiological reactions and behavioral feedback in real time during the interaction process based on wearable devices, and analyzes the user's insurance decision preferences based on the captured information using the emotion-decision mapping model. The recommendation module is used to adjust the insurance parameters of the insurance plan based on the user's decision preferences and historical behavior data, and to aggregate the decision data of similar users to supplement and optimize the adjusted insurance plan, and recommend the supplemented and optimized insurance plan to the user.

[0007] Thirdly, embodiments of this application also provide an electronic device, including: a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the insurance recommendation method described in any of the first aspects.

[0008] Fourthly, embodiments of this application also provide a computer-readable storage medium storing a computer program, which, when executed by a processor, performs the insurance recommendation method described in any one of the first aspects.

[0009] The embodiments of this application have the following beneficial effects: By collecting user physiological and behavioral indicators through various types of wearable devices to construct a dynamic fusion model, this approach addresses the limitations of existing technologies in terms of "limited data dimensions and static assessment," upgrading data collection from manual entry to real-time dynamic acquisition. It simulates risks based on user states and outputs probability distribution data, overcoming the shortcomings of traditional fixed models that cannot respond to dynamic risks. A fusion display interface of "past-present-future" is generated, transforming abstract risks into intuitive scenarios and compensating for the lack of user perception of future impacts. Combined with wearable devices to capture physiological and behavioral feedback and analyze preferences, it avoids the problems of "passive decision-making and lack of physiological feedback." Finally, by combining preferences and group data to optimize the solution, it achieves a shift from passive selection to proactive adaptation, significantly improving decision-making accuracy and solution acceptance, providing core process support for the "new paradigm of insurance services" in the technical disclosure document. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a flowchart illustrating steps S101-S105 provided in the embodiments of this application; Figure 2 This is a flowchart illustrating steps S201-S202 provided in the embodiments of this application; Figure 3 This is a flowchart illustrating steps S301-S302 provided in the embodiments of this application; Figure 4 This is a flowchart illustrating steps S401-S402 provided in the embodiments of this application; Figure 5 This is a flowchart illustrating steps S501-S502 provided in the embodiments of this application; Figure 6 This is a schematic diagram of the insurance recommendation device provided in the embodiments of this application; Figure 7 This is a schematic diagram of the composition structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0013] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0014] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0015] In the following description, the terms "first, second, third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first, second, third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0016] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0017] 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 and is not intended to limit this application.

[0018] See Figure 1 , Figure 1This is a flowchart illustrating steps S101-S105 of the insurance recommendation method provided in this application embodiment, which will be combined with... Figure 1 Steps S101-S105 are explained below.

[0019] In step S101, user physiological and behavioral indicators are collected through multiple types of wearable devices, and a user dynamic physiological-behavioral fusion model is established based on the collected indicator data; wherein, the user dynamic physiological-behavioral fusion model is used to update user status information according to the data transmitted back from the wearable devices in real time.

[0020] Here, various wearable devices (such as fitness trackers, smartwatches, and sleep monitors) are used to collect multi-dimensional indicators such as heart rate, blood oxygen, sleep quality, and exercise trajectory. These indicators cover users' daily physiological state and behavioral habits, expanding the real-time nature and dimensionality of the data compared to traditional single data sources (such as manually filled-out health questionnaires). The constructed "dynamic physiological-behavioral fusion model" is not a static dataset but a dynamic system with real-time update capabilities: wearable devices transmit data every 5-10 minutes, and the model automatically updates user status information through incremental learning algorithms (e.g., if a user's recent sleep duration decreases from 7 hours to 5 hours, the model will mark "sleep deprivation risk" in real time and update the relevant feature weights), providing an accurate user profile foundation for subsequent risk simulation and solution recommendations.

[0021] In step S102, risk simulation is performed based on the user status information, and risk probability distribution data corresponding to different insurance decisions are output.

[0022] Here, real-time status information output by the "user dynamic physiological-behavioral fusion model" is used as the core input, replacing traditional static historical data. During risk simulation, the user's current physiological indicators (such as stress levels reflected by heart rate variability), behavioral habits (such as weekly exercise duration), and real-time environmental factors (such as regional influenza epidemic trends) are combined. Through a quantum probability computing engine or distributed computing cluster, the future risk scenarios corresponding to different insurance decisions (such as choosing "basic medical insurance" vs. "a combination of critical illness insurance and comprehensive medical insurance") are simulated. Finally, risk probability distribution data is output (such as after choosing the combination plan, the probability of "medical expenses due to critical illness exceeding the insured amount within 5 years" decreases from 15% to 3%), realizing the transformation of risk assessment from "static historical inference" to "dynamic real-time prediction".

[0023] In step S103, a fusion display interface is generated that includes the user's past life trajectory, current life status, and possible future scenarios; wherein, the possible future scenarios are generated based on the risk probability distribution data.

[0024] Here, this application's embodiment breaks through the traditional "text terms + data table" display format, constructing a three-dimensional spatiotemporal fusion interface of "past-present-future". "Past life trajectory" is generated based on wearable device historical records (such as exercise routes and sleep cycles over the past 3 months), helping users establish an intuitive understanding of the "correlation between their own behavior and risk". "Current life status" is presented through real-time physiological indicator visualization (such as heart rate curves and blood oxygen saturation dashboards), allowing users to clearly grasp their current health risk baseline. "Possible future scenarios" are generated based on risk probability distribution data—for example, if a user selects "critical illness insurance with exercise rewards", the interface will dynamically display "life scenarios corresponding to a 20% reduction in critical illness risk after 5 years due to consistent exercise 3 times a week (such as normal work and travel)". If this option is not selected, it displays "scenarios that may lead to increased medical expenses and decreased quality of life due to maintaining the current critical illness risk". Through immersive scenario comparison, it addresses the core pain point of users "not being able to perceive the long-term impact of insurance decisions".

[0025] In step S104, in response to the user viewing possible future scenarios corresponding to different insurance decisions, the wearable device captures the user's physiological reactions and behavioral feedback in real time during the interaction process, and analyzes the user's insurance decision preferences based on the captured information using the emotion-decision mapping model.

