Urinary incontinence warning system based on smart chip technology

Abstract

The present invention discloses a urinary incontinence warning system based on intelligent chip technology, and the system includes: a data acquisition module, a data preprocessing module, a module for building a urinary incontinence risk assessment model, a urinary incontinence risk assessment model parameter optimization module and a urinary incontinence real-time warning module. The present invention belongs to the field of risk warning technology, specifically refers to designing an adaptive scaling factor based on a total gradient norm and a historical gradient benchmark, using the adaptive scaling factor to prune the gradient of a neural network parameter, designing an adaptive learning rate, and updating the neural network parameter in combination with an interpolation factor; generating a query value and a reference position based on a local position correction value and a local search ratio value, generating a perturbation position based on a Levy flight strategy and a random individual, and updating the position of all individuals based on the reference position and the perturbation position, determining the optimal parameters, improving the accuracy and reliability of the urinary incontinence risk assessment model, and improving the quality of life and convenience of users.

Description

Technical Field

[0001] The present invention belongs to the technical field of risk warning, and specifically refers to a urinary incontinence warning system based on smart chip technology. Background Art

[0002] The urinary incontinence early warning system uses artificial intelligence and smart chip technology to monitor the risk of urinary incontinence in real time, thereby predicting the occurrence of urinary incontinence and sending reminders to users in a timely manner to improve the quality of life of patients with urinary incontinence. However, the existing urinary incontinence early warning system has the problem that the urinary incontinence monitoring data is complex and it is difficult to accurately extract key features, which increases the difficulty of model evaluation and leads to unstable and inaccurate evaluation results; the existing urinary incontinence early warning system has the problem that the urinary incontinence monitoring data is diverse and it is difficult to cope with patient data with various characteristics, resulting in misjudgment of urinary incontinence risk assessment and the evaluation results are not applicable to some patients. Summary of the invention

[0003] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a urinary incontinence early warning system based on intelligent chip technology. In view of the problems in the existing urinary incontinence early warning system, such as complex urinary incontinence monitoring data, difficulty in accurately extracting key features, increased difficulty in model evaluation, and unstable and inaccurate evaluation results, this solution designs an adaptive scaling factor based on the total gradient norm and the historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of the neural network parameters, designs an adaptive learning rate, and updates the neural network parameters in combination with the interpolation factor, thereby avoiding training anomalies caused by gradient instability and gradient explosion, enabling the learning rate to better adapt to different training stages, helping to better process complex data, and improving the accuracy and reliability of the urinary incontinence risk assessment model, thereby improving user experience. quality of life and convenience; in view of the fact that the existing urinary incontinence early warning system has the problem of diversified urinary incontinence monitoring data and difficulty in dealing with patient data with various characteristics, which leads to misjudgment in urinary incontinence risk assessment and the inapplicability of assessment results to some patients, this scheme generates query values ​​and reference positions based on local position correction values ​​and local search ratio values, generates perturbation positions based on Levy flight strategy and random individuals, and updates the positions of all individuals based on the reference position and perturbation position to determine the optimal parameters, thereby improving the efficiency and accuracy of model parameter optimization, enhancing the adaptability and flexibility of the model, better adapting to the individual differences of different patients, improving the generalization ability of the model, and improving the performance and reliability of the urinary incontinence risk assessment model, which helps to more accurately warn of urinary incontinence risks.

[0004] The urinary incontinence early warning system based on intelligent chip technology provided by the present invention includes a data acquisition module, a data preprocessing module, a urinary incontinence risk assessment model building module, a urinary incontinence risk assessment model parameter optimization module and a urinary incontinence real-time early warning module;

[0005] The data acquisition module collects historical urinary incontinence monitoring data based on the smart chip;

[0006] The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction on the collected data;

[0007] The module for constructing a urinary incontinence risk assessment model designs an adaptive scaling factor based on a total gradient norm and a historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of a neural network parameter, designs an adaptive learning rate, and updates the neural network parameter in combination with an interpolation factor;

[0008] The urinary incontinence risk assessment model parameter optimization module generates a query value and a reference position based on a local position correction value and a local search ratio value, generates a disturbance position based on a Levy flight strategy and a random individual, and updates the positions of all individuals based on the reference position and the disturbance position to determine the optimal parameters;

[0009] The urinary incontinence real-time warning module collects real-time urinary incontinence monitoring data based on a smart chip. When the risk assessment level output by the urinary incontinence risk assessment model is higher than the risk assessment threshold, the smart chip sends a warning message.

