AI driving type electronic cigarette smoking cessation method and system with insurance incentive mechanism

Abstract

The invention relates to the field of electronic cigarettes, in particular to an AI-driven electronic cigarette smoking cessation method and system with an insurance incentive mechanism, and the method comprises the steps: firstly obtaining user multi-modal data and user portrait information, and then inputting the user multi-modal data and the user portrait information into a pre-trained neural network model; the smoking impulse probability of the user in the future time period is predicted, then a tobacco tar mixing ratio modulation parameter matched with the smoking impulse probability is called from a preset database, and the user portrait information and the tobacco tar mixing ratio modulation parameter are input into a reinforcement learning algorithm after excitation iteration for operation; and finally, limiting the start and stop of the electronic cigarette or the adjustment of the tobacco tar mixing ratio according to the smoking cessation intervention information. Compared with the prior art, the method solves the technical problem that the electronic cigarette control technology is difficult to promote the effective smoking cessation effect.

Description

Technical Field

[0001] This invention relates to the field of e-cigarettes. More specifically, this invention relates to an AI-driven e-cigarette smoking cessation method and system with an insurance incentive mechanism. Background Technology

[0002] As a substitute for traditional tobacco, e-cigarettes, with their harm-reduction properties and intelligent functions, are gradually becoming an important choice for smoking cessation aids. However, current e-cigarette technology still suffers from three major flaws, severely limiting its effectiveness in smoking cessation and users' willingness to use it long-term: Firstly, traditional e-cigarette e-liquid cartridges use a single formula design, where the mixing ratio of nicotine salts and flavoring substances is fixed during the production stage. Users cannot dynamically adjust the ingredient ratio according to their own smoking urges, making it difficult to effectively overcome their nicotine addiction.

[0003] Secondly, although existing smart e-cigarettes have basic data collection functions, their hardware layer and cloud services are connected through a one-way transmission protocol, which only realizes data upload without forming a feedback loop and thus does not achieve effective personalized smoking cessation intervention.

[0004] Third, the health insurance industry has not yet established an effective linkage mechanism with e-cigarette smoking cessation behavior, making it difficult for users to have the motivation to quit smoking.

[0005] Therefore, existing e-cigarette control technologies are unlikely to effectively promote smoking cessation. Summary of the Invention

[0006] To address the technical problem that e-cigarette control technology is unable to effectively promote smoking cessation, this invention discloses an AI-driven e-cigarette smoking cessation method and system with an insurance incentive mechanism.

[0007] In a first aspect, this invention discloses an AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism, comprising: Acquire user multimodal data and user profile information; By inputting user multimodal data and user profile information into a pre-trained neural network model, the probability of a user having the urge to smoke in the future time period can be predicted. Retrieve e-liquid mixing parameters that match the probability of the urge to smoke from the preset database; The user profile information and the e-liquid mixing ratio modulation parameters are input into a reinforcement learning algorithm after stimulation iteration to obtain the user's smoking cessation intervention information. Information on smoking cessation interventions can be used to restrict the activation and deactivation of e-cigarettes or the adjustment of e-liquid mixing ratios.

[0008] Beneficial effects: The method of this invention relies on a neural network model to predict the probability of smoking urges, obtain dynamically changing e-liquid mixing ratio modulation parameters, and use them as reference parameters to adjust the e-liquid mixing ratio of e-cigarettes. Then, reinforcement learning algorithms are used to perform further reinforcement learning to obtain smoking cessation intervention information that is more in line with the user's individual situation. Finally, the smoking cessation intervention information is used to limit the start and stop of e-cigarettes or the adjustment of the e-liquid mixing ratio, thereby forming an effective feedback loop and personalized smoking cessation intervention, overcoming the problems existing in the prior art.

[0009] Preferably, user multimodal data and user profile information are input into a pre-trained neural network model to predict the probability of a user's smoking urge in the future, specifically as follows: Extracting users' smoking scenarios from multimodal user data; User multimodal data and user profile information are input into a neural network model based on the Transformer architecture to generate an impulse probability matrix; Based on the smoking scenario, retrieve the corresponding smoking impulse probability from the impulse probability matrix.