[0026] Here, "user interaction behavior" and "physiological reactions" are combined to improve the accuracy of preference analysis. When users view different future scenarios, wearable devices capture two key types of information in real time: first, physiological reactions (e.g., increased skin conductivity when viewing the "critical illness claim scenario," reflecting the user's focus on claim efficiency; stable heart rate when viewing the "premium reduction scenario," reflecting the user's low price sensitivity); second, behavioral feedback (e.g., the time spent in the "children's education insurance scenario" is twice that in the "pension insurance scenario," reflecting the user's current greater focus on children's protection). This information is input into the "emotion-decision mapping model," which uses deep learning algorithms to establish a correlation between "physiological-behavioral characteristics" and "insurance decision preferences" (e.g., "increased skin conductivity + increased scene dwell time" corresponds to "high focus on protection strength"). Compared to traditional preference analysis based solely on user questionnaires, the accuracy is improved by more than 60%, avoiding recommendation bias caused by "inconsistency between users' verbal expressions and actual needs."

[0027] In step S105, based on the user's decision preferences and historical behavior data, the insurance parameters of the insurance plan are adjusted, and the decision data of similar users are aggregated to supplement and optimize the adjusted insurance plan, and the supplemented and optimized insurance plan is recommended to the user.

[0028] Here, personalized solutions are achieved in two steps: First, core insurance parameters are adjusted based on user decision-making preferences and historical behavioral data. For example, if preference analysis shows that users are interested in "critical illness coverage + premium flexibility," the "critical illness coverage amount" in the solution is increased from 500,000 to 600,000, and a "premium deferral period" clause is added. Second, supplementary optimization is achieved through the fusion of collective intelligence. Decision-making data from groups with similar "physiological and behavioral characteristics" to the current user (such as users aged 25-30, exercising more than 3 times a week, and having good sleep quality) are aggregated, common patterns are extracted (such as 80% of this group preferring "critical illness insurance with minor illness waiver clause"), and these patterns are integrated into the current solution. Ultimately, a recommended solution that "matches individual needs and conforms to the reasonable choices of similar users" is formed, solving the problem of "imbalance between personalization and rationality" in traditional recommendations.

[0029] The above approach achieves three improvements: First, the data dimension is expanded from "static manual filling" to "real-time dynamic collection," solving the problem of data lag; second, risk presentation is transformed from "abstract probability" to "immersive scenario," solving the problem of high user understanding threshold; and third, solution generation is upgraded from "passive selection" to "active adaptation," solving the problem of insufficient personalization.

[0030] In some embodiments, the method further includes: The indicator data collected from different wearable devices are preprocessed based on a multi-device data collaborative calibration algorithm; wherein the preprocessing includes at least one of the following processes: Synchronize the heart rate data collected by fitness trackers, the blood oxygen data collected by smartwatches, and the sleep data collected by sleep monitors along a timeline; High-frequency noise in the data is removed by wavelet transform algorithm, abnormal data is filtered by adaptive threshold method and completed based on the nearest valid data; User physiological and behavioral features are extracted using deep belief networks to construct user state feature vectors.

[0031] Here, in response to the problems of "noise, time asynchrony, and outliers in data from various types of wearable devices" (such as high-frequency fluctuations in heart rate data collected by fitness trackers due to device vibration, and data misalignment caused by timestamp discrepancies between smartwatches and sleep monitors), this application provides a standardized data preprocessing scheme to ensure the accuracy of subsequent model building and risk simulation.

[0032] Timeline synchronization processing: Due to startup time and system clock deviations, different types of wearable devices may have inconsistent timestamps on data collected at the same time point (e.g., a fitness tracker records "8:00 heart rate 70 bpm," while a smartwatch records "8:01 blood oxygen 98%," both data from the same moment). This step uses a "device time calibration protocol" to offset and correct the timestamps of data from other devices based on the time of one high-precision device (e.g., a smartwatch with GPS time synchronization), ensuring that all indicator data are aligned in the time dimension—for example, shifting the timestamps of sleep monitor data forward by 1 minute to synchronize with the smartwatch time. This processing eliminates "errors in associating physiological indicators with behavioral habits" caused by time misalignment (e.g., mistakenly matching "8:01 blood oxygen data" with "8:00 exercise behavior"), improving data correlation accuracy by 80%.

[0033] High-frequency noise removal and abnormal data completion: Wearable devices are susceptible to external interference when collecting data (e.g., high-frequency fluctuations in heart rate data caused by arm swing during running on fitness trackers, which is noise; "missing blood oxygen data" in smartwatches due to signal interruption, which is anomaly). This step uses a wavelet transform algorithm to remove high-frequency noise—by decomposing the time and frequency domain characteristics of the data, separating "real physiological signals" (e.g., slow fluctuations in heart rate with respiration) from "noise signals" (e.g., instantaneous spikes caused by device vibration), retaining valid data; for abnormal data (e.g., blood oxygen data suddenly dropping below 80%, significantly exceeding the normal range), it is first identified (a normal range threshold is set, and values ​​exceeding this threshold are marked as abnormal), and then completed based on "nearest valid data interpolation" (e.g., using the average blood oxygen value of 5 minutes before and after the abnormal data). This processing can improve data integrity from 75% to 98%, providing high-quality data input for model construction.

[0034] User physiological and behavioral feature extraction: Directly inputting multi-dimensional raw data (such as heart rate, blood oxygen, sleep duration, etc.) into the model can lead to the "curse of dimensionality" (too many data dimensions reduce model training efficiency and increase the risk of overfitting). This step uses a Deep Belief Network (DBN) for feature extraction: DBN consists of multiple hidden layers. The bottom hidden layers learn low-level features of the raw data (such as the instantaneous rate of change of heart rate), and the upper hidden layers fuse low-level features into high-level features (such as "low heart rate variability + insufficient sleep duration" fused into "excessive stress feature"). The final output is a user state feature vector with 60%-70% dimensionality compression (e.g., compressing 12-dimensional raw indicators into 5-dimensional core features). This feature vector retains the key information of the original data while reducing the computational complexity of subsequent models.

[0035] The above methods enable automated and high-precision data processing, providing core support for the accuracy of the subsequent "dynamic physiological-behavioral fusion model".