[0010] Furthermore, the data acquisition module collects historical urinary incontinence monitoring data based on a smart chip; the historical urinary incontinence monitoring data includes physiological data, user behavior data, health status data and risk assessment level; the physiological data includes urine volume, urethral pressure and bladder residual urine volume; the user behavior data includes urination habits, water intake and exercise volume; the health status data includes heart rate and respiratory rate; the higher the risk assessment level, the greater the risk of urinary incontinence for the user, and the risk assessment level is used as a data label.

[0011] Furthermore, the data preprocessing module performs data cleaning, data conversion, data normalization and data set construction on the collected data; data cleaning includes processing missing values, outliers and duplicate values; data conversion is converting data into vector form; data normalization is based on the maximum and minimum normalization method to unify the data range; data set construction is to construct a risk assessment data set based on the processed data, and transmit the risk assessment data set to the cloud server based on the smart chip.

[0012] Furthermore, the module for constructing the urinary incontinence risk assessment model is to construct a neural network, which includes an input layer, a hidden layer and an output layer. The input layer is responsible for receiving data in the risk assessment data set, the hidden layer performs data processing and feature extraction, and the output layer outputs an assessment label. The neural network parameter update specifically includes the following contents:

[0013] The first and second moments are calculated using the following formula:

[0014] ;

[0015] ;

[0016] ;

[0017] In the formula, g t is the gradient of the loss function to the neural network parameters at the tth update, is the gradient on the parameter θ, J is the loss function, θ t-1 is the neural network parameter at the t-1th update, m t and m t-1 are the first-order moments at the t-th and t-1-th updates, respectively, and v t and v t-1 are the second-order moments at the t-th and t-1-th updates, β1 and β2 are the first-order moment decay rates and the second-order moment decay rates, Yes t The square of , t is the update index of the neural network parameters;

[0018] Bias correction, the formula used is as follows:

[0019] ;

[0020] ;

[0021] In the formula, and are the first-order moment and second-order moment after bias correction at the t-th update, and are the t-th power of β1 and β2 respectively;

[0022] Gradient pruning, designing adaptive scaling factors, pruning the gradients of neural network parameters based on the adaptive scaling factors, including the following:

[0023] Calculate the total gradient norm using the following formula:

[0024] ;

[0025] In the formula, sf t is the total gradient norm at the tth update, d is the parameter index, D is the total number of neural network parameters, is the gradient of the dth parameter at the jth update, is the square of the L2 norm;

[0026] Calculate the historical gradient benchmark, which includes the historical gradient mean and historical gradient variance. The formula used is as follows:

[0027] ;

[0028] ;

[0029] In the formula, lz t and lc t They are the historical gradient mean and historical gradient variance at the tth update, g j is the gradient of the loss function with respect to the neural network parameters at the jth update;

[0030] Design the adaptive scaling factor using the following formula:

[0031] ;

[0032] In the formula, zs t is the adaptive scaling factor at the tth update, gn max is the maximum gradient norm, γ is the control factor;

[0033] Apply trimming using the following formula:

[0034] ;

[0035] In the formula, is the gradient of the dth parameter after pruning at the jth update;

[0036] Design an adaptive learning rate; dynamically adjust the learning rate according to time and set parameters, and the learning rate changes periodically during training; the formula used is as follows:

[0037] ;

[0038] In the formula, xl t is the adaptive learning rate at the tth update, xl0, xl fd and xl zq They are the initial value of the learning rate, the cycle length of the learning rate change, and the amplitude of the learning rate change;

[0039] Neural network parameter update, based on adaptive learning rate and interpolation factor to update the neural network parameters; the formula used is as follows:

[0040] ;

[0041] In the formula, θ t is the neural network parameter at the tth update, σ is the interpolation factor, is the smoothing term.