[0010] Preferably, the smoking cessation intervention information includes at least the daily smoking frequency, duration of each smoking session, the time of herbal cartridge delivery, and / or deep breathing training prompts.

[0011] Preferred methods for using smoking cessation intervention information to restrict e-cigarette activation and deactivation include: In response to a user's cigarette lighting behavior, determine whether the frequency of cigarette lighting on that day exceeds the daily smoking frequency. If so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user's device. If not, record the user's cigarette lighting time, determine if the lighting time exceeds the duration of each smoking session, and if so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user's device.

[0012] Preferably, the restriction of e-liquid mixing ratio adjustment based on smoking cessation intervention information includes: If the e-cigarette is successfully ignited, the time of ignition for that day will be identified. Determine if the smoking time falls within the herbal e-liquid delivery time. If so, output a mixture of herbal e-liquid and nicotine e-liquid in the corresponding proportion.

[0013] Preferably, after acquiring user multimodal data and user profile information, the method of the present invention further includes: User multimodal data and user profile information are input into the random forest algorithm built into the cloud AI to conduct health risk assessment and obtain risk assessment values; If the risk assessment value exceeds the risk threshold, an alert will be triggered and pushed to the user's device.

[0014] Preferably, after obtaining the risk assessment value, the method of the present invention further includes: Extract users' smoking cessation history data over a period of time from user profile information; Input smoking cessation history data into a pre-trained artificial intelligence model to obtain the user's level of active cooperation; If the degree of proactive cooperation exceeds the cooperation threshold, an additional insurance discount will be added and pushed to the user's device.

[0015] Preferably, before inputting the user profile information and e-liquid mixing ratio modulation parameters into the reinforcement learning algorithm after activation iteration for computation, the method of the present invention further includes: The algorithm continuously incentivizes the reinforcement learning algorithm by using the number of days the user quits smoking as the reward function.

[0016] Preferably, the user multimodal data includes at least carbon monoxide concentration, volatile organic compound concentration, heart rate variability, cough frequency, and smoking scenario; the user profile information includes age, smoking history, and / or smoking cessation history data.

[0017] Secondly, the present invention discloses an AI-driven e-cigarette smoking cessation system with an insurance incentive mechanism, which is used in the AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism described in the first aspect. The system of the present invention includes at least a smart e-cigarette terminal and a cloud. Among them, the smart electronic cigarette terminal is used to acquire user multimodal data and user profile information, and is also used to restrict the start and stop of electronic cigarettes or adjust the e-liquid mixing ratio based on smoking cessation intervention information; The execution functions in the cloud include: By inputting user multimodal data and user profile information into a pre-trained neural network model, the probability of a user having the urge to smoke in the future time period can be predicted. Retrieve e-liquid mixing parameters that match the probability of the urge to smoke from the preset database; User profile information and e-liquid mixing ratio parameters are input into a reinforcement learning algorithm after stimulation iteration to obtain user smoking cessation intervention information.

[0018] The beneficial effects of this invention are as follows: (1) Compared with the prior art, the method of the present invention relies on reinforcement learning algorithm to form an effective feedback loop, and uses smoking cessation intervention information to limit the start and stop of e-cigarettes or the adjustment of e-liquid mixing ratio, thereby forming an effective feedback loop and personalized smoking cessation intervention, overcoming the problems existing in the prior art.

[0019] (2) Compared with the prior art, the method of the present invention supports effective linkage with the health insurance industry and uses real-time discount information from insurance feedback to incentivize users to quit smoking.

[0020] (3) Compared with the prior art, the system of the present invention supports real-time feedback between the smart electronic cigarette terminal and the cloud, and can continuously optimize the intervention strategy that fits the actual situation of the user by combining reinforcement learning, thereby improving the user retention rate. Attached Figure Description

[0021] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a flowchart of the AI-driven electronic cigarette smoking cessation method with an insurance incentive mechanism in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the structure of the AI-driven electronic cigarette smoking cessation system with an insurance incentive mechanism in Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of the computer system in Embodiment 3 of the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] This invention discloses an AI-driven e-cigarette smoking cessation method and system with an insurance incentive mechanism, which is used to solve the technical problem that e-cigarette control technology is difficult to promote effective smoking cessation.