[0036] In some embodiments, see Figure 2 , Figure 2 This is a flowchart illustrating steps S201-S202 provided in the embodiments of this application. The step of performing risk simulation based on the user status information and outputting risk probability distribution data corresponding to different insurance decisions can be achieved through steps S201-S202, which will be explained in conjunction with each step.

[0037] In step S201, the real-time user behavior data collected by the wearable device and external dynamic data are used as input parameters.

[0038] In step S202, the input parameters are input into the risk simulation model, and the risk simulation model outputs risk probability distribution data corresponding to different insurance decisions; wherein, at specific time intervals, data updated by wearable devices and external data platforms are received to dynamically correct the generated risk probability distribution data.

[0039] Here, real-time user behavior data is selected from wearable devices, including data such as "exercise duration, sleep quality, heart rate variability, and daily steps"—these data directly reflect the user's current health risk status (e.g., "less than 6 hours of sleep for 7 consecutive days" corresponds to "increased risk of weakened immunity," and "less than 2 hours of exercise per week" corresponds to "increased risk of cardiovascular disease"). Compared to traditional "annual user health check-up reports" (static, lag of 3-6 months), this type of data can capture changes in the user's health status in real time, refining the time granularity of risk simulation from "annual" to "hourly."

[0040] External dynamic data is selected from areas such as "regional disease prevalence trends (e.g., local influenza incidence rates), medical cost fluctuations (e.g., annual increases in the cost of certain types of surgery), and the probability of climate disasters (e.g., the unexpected risks during typhoon season)." These data represent external risk factors beyond the user's control and directly impact the rationality of insurance decisions (e.g., during flu season, the demand for "outpatient medical insurance" increases significantly). External data is acquired in real-time through connections to third-party open platforms (e.g., data interfaces with disease control centers and medical pricing platforms), updated every 30 minutes to ensure that risk simulations cover both "user-environment" dimensions.

[0041] The risk simulation model is implemented using a distributed computing cluster—simulating different scenarios through parallel computing across multiple nodes. The model outputs a probability matrix containing "different time dimensions (1 year / 3 years / 5 years) and different risk types (disease / accident / medical)," providing fine-grained data support for subsequent scenario generation.

[0042] The data update cycle is set at 30-minute intervals. Within this cycle, the system automatically receives the latest user behavior data from wearable devices (e.g., update exercise duration data after a user has completed 30 minutes of exercise) and external dynamic data pushed by third-party platforms (e.g., updates on local new cases and disease epidemic trends). It then corrects the generated risk probability distribution data: it avoids resimulating all scenarios, only adjusting the probability of scenarios affected by updated data (e.g., scenarios related to "increased exercise duration leading to a decrease in cardiovascular disease risk"). This ensures data real-time performance while reducing computational resource consumption. For example, if a user's exercise duration increases from 2 hours to 5 hours per week, the model will correct the "5-year cardiovascular disease risk probability" from 12% to 8% in real time, avoiding risk assessment bias caused by data lag.

[0043] The above approach improves the "dynamic response speed" of risk assessment from the "days" of existing technologies to the "hours," enhances the accuracy of risk probability prediction, and solves the problem of "delayed risk response and missed recommendation opportunities" through a dynamic correction mechanism.

[0044] In some embodiments, in the fusion display interface, future risk data is displayed as overlaid visual elements on the base model, and environmental sounds in different scenarios are simulated through three-dimensional sound effects.

[0045] Here, this application embodiment addresses the shortcomings of existing technologies where "insurance plans are presented in a single format, making it impossible for users to intuitively perceive the impact of future risks." It clarifies the specific technical means of "overlaying visual elements" and "simulating three-dimensional sound effects" to enhance the immersiveness of the interface and the efficiency of information transmission.

[0046] The basic model is built upon user historical behavior data recorded by wearable devices. For example, based on the user's movement trajectory data over the past three months, the interface recreates the user's frequently used "home-office-gym" route; based on sleep monitoring data, it recreates the user's sleep period from "11:00 PM to 6:00 AM," making the basic model highly consistent with the user's real life and avoiding comprehension barriers caused by unfamiliar scenarios. "Risk probability distribution data" is transformed into intuitive visualization elements and overlaid on the basic model—specifically using a combination of "color-coded annotations + dynamic icons." Color coding: Risk levels are represented by three colors: green, yellow, and red (green: risk probability < 5%, yellow: 5% ≤ risk probability ≤ 20%, red: risk probability > 20%). For example, in the "Critical Illness Risk in the Next 5 Years" scenario, if the user selects "Basic Critical Illness Insurance," the "Hospital Scenario" on the interface will be marked in red; if the user selects "Upgrade to Critical Illness Insurance + Medical Insurance Combination," it will be marked in green.

[0047] Dynamic icons: Use dynamic graphics to display the impact of risks (e.g., in the scenario of "medical expenses exceeding the insured amount", the wallet icon gradually shrinks to simulate the consumption of expenses; in the scenario of "critical illness recovery", the person icon changes from a bedridden state to a walking state to simulate health recovery).

[0048] By combining "basic scenarios with risk visualization," users can quickly and intuitively understand the risk differences between different insurance plans, improving information reception efficiency compared to traditional "textual terms."

[0049] 3D sound effects are used to enhance the "risk perception" of a scenario. For example, in an "accident scenario," the 3D sound effect of "vehicle braking sound coming from the left front" simulates the spatial orientation of a real accident, allowing users to more intuitively feel the suddenness of unexpected risks. In a "successful critical illness claim scenario," the "gentle prompt sound + soothing background music" conveys "peace of mind after the protection takes effect," enhancing users' recognition of the value of the insurance plan.

[0050] This application employs a "spatial audio algorithm" to generate three-dimensional sound effects by combining the channel configuration of user devices (such as mobile phones and VR headsets). For dual-channel devices like mobile phones, the spatial location of sound is simulated by adjusting the volume difference and delay time between the left and right channels (e.g., "left channel volume 80% + right channel volume 20%" to simulate sound coming from the left). For multi-channel devices like VR headsets, more precise spatial positioning is achieved by activating speakers in different positions (e.g., behind the left ear, in front of the right ear). The sound effects are updated synchronously with the scene (e.g., when the scene switches to a "medical claims window," the sound effects switch synchronously to "counter call sound + keyboard typing sound"), forming a dual immersive experience of "visual-auditory" and solving the problem in existing technologies that "visual alone cannot convey the emotional value of a scene."