[0042] Furthermore, the urinary incontinence risk assessment model parameter optimization module specifically includes the following contents:

[0043] Initialize the individual position, and adjust the first-order moment decay rate β1, second-order moment decay rate β2, and maximum gradient norm gn in the urinary incontinence risk assessment model max , control factor γ, initial value of learning rate xl0, cycle length of learning rate change xl fd , the magnitude of the learning rate change xl zq and the interpolation factor σ to establish a search space, randomly initialize A individual positions in the search space, use the individual positions as representatives of the parameters of the urinary incontinence risk assessment model, and use the loss function of the urinary incontinence risk assessment model established based on the model parameters as the individual fitness value;

[0044] Design reference positions, including the following:

[0045] Calculate the local search ratio value using the following formula:

[0046] ;

[0047] In the formula, is the local search ratio value at the u-th iteration, is the position of the global optimal individual at the u-th iteration, is the average position of all individuals at the u-th iteration, u is the current iteration number, i is the individual index, rand(·) is the random number generation function, is the position of the i-th individual at the u-th iteration;

[0048] To calculate the local position correction, the formula used is as follows:

[0049] ;

[0050] In the formula, is the local position correction value of the ith individual at the uth iteration, is the position of the global worst individual at the u-th iteration;

[0051] Calculate the query value and generate two different query values ​​for each individual based on the local position correction value and the local search ratio value. The formula used is as follows:

[0052] ;

[0053] ;

[0054] In the formula, and are the first query value and the second query value of the i-th individual at the u-th iteration respectively;

[0055] Calculate two reference positions, generate two different reference positions for each individual based on the query value, and guide the individual to move in a more promising direction. The formula used is as follows:

[0056] ;

[0057] ;

[0058] In the formula, and are the first reference individual and the second reference position of the i-th individual at the u-th iteration, is the position of the first random individual at the uth iteration;

[0059] Calculate the perturbation position and generate a perturbation position for each individual based on the Levy flight strategy and random individuals; the formula used is as follows:

[0060] ;

[0061] ;

[0062] ;

[0063] In the formula, is the perturbed position of the i-th individual at the u-th iteration, a is a random number generated based on the Levy flight strategy, b and p are taken from the normal distribution, , , λ is the shape parameter, is the gamma function, is the standard deviation of the normal distribution, and are the positions of the second and third random individuals at the u-th iteration, is the ceiling function;

[0064] Individual position update: All individual positions are updated based on the reference position and the disturbance position. The formula used is as follows:

[0065] ;

[0066] In the formula, is the position of the i-th individual at the u+1-th iteration, s1 and s2 are the first and second random values ​​in the range (0, 1), respectively;

[0067] The optimal parameters are determined, the fitness threshold is set in advance, the individual fitness value and the global optimal individual are updated, and when the fitness value corresponding to the global optimal individual is lower than the fitness threshold, the model parameters corresponding to the global optimal individual position are the optimal parameters, and the urinary incontinence risk assessment model is constructed based on the optimal parameters; otherwise, if the maximum number of iterations is reached, the individual position is reinitialized; otherwise, the number of iterations is increased by 1, the reference position and the disturbance position are recalculated, and the individual position is updated.

[0068] Furthermore, the urinary incontinence real-time warning module collects real-time urinary incontinence monitoring data based on a smart chip; the real-time urinary incontinence monitoring data includes physiological data, user behavior data and health status data. After data preprocessing, the real-time urinary incontinence monitoring data is input into a urinary incontinence risk assessment model, and a risk assessment threshold is set in advance. When the risk assessment level output by the model is higher than the risk assessment threshold, the smart chip sends a warning message.

[0069] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0070] (1) In view of the problem that the existing urinary incontinence early warning system has complex urinary incontinence monitoring data and it is difficult to accurately extract key features, which increases the difficulty of model evaluation and leads to unstable and inaccurate evaluation results, this scheme designs an adaptive scaling factor based on the total gradient norm and historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of the neural network parameters, designs an adaptive learning rate, and combines the interpolation factor to update the neural network parameters, avoiding training anomalies caused by gradient instability and gradient explosion, making the learning rate better adapted to different training stages, helping to better handle complex data, improving the accuracy and reliability of the urinary incontinence risk assessment model, and improving the quality of life and convenience of users.