[0024] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0025] Example 1 like Figure 1 As shown, this embodiment discloses an AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism, including: S10: Obtain user multimodal data and user profile information.

[0026] In this embodiment, the aforementioned user multimodal data includes carbon monoxide concentration, volatile organic compound (VOC) concentration, heart rate variability, cough frequency, blood oxygen saturation, breath sounds, and smoking scenario. User profile information includes age, smoking history, and smoking cessation history. Specifically, carbon monoxide concentration can be collected by embedding a carbon monoxide sensor in the e-cigarette product; VOC concentration can be collected by embedding a VOC sensor; heart rate variability can be collected by embedding a PPG sensor; and user voice information can be collected via a microphone. Blood oxygen saturation can be collected using a fitness tracker. The smoking scenario can be determined using a positioning module (e.g., GPS) built into the smart e-cigarette terminal. User profile information can be obtained through a user-authorized application. All the above information or data is ultimately aggregated to the smart e-cigarette terminal via wireless communication technology.

[0027] In other embodiments, the specific content of the aforementioned user multimodal data and user profile information can be expanded according to actual circumstances. The methods for acquiring this data or information can be adaptively adjusted.

[0028] S20: Input user multimodal data and user profile information into the pre-trained neural network model to predict the probability of a user's smoking urge in the future time period.

[0029] It should be noted that the above neural network model adopts a neural network model based on the Transformer architecture, which can be pre-trained to adapt to the application scenario of predicting the probability of smoking impulse in this embodiment.

[0030] Step S20 above includes: S21: Extract users' smoking scenarios from user multimodal data.

[0031] Specifically, once the e-cigarette device obtains its location, a map app can identify the user's specific smoking scenario. These scenarios can include office areas, public areas, and designated smoking areas.

[0032] S22: Input user multimodal data and user profile information into a neural network model based on the Transformer architecture to generate an impulse probability matrix.

[0033] It should be noted that the aforementioned user multimodal data and user profile information will be aggregated into a data matrix through edge computing, and then encrypted and transmitted from the e-cigarette terminal to the cloud AI for deep processing in step S22 using LoRa encryption. During this deep processing, the cloud AI uses Transformer to predict the impulse probability matrix for each scenario.

[0034] S23: Based on the smoking scenario, retrieve the corresponding smoking impulse probability from the impulse probability matrix.

[0035] Furthermore, by combining the smoking scenario with the corresponding smoking impulse probability, and determining whether the smoking impulse probability is greater than the smoking threshold, the following step S30 is executed.

[0036] S30: Retrieve e-liquid mixing ratio modulation parameters from the preset database that match the probability of smoking urge.

[0037] It should be noted that the electronic cigarette in this embodiment has an A e-liquid compartment and a B e-liquid compartment. The A e-liquid compartment contains nicotine, while the B e-liquid compartment contains a sedative ingredient. In step S30, the e-liquid mixing ratio modulation parameters of the electronic cigarette are dynamically adjusted according to the aforementioned smoking urge probability. The initial state of the e-liquid mixing ratio modulation parameters depends on the aforementioned database matching, and subsequent adjustments are made dynamically using a reinforcement learning mechanism.

[0038] For example, when the probability of a smoking urge is high, reduce the nicotine in tank A by 20% and increase the sedative component in tank B by 15%.

[0039] S40: Input the user profile information and e-liquid mixing ratio modulation parameters into the reinforcement learning algorithm after stimulation iteration to obtain the user's smoking cessation intervention information.

[0040] In this embodiment, before executing step S40, a corresponding reward function needs to be established. Specifically, the number of days the user has abstained from smoking is used as the reward function to continuously incentivize the iterative reinforcement learning algorithm. The aforementioned smoking cessation intervention information includes daily smoking frequency, duration of each smoking session, herbal e-cigarette cartridge push time, and deep breathing training prompts. The reinforcement learning algorithm can employ the PPO algorithm.

[0041] In other embodiments, other reinforcement learning algorithms may also be used.