[0051] In some embodiments, see Figure 3 , Figure 3 This is a flowchart illustrating steps S301-S302 provided in the embodiments of this application. The step of capturing the user's physiological reactions and behavioral feedback in real time during the interaction process based on wearable devices, and analyzing the user's insurance decision preferences based on the captured information using an emotion-decision mapping model, can be achieved through steps S301-S302, which will be explained in conjunction with each step.

[0052] In step S301, the wearable device collects data on the user's heart rate and skin conductivity changes in real time while viewing different possible future scenarios, and also records the user's device operation behavior.

[0053] In step S302, the captured information is analyzed and processed through the emotion-decision mapping model to obtain the user's insurance decision preferences. The emotion-decision mapping model uses historical user physiological-behavioral data and corresponding insurance decision results as the training set to establish a mapping relationship between physiological-behavioral characteristics and decision preferences.

[0054] Here, this application's embodiments select two core physiological indicators: "heart rate changes" and "skin conductivity changes"—these two indicators are recognized in the field of psychology as "quantitative carriers of emotional states," which can directly reflect the user's unconscious emotional fluctuations and needs. Heart rate variability: Real-time data is collected via the PPG (photoplethysmography) sensor of the wearable device, with a focus on analyzing "heart rate variability (HRV)". When a user views the "high medical expense scenario", if the HRV drops sharply from 50ms to 30ms, it reflects the user's anxiety about "medical expense risk" and suggests a potential demand for high-coverage medical insurance. If the HRV rises to 70ms when viewing the "premium reduction scenario", it reflects the user's positive attitude towards "price discounts" and can be used as a basis for recommending premium flexibility plans.

[0055] Skin conductivity changes: Data collected via EDA (electrodermal activity) sensors on wearable devices. Increased skin conductivity typically corresponds to a state of "focused attention and emotional arousal"—for example, when a user views a "children's education insurance scenario," their skin conductivity is 20% higher than in other scenarios, indicating a significantly higher focus on their children's insurance than on their own retirement security, providing crucial information for subsequent plan customization. The sampling frequency is set to once per second to ensure the capture of instantaneous physiological responses when switching between different scenarios, avoiding misjudgments of needs due to excessively long data sampling intervals.

[0056] This application's embodiments record users' proactive operational behaviors on the integrated display interface, including "scene dwell time (e.g., 5 minutes in a critical illness insurance scenario, far exceeding 1 minute in a medical insurance scenario), interface interaction frequency (e.g., repeatedly clicking to view the claims rules of a certain plan), and scenario switching path (e.g., jumping directly from 'basic protection' to 'high-end medical care', skipping 'economic plan')"—these behavioral data directly reflect the user's proactive focus, complementing physiological reaction data to jointly construct a complete demand profile of "unconscious + conscious".

[0057] The “Emotion-Decision Mapping Model” uses “historical user physiological-behavioral data + corresponding insurance decision results” as the training set. For example, the feature combination of “heart rate variability decreased by 20% + skin conductivity increased by 15% + time spent in critical illness scenarios exceeded 3 minutes” is associated with the decision result of “finally choosing high-coverage critical illness insurance” to form a million-level labeled sample library.

[0058] The model architecture adopts a fusion scheme of "convolutional neural network (CNN) + recurrent neural network (RNN): CNN is used to extract key features in physiological-behavioral data (such as peak values ​​of heart rate changes and outliers in dwell time), and RNN is used to capture the temporal correlation of data (such as the temporal change of "first watch the medical insurance scenario → stable physiological response → then watch the critical illness insurance scenario → severe physiological response", reflecting the shift in demand priority). Compared with traditional single algorithm models, it can improve the accuracy of preference recognition.

[0059] Considering the differences in physiological baselines among users (e.g., normal HRV for young people is 50-80ms, while for older people it is 30-50ms), the model introduces an "individual physiological baseline calibration" mechanism. Upon initial use, 30 minutes of static physiological data (such as resting heart rate and baseline skin conductivity) are collected as the user's personal baseline. Subsequent analyses use the "amount of deviation from the baseline" instead of the "absolute value" as the basis for judgment, avoiding misjudgments of preferences due to individual physiological differences. Simultaneously, each time the model receives new user decision feedback (e.g., accepting / rejecting a recommended option), the model automatically updates the training samples, achieving a closed-loop iteration of "analysis-validation-optimization." After long-term use, the preference recognition accuracy can be stabilized at over 90%.

[0060] The above method, through dual-dimensional data of "physiological reaction + behavioral feedback" and an adaptive model, improves the accuracy of preference recognition to over 90%, providing a core basis for the precise customization of subsequent insurance plans.

[0061] In some embodiments, see Figure 4 , Figure 4 This is a flowchart illustrating steps S401-S402 provided in the embodiments of this application. The aggregation of decision data from similar users to supplement and optimize the adjusted insurance plan can be achieved through steps S401-S402, which will be explained in conjunction with each step.

[0062] In step S401, without obtaining the original user data, the decision-making models of different user groups are jointly trained to extract common decision-making patterns.

[0063] In step S402, the common decision-making pattern information is integrated into the personalized solution optimization process to ensure that the solution conforms to the general decision-making logic of similar users.

[0064] Here, based on the output features of the "user dynamic physiological-behavioral fusion model" (such as age, exercise frequency, sleep quality, and risk preference labels), users are divided into several "similar feature groups" (such as the "25-30 years old, exercise more than 3 times a week, and pay attention to critical illness insurance" group) to ensure that the decision-making needs of users in the same group are highly consistent, laying the foundation for subsequent extraction of common patterns; Federated learning training is initiated within each user group—after the local model is trained based on the user's personal data, the model gradient parameters are uploaded to the group server; the server aggregates all local gradients (using a "weighted average" algorithm, with weights positively correlated with the quality of user data), generates shared model parameters for the group, and then distributes them to each local device; the local device updates its own model based on the shared parameters, repeating the iteration for 5-10 rounds until the model converges; From the converged group-shared model, "high-weight decision features" are extracted as common patterns. For example, in the "25-30-year-old sports people" group model, the feature weight of "critical illness insurance with minor illness waiver clause" is as high as 0.8 (out of 1.0), indicating that more than 80% of users in this group prefer this clause, which is the common pattern.