[0071] (2) In view of the fact that the existing urinary incontinence early warning system has the problem of diversified urinary incontinence monitoring data and difficulty in dealing with patient data with various characteristics, which leads to misjudgment in urinary incontinence risk assessment and the assessment results are not applicable to some patients, this scheme generates query values ​​and reference positions based on local position correction values ​​and local search ratio values, generates perturbation positions based on Levy flight strategy and random individuals, and updates the positions of all individuals based on the reference position and perturbation position to determine the optimal parameters, thereby improving the efficiency and accuracy of model parameter optimization, enhancing the adaptability and flexibility of the model, better adapting to the individual differences of different patients, improving the generalization ability of the model, and improving the performance and reliability of the urinary incontinence risk assessment model, which helps to more accurately warn of urinary incontinence risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0072] Figure 1 A schematic diagram of a urinary incontinence early warning system based on smart chip technology provided by the present invention;

[0073] Figure 2 A schematic diagram of the module for constructing a urinary incontinence risk assessment model;

[0074] Figure 3 Schematic diagram of the parameter optimization module for the urinary incontinence risk assessment model.

[0075] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention. DETAILED DESCRIPTION

[0076] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0077] In the description of the present invention, it should be understood that terms such as “upper”, “lower”, “front”, “back”, “left”, “right”, “top”, “bottom”, “inside” and “outside” indicating directions or positional relationships are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as limiting the present invention.

[0078] Example 1, see Figure 1 The urinary incontinence early warning system based on the smart chip technology provided by the present invention includes a data acquisition module, a data preprocessing module, a urinary incontinence risk assessment model building module, a urinary incontinence risk assessment model parameter optimization module and a urinary incontinence real-time early warning module;

[0079] The data acquisition module collects historical urinary incontinence monitoring data based on the smart chip and sends the data to the data preprocessing module;

[0080] The data preprocessing module receives the data sent by the data acquisition module, performs data cleaning, data conversion, data normalization and data set construction on the collected data, and sends the data to the urinary incontinence risk assessment model construction module;

[0081] The module for constructing a urinary incontinence risk assessment model receives data sent by the data preprocessing module, designs an adaptive scaling factor based on the total gradient norm and the historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of the neural network parameters, designs an adaptive learning rate, and updates the neural network parameters in combination with the interpolation factor, and sends the data to the urinary incontinence risk assessment model parameter optimization module;

[0082] The urinary incontinence risk assessment model parameter optimization module receives the data sent by the urinary incontinence risk assessment model construction module, generates a query value and a reference position based on a local position correction value and a local search ratio value, generates a disturbance position based on a Levy flight strategy and a random individual, and updates the positions of all individuals based on the reference position and the disturbance position, determines the optimal parameters, and sends the data to the urinary incontinence real-time warning module;

[0083] The urinary incontinence real-time warning module receives data sent by the urinary incontinence risk assessment model parameter optimization module, collects real-time urinary incontinence monitoring data based on the smart chip, and when the risk assessment level output by the urinary incontinence risk assessment model is higher than the risk assessment threshold, the smart chip sends a warning message.

[0084] Example 2, see Figure 1 This embodiment is based on the above embodiment. The data acquisition module collects historical urinary incontinence monitoring data based on a smart chip; the historical urinary incontinence monitoring data includes physiological data, user behavior data, health status data and risk assessment level; physiological data includes urine volume, urethral pressure and bladder residual urine volume; user behavior data includes urination habits, water intake and exercise volume; health status data includes heart rate and respiratory rate; the higher the risk assessment level, the greater the risk of urinary incontinence for the user, and the risk assessment level is used as a data label.

[0085] Example 3, see Figure 1 This embodiment is based on the above embodiment, and the data preprocessing module performs data cleaning, data conversion, data normalization and data set construction on the collected data; data cleaning includes processing missing values, outliers and duplicate values; data conversion is to convert the data into vector form; data normalization is to unify the data range based on the maximum and minimum normalization method; data set construction is to construct a risk assessment data set based on the processed data, and transmit the risk assessment data set to the cloud server based on the smart chip.