[0042] Specifically, after inputting user profile information and e-liquid mixing ratio parameters into the PPO algorithm, the PPO algorithm is continuously iterated using "user's number of days since quitting smoking" as the reward function. Finally, based on the convergence of the PPO algorithm, the corresponding daily smoking frequency, duration of each smoking session, herbal cartridge push time, and deep breathing training prompts are output.

[0043] S50: Use smoking cessation intervention information to restrict the start / stop of e-cigarettes or the adjustment of e-liquid mixing ratio.

[0044] Regarding the restrictions on starting and stopping electronic cigarettes, step S50 includes: S510: In response to the user's cigarette lighting behavior, determine whether the frequency of cigarette lighting on that day exceeds the daily smoking frequency. If so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user.

[0045] S511: If not, record the user's cigarette lighting time, determine whether the cigarette lighting time exceeds the duration of each smoking session, and if so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user.

[0046] By limiting the frequency and duration of a user's smoking through steps S510 and S511, and taking into account the user's health condition, we can proactively and personally manage the user's smoking behavior.

[0047] Regarding the dynamic adjustment of the e-liquid mixing ratio, step S50 above includes: S520: If the e-cigarette is successfully ignited, identify the time of ignition for that day.

[0048] S521: Determine whether the smoking time of the day falls within the herbal e-liquid cartridge delivery time. If so, output a mixture of herbal e-liquid and nicotine e-liquid in the corresponding proportion.

[0049] For example, if the smoking time is 14:00 on the day, the herbal tobacco cartridge is pushed and a deep breathing instruction is issued on the e-cigarette terminal.

[0050] Through the above steps S520-S521, users can alleviate discomfort during the smoking cessation process and help reduce their dependence on nicotine.

[0051] Based on steps S10-S50 above, the method in this embodiment first uses a neural network model to predict the probability of smoking urges, then obtains dynamically changing e-liquid mixing ratio modulation parameters as reference parameters for adjusting the e-liquid mixing ratio of the e-cigarette. Next, a reinforcement learning algorithm is used for further reinforcement learning to obtain smoking cessation intervention information that better suits the user's individual situation. Finally, the smoking cessation intervention information is used to limit the activation / deactivation of the e-cigarette or the adjustment of the e-liquid mixing ratio, thus forming an effective feedback loop and personalized smoking cessation intervention. Compared with existing technologies, the method in this embodiment overcomes the technical problem that e-cigarette control technology struggles to achieve effective smoking cessation.

[0052] Furthermore, after step S50 above, the method of this embodiment further includes: S60: Input user multimodal data and user profile information into the cloud AI's built-in random forest algorithm to conduct health risk assessment and obtain risk assessment values.

[0053] In this embodiment, the aforementioned risk assessment value will be synchronized to the hospital's HS system authorized by the user information in order to achieve user information archiving.

[0054] S70: If the risk assessment value is greater than the risk threshold, an alert will be triggered and pushed to the user's terminal.

[0055] For example, based on the trend of carbon monoxide concentration changes, if the carbon monoxide concentration is greater than 10 ppm for three consecutive days and the user's breathing sounds include wheezing, the risk of acute COPD exacerbation is assessed using a random forest algorithm, and an early warning is triggered if the risk assessment value is greater than 0.7.

[0056] Furthermore, after step S60 above, the method of this embodiment further includes: S600: Extracts a user's smoking cessation history data over a period of time from user profile information.

[0057] S601: Input smoking cessation history data into a pre-trained artificial intelligence model to obtain the user's level of active cooperation.

[0058] S602: If the degree of active cooperation is greater than the cooperation threshold, add an insurance discount and push the insurance discount to the user's terminal.

[0059] Through steps S600-S601 described above, the method of this embodiment can push premiums and discounts based on the user's level of cooperation to incentivize the user to quit smoking. Specifically, the calculation of the premium can be as follows: Basic discount: 5% off premium after 30 days of quitting smoking; 15% off after 90 days of quitting smoking.

[0060] Behavioral weighting: If a user’s “proactive cooperation” (such as completing intervention tasks on time) is greater than 80 points, an additional 5% discount will be applied.

[0061] Risk hedging: Insurance companies dynamically adjust the discount cap based on the user's health risk value.