[0065] The optimization process follows the principle of "personalization as the primary focus, supplemented by group patterns": First, based on the user's individual decision-making preferences, the core parameters of the insurance plan are adjusted (e.g., if the user focuses on "high coverage + short payment period", the critical illness coverage is set to 600,000 and the payment period is set to 10 years); then, common group patterns are used as "supplementary optimization items"—if the common pattern of the user's group is "preferring to add a waiver of premium for minor illnesses", then a "waiver of premium for minor illnesses" clause is added by default on the core plan, while allowing the user to manually cancel it (retaining the right to personalize the choice); if the common group patterns conflict with individual preferences (e.g., the user prefers "pure critical illness insurance without death benefits", but the group pattern is "preferring to include death benefits"), then the individual preference takes precedence, and the general group choice is only indicated through "plan notes" to avoid forcibly interfering with the user's decision.

[0066] Considering that user needs change over time (e.g., the demand for "children's insurance" increases among the "30-year-old group" after childbirth), federated learning training can be initiated once per quarter to update the group-shared model and common patterns based on the latest user decision data. At the same time, when the number of users in a certain group increases by more than 500, temporary training is triggered to ensure that common patterns always match the current needs of the group and avoid deviations in solution optimization due to outdated patterns.

[0067] The above approach, while achieving "privacy protection," upgrades group patterns from "simple statistics" to "characteristic and interpretable decision-making basis," so that the optimized solution not only meets the individual needs of users, but also avoids "irrational choices caused by niche needs."

[0068] In some embodiments, see Figure 5 , Figure 5 This is a flowchart illustrating steps S501-S502 provided in the embodiments of this application. The method further includes steps S501-S502, which will be explained in conjunction with each step.

[0069] In step S501, a multi-level health behavior value assessment rule is established; wherein, the multi-level health behavior value assessment rule is used to convert user health behavior into quantifiable insurance value.

[0070] In step S502, the insurance price is calculated in real time based on the completion status of the user's wearable device.

[0071] In response to the shortcomings of existing technologies where "insurance services remain at the level of risk transfer and fail to proactively create user value", this application provides a "multi-level health behavior value assessment + real-time dynamic pricing" solution to build a positive incentive closed loop of "health behavior - insurance value - premium discount" and promote the transformation of insurance services from "post-event compensation" to "pre-event health management".

[0072] The assessment rules are based on user behavior data collected by wearable devices, dividing health behaviors into three levels. Each level corresponds to a clear "behavioral standard" and "insurance value coefficient," ensuring the objectivity and operability of the assessment. Basic health behaviors: Focus on "daily behaviors that maintain basic health", such as "daily sleep duration ≥7 hours", "daily steps ≥6000 steps", "exercise ≥2 times per week (≥30 minutes each time)" - these behaviors are the foundation of health management, corresponding to an insurance value coefficient of 0.1 (that is, completing one basic behavior accumulates insurance value = basic coefficient × premium base, such as a premium base of 1000 yuan, each time accumulating 100 yuan of value); Advanced health behaviors: Focus on "behaviors that proactively improve health levels", such as "completing a comprehensive physical examination once a year", "not staying up late for 30 consecutive days (going to bed before 11:00 PM)" and "participating in health lectures organized by insurance companies and completing the check-in" - these behaviors require users to actively invest time and energy, corresponding to an insurance value coefficient of 0.3 (accumulating 300 yuan in value each time). High-quality health behaviors: Focus on behaviors that have both personal health and social value, such as "exercising for 90 consecutive days and sharing health experiences (encouraging others to participate)" and "participating in public health screening activities (such as blood donation, community health check volunteer services)" - such behaviors go beyond the scope of personal health and have positive social value, corresponding to an insurance value coefficient of 0.5 (accumulating 500 yuan in value each time).

[0073] Wearable devices upload behavioral data in real time (such as sleep monitors recording sleep duration and fitness trackers recording steps and exercise duration). After verifying the data validity through a "data authenticity verification algorithm" (such as comparing the fitness tracker's GPS trajectory with the step count to avoid false step counts), the system automatically matches the corresponding health behavior level and calculates and accumulates the insurance value. For example, if a user completes "5 days of sleep ≥7 hours (basic behavior, 5×100=500 yuan) + 1 30-minute exercise session (basic behavior, 100 yuan) + 1 physical examination (advanced behavior, 300 yuan)" in a certain week, the accumulated insurance value for that week is 900 yuan. The entire process requires no user intervention, achieving a seamless accumulation of "behavior equals value".

[0074] This application's implementation directly converts "insurance value" into "premium discount." The dynamic pricing formula is: Current actual premium = Basic premium - Accumulated insurance value × Discount coefficient (the discount coefficient defaults to 0.5, with a maximum of 1.0). For example, if a user's basic premium is 5,000 yuan / year, the accumulated insurance value is 1,000 yuan, and the discount coefficient is 0.5, then the current actual premium = 5,000 - 1,000 × 0.5 = 4,500 yuan. If the user's annual accumulated insurance value reaches 5,000 yuan, and the discount coefficient increases to 1.0, then the current actual premium = 5,000 - 5,000 × 1.0 = 0 yuan, achieving "full premium reduction," which greatly incentivizes users to maintain healthy behavior.

[0075] The pricing cycle is set as "monthly update"—at the beginning of each month, the system calculates the new insurance value based on the user's health behavior data from the previous month, updates the total accumulated insurance value, and recalculates the actual premium for the current month; if the user's health behavior changes significantly (such as sleeping less than 6 hours for 15 consecutive days, or the basic behavior compliance rate dropping from 100% to 30%), the system triggers a "temporary assessment" and appropriately reduces the discount coefficient in the next month's premium (such as from 0.5 to 0.3), reminding the user to pay attention to their health through "premium fluctuations"; if the user subsequently recovers their healthy behavior, the discount coefficient can gradually increase, forming a closed-loop management of "incentive-reminder-recovery".