[0086] Example 4, see Figure 1 and Figure 2 This embodiment is based on the above embodiment. The urinary incontinence risk assessment model module is to construct a neural network. The neural network includes an input layer, a hidden layer and an output layer. The input layer is responsible for receiving data in the risk assessment data set. The hidden layer performs data processing and feature extraction. The output layer outputs the assessment label. The neural network parameter update specifically includes the following contents:

[0087] The first and second moments are calculated using the following formula:

[0088] ;

[0089] ;

[0090] ;

[0091] In the formula, g t is the gradient of the loss function to the neural network parameters at the tth update, is the gradient on the parameter θ, J is the loss function, θ t-1 is the neural network parameter at the t-1th update, m t and m t-1 are the first-order moments at the t-th and t-1-th updates, respectively, and v t and v t-1 are the second-order moments at the t-th and t-1-th updates, β1 and β2 are the first-order moment decay rates and the second-order moment decay rates, Yes t The square of , t is the update index of the neural network parameters;

[0092] Bias correction, the formula used is as follows:

[0093] ;

[0094] ;

[0095] In the formula, and are the first-order moment and second-order moment after bias correction at the t-th update, and are the t-th power of β1 and β2 respectively;

[0096] Gradient pruning, designing adaptive scaling factors, pruning the gradients of neural network parameters based on the adaptive scaling factors, including the following:

[0097] Calculate the total gradient norm using the following formula:

[0098] ;

[0099] In the formula, sf t is the total gradient norm at the tth update, d is the parameter index, D is the total number of neural network parameters, is the gradient of the dth parameter at the jth update, is the square of the L2 norm;

[0100] Calculate the historical gradient benchmark, which includes the historical gradient mean and historical gradient variance. The formula used is as follows:

[0101] ;

[0102] ;

[0103] In the formula, lz t and lc t They are the historical gradient mean and historical gradient variance at the tth update, g j is the gradient of the loss function with respect to the neural network parameters at the jth update;

[0104] Design an adaptive scaling factor; scale when the gradient is too large to prevent gradient explosion; keep it unchanged when the gradient is small, so as to flexibly adjust the size of the gradient, avoid over-adjustment and under-adjustment, and improve the stability and convergence speed of model training; the formula used is as follows:

[0105] ;

[0106] In the formula, zs t is the adaptive scaling factor at the tth update, gn max is the maximum gradient norm, γ is the control factor;

[0107] Apply clipping to ensure that the gradient is within a reasonable range, using the following formula:

[0108] ;

[0109] In the formula, is the gradient of the dth parameter after pruning at the jth update;

[0110] Design an adaptive learning rate; dynamically adjust the learning rate according to the time and set parameters. The learning rate changes periodically during the training process to improve the adaptability of the model to training at different stages. Dynamically adjusting the learning rate helps accelerate the convergence of the model and avoid falling into the local optimum. The formula used is as follows:

[0111] ;

[0112] In the formula, xl t is the adaptive learning rate at the tth update, xl0, xl fd and xl zq They are the initial value of the learning rate, the cycle length of the learning rate change, and the amplitude of the learning rate change;

[0113] Neural network parameter update: update the neural network parameters based on adaptive learning rate and interpolation factor; make parameter update more reasonable and effective, and maintain the stability and coherence of the model; the formula used is as follows:

[0114] ;

[0115] In the formula, θ t is the neural network parameter at the tth update, σ is the interpolation factor, is the smoothing term.

[0116] By performing the above operations, in order to solve the problem that the urinary incontinence monitoring data in the existing urinary incontinence early warning system is complex and it is difficult to accurately extract key features, which increases the difficulty of model evaluation and leads to unstable and inaccurate evaluation results, this scheme designs an adaptive scaling factor based on the total gradient norm and the historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of the neural network parameters, designs an adaptive learning rate, and updates the neural network parameters in combination with the interpolation factor, thereby avoiding training anomalies caused by gradient instability and gradient explosion, making the learning rate better adapt to different training stages, helping to better handle complex data, improving the accuracy and reliability of the urinary incontinence risk assessment model, and improving the quality of life and convenience of users.