[0062] Based on the above technical discussion, the method of this embodiment has at least the following advantages: (1) Existing smart e-cigarettes rely on one-way data transmission and rule-based intervention, resulting in a user retention rate of less than 30%. However, the method in this embodiment uses real-time feedback between hardware and cloud, combined with PPO reinforcement learning to continuously optimize the intervention strategy, thereby increasing the 12-month user retention rate to 67% and the intervention task completion rate from 42% to 89%. Therefore, the method in this embodiment has a closed-loop intervention effect and high user compliance.

[0063] (2) Traditional insurance plans are disconnected from smoking cessation behavior, requiring users to purchase additional health management services. However, the method in this embodiment reduces the annual premium expenditure of highly cooperative users by 35% through a dynamic premium discount system, while reducing the insurance company's loss ratio by 18%, forming a win-win closed loop between users and insurance companies.

[0064] Example 2 like Figure 2As shown, this embodiment also discloses an AI-driven e-cigarette smoking cessation system with an insurance incentive mechanism, which is used in the AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism described in Embodiment 1. The system in this embodiment includes a smart e-cigarette terminal, a cloud, an insurance server, and a user terminal.

[0065] Among them, the smart electronic cigarette terminal is used to acquire user multimodal data and user profile information, and is also used to restrict the start and stop of electronic cigarettes or adjust the e-liquid mixing ratio based on smoking cessation intervention information.

[0066] The execution functions in the cloud include: By inputting user multimodal data and user profile information into a pre-trained neural network model, the probability of a user having the urge to smoke in the future time period can be predicted.

[0067] Retrieve e-liquid mixing ratio parameters that match the probability of smoking urges from the preset database.

[0068] User profile information and e-liquid mixing ratio parameters are input into a reinforcement learning algorithm after stimulation iteration to obtain user smoking cessation intervention information.

[0069] Furthermore, the aforementioned smart electronic cigarette independently controls the flow rate of tanks A and B via solenoid valves, with a flow rate adjustment range of 0.1-1.0 mL / min to 1.0 mL / min. Combined with the high-frequency vibration (10 kHz-50 kHz) of the piezoelectric pump, it achieves millisecond-level uniform mixing of components, avoiding the delay (greater than 500 ms) and uneven mixing (concentration fluctuation greater than 15%) problems of traditional mechanical pumps, ensuring that the components inhaled by the user are completely synchronized with the AI ​​commands.

[0070] Furthermore, the aforementioned insurance service platform is used to push premiums and discounts to users based on their level of cooperation and relevant smoking cessation status. The user terminal can be a user's mobile phone or a fitness watch.

[0071] Example 3 like Figure 3 As shown, this embodiment also discloses a computer system. The system includes a processor and a memory, the memory storing computer program instructions. When the computer program instructions are executed by the processor, the AI-driven electronic cigarette smoking cessation method with an insurance incentive mechanism described in Embodiment 1 is implemented.

[0072] The system also includes other components well known to those skilled in the art, such as communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0073] In this invention, the aforementioned memory can be any tangible medium containing or storing a program that can be used or combined with an instruction execution system, apparatus, or device. For example, a computer-readable storage medium can be any suitable magnetic or magneto-optical storage medium, such as Resistive Random Access Memory (RRAM), Dynamic Random Access Memory (DRAM), Static Random Access Memory (SRAM), Enhanced Dynamic Random Access Memory (EDRAM), High-Bandwidth Memory (HBM), Hybrid Memory Cube (HMC), etc., or any other medium that can be used to store desired information and can be accessed by an application, module, or both. Any such computer storage medium can be part of a device or accessible to or connected to a device. Any application or module described in this invention can be implemented using computer-readable / executable instructions that can be stored or otherwise maintained by such a computer-readable medium.

[0074] In the description of this specification, "multiple" means at least two, such as two, three or more, etc., unless otherwise expressly and specifically defined.