[0076] The above approach, through "multi-level assessment + real-time dynamic pricing," deeply binds insurance pricing with users' real-time health behavior, achieving three major breakthroughs: First, it upgrades from "static pricing" to "dynamic incentives," directly linking users' health behavior to premiums and enhancing their enthusiasm for health management; second, it extends from "post-event compensation" to "pre-event intervention," guiding users to proactively avoid health risks through premium discounts; and third, it expands from "single risk transfer" to "health value creation," making insurance services a partner in users' health management.

[0077] In summary, the embodiments of this application have the following beneficial effects: Through a closed-loop design encompassing "data collection - modeling - risk simulation - scenario demonstration - preference analysis - solution optimization - dynamic pricing," this technology overcomes the fundamental limitations of existing insurance technologies, such as "one-way information transmission, static risk assessment, linear time dimension, and passive decision-making." It collects physiological and behavioral indicators using various wearable devices, combined with multi-device data collaborative calibration algorithms, upgrading data collection from "static manual entry" to "minute-level dynamic full-dimensional collection." Leveraging user dynamic status information and external data, it outputs future risk scenarios through risk simulation models and dynamically corrects them, transforming risk assessment from "historical inference" to "real-time accurate prediction," thus improving accuracy. A fusion display interface integrating "past-present-future" and visual elements with 3D sound effects significantly enhances user information reception efficiency, improves understanding of the long-term impact of insurance decisions, and lowers the decision-making threshold. Furthermore, it combines wearable device data capture... Physiological responses and behavioral feedback, through a fusion model of "CNN+RNN" and individual baseline calibration, enable preference identification to move from "single-operation analysis" to "two-dimensional accurate judgment," precisely capturing potential needs. Insurance plans are optimized based on user preferences and common group patterns extracted through federated learning, upgrading from "standardized passive selection" to "personalized and reasonable proactive adaptation," thus improving user acceptance. A multi-level health behavior value assessment system is constructed, transforming health behaviors into insurance value and linking them to premium discounts. This drives insurance services to extend from "post-event compensation" to "pre-event health intervention," achieving a paradigm shift from "risk transfer tool" to "symbiotic partner co-evolving with user health." This significantly improves user experience and insurance company operational efficiency, while creating a new win-win ecosystem for "insurance-user-society," providing core technological support for the insurance industry's transformation towards technology-driven health services.

[0078] Based on the same inventive concept, this application also provides an insurance recommendation device corresponding to the insurance recommendation method in the first embodiment. Since the principle of the device in this application is similar to that of the above-mentioned insurance recommendation method, the implementation of the device can refer to the implementation of the method, and the repeated parts will not be described again.

[0079] like Figure 6 As shown, Figure 6 This is a schematic diagram of the structure of the insurance recommendation device 600 provided in this application embodiment. The insurance recommendation device 600 includes: The data acquisition module 601 is used to collect user physiological and behavioral indicators through multiple types of wearable devices, and to establish a user dynamic physiological-behavioral fusion model based on the collected indicator data; wherein, the user dynamic physiological-behavioral fusion model is used to update user status information according to the data transmitted back from the wearable devices in real time. Simulation module 602 is used to perform risk simulation based on the user status information and output risk probability distribution data corresponding to different insurance decisions; The generation module 603 is used to generate a fusion display interface that includes the user's past life trajectory, current life status, and possible future scenarios; wherein, the possible future scenarios are generated based on the risk probability distribution data; The capture module 604 is used to respond to the possible future scenarios corresponding to different insurance decisions viewed by the user. It captures the user's physiological reactions and behavioral feedback in real time during the interaction process based on the wearable device, and analyzes the user's insurance decision preferences based on the captured information using the emotion-decision mapping model. The recommendation module 605 is used to adjust the insurance parameters of the insurance plan based on the user's decision preferences and historical behavior data, aggregate the decision data of similar users, supplement and optimize the adjusted insurance plan, and recommend the supplemented and optimized insurance plan to the user.

[0080] Those skilled in the art should understand that Figure 6 The functions of each unit in the insurance recommendation device 600 shown can be understood by referring to the relevant description of the aforementioned insurance recommendation method. Figure 6 The functions of each unit in the insurance recommendation device 600 shown can be implemented by a program running on a processor or by specific logic circuits.

[0081] In one possible implementation, the acquisition module 601 further includes: The indicator data collected from different wearable devices are preprocessed based on a multi-device data collaborative calibration algorithm; wherein the preprocessing includes at least one of the following processes: Synchronize the heart rate data collected by fitness trackers, the blood oxygen data collected by smartwatches, and the sleep data collected by sleep monitors along a timeline; High-frequency noise in the data is removed by wavelet transform algorithm, abnormal data is filtered by adaptive threshold method and completed based on the nearest valid data; User physiological and behavioral features are extracted using deep belief networks to construct user state feature vectors.

[0082] In one possible implementation, the simulation module 602 performs risk simulation based on the user state information and outputs risk probability distribution data corresponding to different insurance decisions, including: Use real-time user behavior data collected by wearable devices and external dynamic data as input parameters; The input parameters are fed into the risk simulation model, which outputs risk probability distribution data corresponding to different insurance decisions. The model receives data updates from wearable devices and external data platforms at specific time intervals to dynamically correct the generated risk probability distribution data.

[0083] In one possible implementation, the fusion display interface displays future risk data as overlaid visual elements on the base model, and simulates environmental sounds in different scenarios using three-dimensional sound effects.

[0084] In one possible implementation, the capture module 604 captures the user's physiological reactions and behavioral feedback in real time during the interaction process based on the wearable device, and analyzes the user's insurance decision preferences based on the captured information using an emotion-decision mapping model, including: The device collects real-time data on changes in heart rate and skin conductivity as users view different possible future scenarios, while also recording the user's device operation behavior. The captured information is analyzed and processed through an emotion-decision mapping model to obtain users' insurance decision preferences. The emotion-decision mapping model uses historical users' physiological and behavioral data and corresponding insurance decision results as a training set to establish a mapping relationship between physiological and behavioral characteristics and decision preferences.

[0085] In one possible implementation, the recommendation module 605 aggregates decision data from similar users to supplement and optimize the adjusted insurance plan, including: Without obtaining the original user data, decision-making models of different user groups are jointly trained to extract common decision-making patterns. The information on common decision-making patterns is incorporated into the personalized solution optimization process to ensure that the solution conforms to the general decision-making logic of similar users.