[0117] Example 5, see Figure 1 and Figure 3 Based on the above embodiment, this embodiment, the parameter optimization module of the urinary incontinence risk assessment model specifically includes the following contents:

[0118] Initialize the individual position, and adjust the first-order moment decay rate β1, second-order moment decay rate β2, and maximum gradient norm gn in the urinary incontinence risk assessment model max , control factor γ, initial value of learning rate xl0, cycle length of learning rate change xl fd , the magnitude of the learning rate change xl zq and the interpolation factor σ to establish a search space, randomly initialize A individual positions in the search space, use the individual positions as representatives of the parameters of the urinary incontinence risk assessment model, and use the loss function of the urinary incontinence risk assessment model established based on the model parameters as the individual fitness value;

[0119] Design reference locations, including the following:

[0120] Calculate the local search ratio value using the following formula:

[0121] ;

[0122] In the formula, is the local search ratio value at the u-th iteration, is the position of the global optimal individual at the u-th iteration, is the average position of all individuals at the u-th iteration, u is the current iteration number, i is the individual index, rand(·) is the random number generation function, is the position of the i-th individual at the u-th iteration;

[0123] To calculate the local position correction, the formula used is as follows:

[0124] ;

[0125] In the formula, is the local position correction value of the ith individual at the uth iteration, is the position of the global worst individual at the u-th iteration;

[0126] Calculate the query value and generate two different query values ​​for each individual based on the local position correction value and the local search ratio value; different search directions and ranges can be obtained, which increases the diversity and flexibility of the search and helps to find better parameter combinations; the formula used is as follows:

[0127] ;

[0128] ;

[0129] In the formula, and are the first query value and the second query value of the i-th individual at the u-th iteration respectively;

[0130] Calculate two reference positions and generate two different reference positions for each individual based on the query value; guide the individual to move in a more promising direction to improve the efficiency and accuracy of the search; the formula used is as follows:

[0131] ;

[0132] ;

[0133] In the formula, and are the first reference individual and the second reference position of the i-th individual at the u-th iteration, is the position of the first random individual at the uth iteration;

[0134] Calculate the disturbance position and generate a disturbance position for each individual based on the Levy flight strategy and random individuals; increase the randomness and exploration ability of the search to avoid the algorithm falling into the local optimum; the formula used is as follows:

[0135] ;

[0136] ;

[0137] ;

[0138] In the formula, is the perturbed position of the i-th individual at the u-th iteration, a is a random number generated based on the Levy flight strategy, b and p are taken from the normal distribution, , , λ is the shape parameter, is the gamma function, is the standard deviation of the normal distribution, and are the positions of the second and third random individuals at the u-th iteration, is the ceiling function;

[0139] Individual position update: update the positions of all individuals based on the reference position and the disturbance position; enable individuals to move more efficiently in the search space and find a better combination of parameters; the formula used is as follows:

[0140] ;

[0141] In the formula, is the position of the i-th individual at the u+1-th iteration, s1 and s2 are the first and second random values ​​in the range (0, 1), respectively;

[0142] The optimal parameters are determined, the fitness threshold is set in advance, the individual fitness value and the global optimal individual are updated, and when the fitness value corresponding to the global optimal individual is lower than the fitness threshold, the model parameters corresponding to the global optimal individual position are the optimal parameters, and the urinary incontinence risk assessment model is constructed based on the optimal parameters; otherwise, if the maximum number of iterations is reached, the individual position is reinitialized; otherwise, the number of iterations is increased by 1, the reference position and the disturbance position are recalculated, and the individual position is updated.

[0143] By performing the above operations, in order to solve the problem that the existing urinary incontinence early warning system has diversified urinary incontinence monitoring data and is difficult to cope with patient data with various characteristics, resulting in misjudgment of urinary incontinence risk assessment and the assessment results are not applicable to some patients, this scheme generates query values ​​and reference positions based on local position correction values ​​and local search ratio values, generates perturbation positions based on Levy flight strategy and random individuals, and updates the positions of all individuals based on the reference position and perturbation position to determine the optimal parameters, thereby improving the efficiency and accuracy of model parameter optimization, enhancing the adaptability and flexibility of the model, better adapting to the individual differences of different patients, improving the generalization ability of the model, and improving the performance and reliability of the urinary incontinence risk assessment model, which helps to more accurately warn of urinary incontinence risks.