[0075] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. An AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism, characterized in that, include: Acquire user multimodal data and user profile information; The user's multimodal data and user profile information are input into a pre-trained neural network model to predict the probability of the user's smoking urge in the future time period. Retrieve e-liquid mixing ratio modulation parameters that match the probability of the smoking urge from the preset database; The user profile information and the e-liquid mixing ratio modulation parameters are input into a reinforcement learning algorithm after stimulation iteration to obtain the user's smoking cessation intervention information. The smoking cessation intervention information is used to restrict the activation and deactivation of e-cigarettes or the adjustment of the e-liquid mixing ratio.

2. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 1, characterized in that, The user's multimodal data and user profile information are input into a pre-trained neural network model to predict the probability of a user's smoking urge in the future time period, specifically: Extract the user's smoking scenario from the user's multimodal data; The user multimodal data and the user profile information are input into a neural network model based on the Transformer architecture to generate an impulse probability matrix; Based on the smoking scenario, retrieve the corresponding smoking impulse probability from the impulse probability matrix.

3. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 1, characterized in that, The smoking cessation intervention information includes at least the daily smoking frequency, duration of each smoking session, the time of herbal e-cigarette cartridge delivery, and / or deep breathing training prompts.

4. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 3, characterized in that, Restricting e-cigarette activation and deactivation using the aforementioned smoking cessation intervention information includes: In response to a user's cigarette lighting behavior, determine whether the frequency of cigarette lighting on that day exceeds the daily smoking frequency. If so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user's device. If not, record the user's cigarette lighting time, determine whether the cigarette lighting time exceeds the duration of each smoking session, and if so, interrupt the cigarette lighting process and push a cigarette lighting warning to the user's device.

5. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 3, characterized in that, Limiting the e-liquid mixing ratio adjustment based on the aforementioned smoking cessation intervention information includes: If the e-cigarette is successfully ignited, the time of ignition for that day will be identified. Determine whether the smoking time of the day falls within the time period for pushing the herbal e-liquid cartridge. If so, output a mixture of herbal e-liquid and nicotine e-liquid in the corresponding proportion.

6. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 1, characterized in that, After acquiring user multimodal data and user profile information, the method further includes: The user's multimodal data and user profile information are input into the random forest algorithm built into the cloud AI to perform health risk assessment and obtain a risk assessment value. If the risk assessment value is greater than the risk threshold, an alert will be triggered and pushed to the user's device.

7. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 6, characterized in that, After obtaining the risk assessment value, the method further includes: Extract the user's smoking cessation history data over a period of time from the user profile information; The smoking cessation history data is input into a pre-trained artificial intelligence model to obtain the user's level of active cooperation; If the degree of active cooperation is greater than the cooperation threshold, an additional insurance discount is applied, and the insurance discount is pushed to the user's terminal.

8. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 1, characterized in that, Before inputting the user profile information and the e-liquid mixing ratio modulation parameters into the reinforcement learning algorithm after activation iteration for computation, the method further includes: The reinforcement learning algorithm is continuously incentivized iteratively using the number of days the user has quit smoking as a reward function.

9. The AI-driven e-cigarette smoking cessation method with an insurance incentive mechanism according to claim 1, characterized in that, The user multimodal data includes at least carbon monoxide concentration, volatile organic compound concentration, heart rate variability, cough frequency, and smoking scenario; the user profile information includes age, smoking history, and / or smoking cessation history data.

10. An AI-driven electronic cigarette smoking cessation system with an insurance incentive mechanism, characterized in that, The AI-driven electronic cigarette smoking cessation method with an insurance incentive mechanism as described in any one of claims 1-9, wherein the system includes at least a smart electronic cigarette terminal and a cloud; The intelligent electronic cigarette terminal is used to acquire user multimodal data and user profile information, and is also used to restrict the start / stop of electronic cigarettes or adjust the e-liquid mixing ratio based on smoking cessation intervention information. The cloud-based execution functions include: The user's multimodal data and user profile information are input into a pre-trained neural network model to predict the probability of the user's smoking urge in the future time period. Retrieve e-liquid mixing ratio modulation parameters that match the probability of the smoking urge from the preset database; The user profile information and the e-liquid mixing ratio modulation parameters are input into a reinforcement learning algorithm after stimulation iteration to obtain the user's smoking cessation intervention information.

Images