[0086] In one possible implementation, the recommendation module 605 further includes: Establish multi-level health behavior value assessment rules; wherein, the multi-level health behavior value assessment rules are used to convert user health behaviors into quantifiable insurance value; Insurance prices are calculated in real time based on the user's wearable device-recorded behavior.

[0087] The aforementioned insurance recommendation device overcomes the fundamental limitations of existing insurance technologies—namely, "one-way information transmission, static risk assessment, linear time dimension, and passive decision-making mode"—through a closed-loop design encompassing "data collection, modeling, risk simulation, scenario display, preference analysis, scheme optimization, and dynamic pricing." It collects physiological and behavioral indicators using various wearable devices and combines this with a multi-device data collaborative calibration algorithm, upgrading data collection from "static manual entry" to "minute-level dynamic full-dimensional collection." Leveraging user dynamic status information and external data, it outputs future risk scenarios through a risk simulation model and dynamically corrects them, transforming risk assessment from "historical inference" to "real-time accurate prediction," thus improving accuracy. It constructs a fusion display interface encompassing "past, present, and future," coupled with visual elements and 3D sound effects, significantly improving user information reception efficiency, enhancing understanding of the long-term impact of insurance decisions, and lowering the decision-making threshold. Combined with wearable devices… The device captures physiological responses and behavioral feedback, and through a CNN+RNN fusion model and individual baseline calibration, it achieves preference recognition from "single operation analysis" to "two-dimensional accurate judgment," accurately capturing potential needs. Based on user preferences and group common patterns extracted by federated learning, it optimizes insurance plans, upgrading from "standardized passive selection" to "personalized and reasonable proactive adaptation," improving user acceptance. It constructs a multi-level health behavior value assessment system, transforming health behaviors into insurance value and linking them to premium discounts, promoting the extension of insurance services from "post-event compensation" to "pre-event health intervention," realizing a paradigm upgrade of insurance services from "risk transfer tool" to "symbiotic partner that co-evolves with user health." This significantly improves user experience and insurance company operational efficiency, and creates a new win-win ecosystem for "insurance-user-society," providing core technological support for the insurance industry's transformation towards technology-driven health services.

[0088] like Figure 7 As shown, Figure 7 This is a schematic diagram of the composition structure of the electronic device 700 provided in the embodiments of this application. The electronic device 700 includes: The device 700 includes a processor 701, a storage medium 702, and a bus 703. The storage medium 702 stores machine-readable instructions executable by the processor 701. When the electronic device 700 is running, the processor 701 communicates with the storage medium 702 via the bus 703. The processor 701 executes the machine-readable instructions to perform the steps of the insurance recommendation method described in the embodiments of this application.

[0089] In practical applications, the various components in the electronic device 700 are coupled together via a bus 703. It is understood that the bus 703 is used to achieve communication between these components. In addition to a data bus, the bus 703 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in... Figure 7The general designated all buses as Bus 703.

[0090] The aforementioned electronic devices, through a closed-loop design of "data acquisition - modeling - risk simulation - scenario display - preference analysis - solution optimization - dynamic pricing," overcome the fundamental limitations of existing insurance technologies, such as "one-way information transmission, static risk assessment, linear time dimension, and passive decision-making mode." They collect physiological and behavioral indicators through multiple types of wearable devices, combined with multi-device data collaborative calibration algorithms, upgrading data from "static manual reporting" to "minute-level dynamic full-dimensional acquisition." Relying on user dynamic status information and external data, they output future risk scenarios through risk simulation models and dynamically correct them, transforming risk assessment from "historical inference" to "real-time accurate prediction," thus improving accuracy. They construct a "past-present-future" integrated display interface, coupled with visual elements and 3D sound effects, greatly improving the efficiency of user information reception, enhancing understanding of the long-term impact of insurance decisions, and lowering the decision-making threshold. Combined with wearable devices... It captures physiological responses and behavioral feedback, and through a CNN+RNN fusion model and individual baseline calibration, it achieves preference recognition from "single-operation analysis" to "two-dimensional accurate judgment," accurately capturing potential needs. Based on user preferences and group commonalities extracted by federated learning, it optimizes insurance plans, upgrading from "standardized passive selection" to "personalized and reasonable proactive adaptation," improving user acceptance. It constructs a multi-level health behavior value assessment system, transforming health behaviors into insurance value and linking them to premium discounts, promoting the extension of insurance services from "post-event compensation" to "pre-event health intervention," realizing a paradigm upgrade of insurance services from "risk transfer tool" to "symbiotic partner that co-evolves with user health." This significantly improves user experience and insurance company operational efficiency, and creates a new win-win ecosystem for "insurance-user-society," providing core technical support for the insurance industry's transformation towards technology-driven health services.

[0091] This application also provides a computer-readable storage medium storing executable instructions that, when executed by at least one processor 701, implement the insurance recommendation method described in this application.

[0092] In some embodiments, the storage medium may be a magnetic random access memory (FRAM), a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM), etc.; or it may be a device that includes one or any combination of the above-mentioned memories.

[0093] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0094] As an example, executable instructions may, but do not necessarily, correspond to files in the file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a HyperText Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0095] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0096] The aforementioned computer-readable storage medium, through a closed-loop design of "data acquisition - modeling - risk simulation - scenario display - preference analysis - solution optimization - dynamic pricing," overcomes the fundamental limitations of existing insurance technologies, such as "one-way information transmission, static risk assessment, linear time dimension, and passive decision-making mode." It collects physiological and behavioral indicators through various types of wearable devices, combined with multi-device data collaborative calibration algorithms, upgrading data from "static manual reporting" to "minute-level dynamic full-dimensional collection." Relying on user dynamic status information and external data, it outputs future risk scenarios through risk simulation models and dynamically corrects them, transforming risk assessment from "historical inference" to "real-time accurate prediction," thus improving accuracy. It constructs a "past-present-future" integrated display interface, coupled with visual elements and 3D sound effects, greatly improving the efficiency of user information reception, enhancing understanding of the long-term impact of insurance decisions, and lowering the decision-making threshold. Combined with wearable... Wearable devices capture physiological responses and behavioral feedback. Through a CNN+RNN fusion model and individual baseline calibration, preference recognition is transformed from "single-operation analysis" to "two-dimensional accurate judgment," precisely capturing potential needs. Based on user preferences and group commonalities extracted through federated learning, insurance plans are optimized, upgrading from "standardized passive selection" to "personalized and reasonable proactive adaptation," improving user acceptance. A multi-level health behavior value assessment system is constructed, transforming health behaviors into insurance value and linking them to premium discounts. This drives insurance services to extend from "post-event compensation" to "pre-event health intervention," realizing a paradigm shift from "risk transfer tool" to "symbiotic partner co-evolving with user health." This significantly improves user experience and insurance company operational efficiency, while creating a new win-win ecosystem for "insurance-user-society," providing core technological support for the insurance industry's transformation towards technology-driven health services.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed methods and electronic devices can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components may be combined, or integrated into another system, or some features may be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed may be through some interfaces, and the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0098] The modules described as separate components may or may not be physically separate. The components shown as modules 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.