[0144] Example 6, see Figure 1This embodiment is based on the above embodiment. The urinary incontinence real-time warning module collects real-time urinary incontinence monitoring data based on a smart chip; the real-time urinary incontinence monitoring data includes physiological data, user behavior data and health status data. After data preprocessing, the real-time urinary incontinence monitoring data is input into a urinary incontinence risk assessment model, and a risk assessment threshold is pre-set. When the risk assessment level output by the model is higher than the risk assessment threshold, the smart chip sends a warning message.

[0145] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0146] While the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations can be made to the embodiments without departing from the principles and spirit of the invention.

[0147] The present invention and its embodiments are described above, and such description is not restrictive. The drawings show only one embodiment of the present invention, and the actual structure is not limited thereto. In short, if ordinary technicians in the field are inspired by it, without departing from the purpose of the invention, they can design a structure and embodiment similar to the technical solution without creativity, which should belong to the protection scope of the present invention.

Claims

1. Urinary incontinence warning system based on smart chip technology, characterized by: It includes a data acquisition module, a data preprocessing module, a urinary incontinence risk assessment model building module, a urinary incontinence risk assessment model parameter optimization module and a urinary incontinence real-time warning module; The data acquisition module collects historical urinary incontinence monitoring data based on the smart chip; The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction on the collected data; The module for constructing a urinary incontinence risk assessment model designs an adaptive scaling factor based on a total gradient norm and a historical gradient benchmark, uses the adaptive scaling factor to prune the gradient of a neural network parameter, designs an adaptive learning rate, and updates the neural network parameter in combination with an interpolation factor; The urinary incontinence risk assessment model parameter optimization module generates a query value and a reference position based on a local position correction value and a local search ratio value, generates a disturbance position based on a Levy flight strategy and a random individual, and updates the positions of all individuals based on the reference position and the disturbance position to determine the optimal parameters; The urinary incontinence real-time warning module collects real-time urinary incontinence monitoring data based on the smart chip. When the risk assessment level output by the urinary incontinence risk assessment model is higher than the risk assessment threshold, the smart chip sends a warning message; The module for constructing the urinary incontinence risk assessment model includes gradient pruning, which specifically includes the following contents: Calculate the total gradient norm using the following formula: ； In the formula, sf t is the total gradient norm at the tth update, d is the parameter index, D is the total number of neural network parameters, is the gradient of the dth parameter at the tth update, is the square of the L2 norm; Calculate the historical gradient benchmark, which includes the historical gradient mean and historical gradient variance. The formula used is as follows: ； ； In the formula, lz t and lc t They are the historical gradient mean and historical gradient variance at the tth update, g j is the gradient of the loss function to the neural network parameters at the jth update, and t is the update index of the neural network parameters; Design the adaptive scaling factor using the following formula: ； In the formula, zs t is the adaptive scaling factor at the tth update, gn max is the maximum gradient norm, γ is the control factor; Apply trimming using the following formula: ； In the formula, is the gradient of the dth parameter after clipping at the jth update.

2. The urinary incontinence early warning system based on smart chip technology according to claim 1, characterized in that: The module for constructing the urinary incontinence risk assessment model is to construct a neural network, which includes an input layer, a hidden layer and an output layer. The input layer is responsible for receiving data in the risk assessment data set, the hidden layer performs data processing and feature extraction, and the output layer outputs an assessment label. The neural network parameter update specifically includes the following contents: The first and second moments are calculated using the following formula: ； ； ； In the formula, g t is the gradient of the loss function to the neural network parameters at the tth update, is the gradient on the parameter θ, J is the loss function, θ t-1 is the neural network parameter at the t-1th update, m t and m t-1 are the first-order moments at the t-th and t-1-th updates, respectively, and v t and v t-1 are the second-order moments at the t-th and t-1-th updates, β1 and β2 are the first-order moment decay rates and the second-order moment decay rates, Yes t The square of , t is the update index of the neural network parameters; Bias correction, the formula used is as follows: ； ； In the formula, and are the first-order moment and second-order moment after bias correction at the t-th update, and are the t-th power of β1 and β2 respectively; Gradient pruning, designing adaptive scaling factors and pruning the gradients of neural network parameters based on the adaptive scaling factors; Design an adaptive learning rate using the following formula: ； In the formula, xl t is the adaptive learning rate at the tth update, xl0, xl fd and xl zq They are the initial value of the learning rate, the cycle length of the learning rate change, and the amplitude of the learning rate change; Neural network parameter update: Neural network parameters are updated based on adaptive learning rate and interpolation factor. The formula used is as follows: ； In the formula, θ t is the neural network parameter at the tth update, σ is the interpolation factor, is the smoothing term.