[0099] In addition, 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.

[0100] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a processor-executable, non-volatile, 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 a portion 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 several instructions to cause a computer device (which may be a personal computer, a platform server, or a 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 program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0101] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. An insurance recommendation method, characterized in that, The method includes: By collecting user physiological and behavioral indicators through various types of wearable devices, a dynamic physiological-behavioral fusion model for users is established based on the collected indicator data; wherein, the dynamic physiological-behavioral fusion model for users is used to update user status information according to the data transmitted back from wearable devices in real time. Based on the user status information, risk simulation is performed, and risk probability distribution data corresponding to different insurance decisions are output. Generate a fusion display interface that includes the user's past life trajectory, current life status, and possible future scenarios; wherein, the possible future scenarios are generated based on the risk probability distribution data; In response to users viewing possible future scenarios corresponding to different insurance decisions, the system captures users' physiological reactions and behavioral feedback in real time during the interaction process using wearable devices, and analyzes users' insurance decision preferences based on the captured information using an emotion-decision mapping model. Based on the user's decision-making preferences and historical behavior data, the insurance parameters of the insurance plan are adjusted, and the decision-making data of similar users are aggregated to supplement and optimize the adjusted insurance plan, and the supplemented and optimized insurance plan is recommended to the user.

2. The method according to claim 1, characterized in that, The method also includes, The indicator data collected from different wearable devices are preprocessed based on a multi-device data collaborative calibration algorithm; wherein the preprocessing includes at least one of the following processes: Synchronize the heart rate data collected by fitness trackers, the blood oxygen data collected by smartwatches, and the sleep data collected by sleep monitors along a timeline; High-frequency noise in the data is removed by wavelet transform algorithm, abnormal data is filtered by adaptive threshold method and completed based on the nearest valid data; User physiological and behavioral features are extracted using deep belief networks to construct user state feature vectors.

3. The method according to claim 1, characterized in that, The step of performing risk simulation based on the user status information and outputting risk probability distribution data corresponding to different insurance decisions includes: Use real-time user behavior data collected by wearable devices and external dynamic data as input parameters; The input parameters are fed into the risk simulation model, which outputs risk probability distribution data corresponding to different insurance decisions. The model receives data updates from wearable devices and external data platforms at specific time intervals to dynamically correct the generated risk probability distribution data.

4. The method according to claim 1, characterized in that, In the integrated display interface, future risk data is displayed as visual elements overlaid on the basic model, and environmental sounds in different scenarios are simulated through three-dimensional sound effects.

5. The method according to claim 1, characterized in that, The method involves real-time capture of users' physiological reactions and behavioral feedback during interaction using wearable devices, and analysis of users' insurance decision preferences based on the captured information using an emotion-decision mapping model, including: The device collects real-time data on changes in heart rate and skin conductivity as users view different possible future scenarios, while also recording the user's device operation behavior. The captured information is analyzed and processed through an emotion-decision mapping model to obtain users' insurance decision preferences. The emotion-decision mapping model uses historical users' physiological and behavioral data and corresponding insurance decision results as a training set to establish a mapping relationship between physiological and behavioral characteristics and decision preferences.

6. The method according to claim 1, characterized in that, The aggregation of decision data from similar users is used to supplement and optimize the adjusted insurance plan, including: Without obtaining the original user data, decision-making models of different user groups are jointly trained to extract common decision-making patterns. The information on common decision-making patterns is incorporated into the personalized solution optimization process to ensure that the solution conforms to the general decision-making logic of similar users.

7. The method according to claim 1, characterized in that, The method further includes: Establish multi-level health behavior value assessment rules; wherein, the multi-level health behavior value assessment rules are used to convert user health behaviors into quantifiable insurance value; Insurance prices are calculated in real time based on the user's wearable device-recorded behavior.

8. An insurance recommendation device, characterized in that, The device includes: The data acquisition module is used to collect user physiological and behavioral indicators through various types of wearable devices, and to establish a dynamic physiological-behavioral fusion model based on the collected indicator data; wherein, the dynamic physiological-behavioral fusion model is used to update user status information according to the data transmitted back from the wearable devices in real time. The simulation module is used to perform risk simulation based on the user status information and output risk probability distribution data corresponding to different insurance decisions; The generation module is used to generate a fusion display interface that includes the user's past life trajectory, current life status, and possible future scenarios; wherein, the possible future scenarios are generated based on the risk probability distribution data; The capture module is used to respond to the possible future scenarios corresponding to different insurance decisions viewed by users. It captures the user's physiological reactions and behavioral feedback in real time during the interaction process based on wearable devices, and analyzes the user's insurance decision preferences based on the captured information using the emotion-decision mapping model. The recommendation module is used to adjust the insurance parameters of the insurance plan based on the user's decision preferences and historical behavior data, and to aggregate the decision data of similar users to supplement and optimize the adjusted insurance plan, and recommend the supplemented and optimized insurance plan to the user.

9. An electronic device, characterized in that, include: The device includes a processor, a storage medium, and a bus, wherein the storage medium stores machine-readable instructions executable by the processor, and when the electronic device is in operation, the processor communicates with the storage medium via the bus, and the processor executes the machine-readable instructions to perform the insurance recommendation method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, performs the insurance recommendation method as described in any one of claims 1 to 7.

Images