3. The urinary incontinence early warning system based on smart chip technology according to claim 1, characterized in that: The urinary incontinence risk assessment model parameter optimization module specifically includes the following contents: Initialize the individual position, and adjust the first-order moment decay rate β1, second-order moment decay rate β2, and maximum gradient norm gn in the urinary incontinence risk assessment model max , control factor γ, initial value of learning rate xl0, cycle length of learning rate change xl fd , the magnitude of the learning rate change xl zq and the interpolation factor σ to establish a search space, randomly initialize A individual positions in the search space, use the individual positions as representatives of the parameters of the urinary incontinence risk assessment model, and use the loss function of the urinary incontinence risk assessment model established based on the model parameters as the individual fitness value; Design reference position; Calculate the disturbance position and generate a disturbance position for each individual based on the Levy flight strategy and random individuals. The formula used is as follows: ； In the formula, is the perturbed position of the i-th individual at the u-th iteration, a is a random number generated based on the Levy flight strategy, and are the positions of the second and third random individuals at the uth iteration, u is the current iteration number, i is the individual index, rand(·) is the random number generation function, is the position of the i-th individual at the u-th iteration; Individual position update: All individual positions are updated based on the reference position and the disturbance position. The formula used is as follows: ； In the formula, is the position of the i-th individual at the u+1-th iteration, s1 and s2 are the first and second random values, respectively. and are the first reference individual and the second reference position of the i-th individual at the u-th iteration, respectively; The optimal parameters are determined, the fitness threshold is pre-set, the individual fitness value and the global optimal individual are updated, and when the fitness value corresponding to the global optimal individual is lower than the fitness threshold, the model parameters corresponding to the global optimal individual position are the optimal parameters, and the urinary incontinence risk assessment model is constructed based on the optimal parameters; otherwise, if the maximum number of iterations is reached, the individual position is reinitialized; Otherwise, the number of iterations is increased by 1 and the reference position and perturbation position are recalculated and individual position updates are performed.

4. The urinary incontinence early warning system based on smart chip technology according to claim 3 is characterized by: The design reference position specifically includes the following contents: Calculate the local search ratio value using the following formula: ； In the formula, is the local search ratio value at the u-th iteration, is the position of the global optimal individual at the u-th iteration, is the average position of all individuals at the uth iteration; To calculate the local position correction, the formula used is as follows: ； In the formula, is the local position correction value of the ith individual at the uth iteration, is the position of the global worst individual at the u-th iteration; Calculate the query value and generate two different query values ​​for each individual based on the local position correction value and the local search ratio value. The formula used is as follows: ； ； In the formula, and are the first query value and the second query value of the i-th individual at the u-th iteration respectively; Calculate two reference positions to generate two different reference positions for each individual based on the query value using the following formula: ； ； In the formula, and are the first reference individual and the second reference position of the i-th individual at the u-th iteration, is the position of the first random individual at the u-th iteration.

5. The urinary incontinence early warning system based on smart chip technology according to claim 1, characterized in that: The urinary incontinence real-time warning module collects real-time urinary incontinence monitoring data based on a smart chip; the real-time urinary incontinence monitoring data includes physiological data, user behavior data and health status data. After data preprocessing, the real-time urinary incontinence monitoring data is input into a urinary incontinence risk assessment model, and a risk assessment threshold is pre-set. When the risk assessment level output by the model is higher than the risk assessment threshold, the smart chip sends a warning message.

6. The urinary incontinence early warning system based on smart chip technology according to claim 1, characterized in that: The data acquisition module collects historical urinary incontinence monitoring data based on a smart chip; the historical urinary incontinence monitoring data includes physiological data, user behavior data, health status data and risk assessment level; the physiological data includes urine volume, urethral pressure and bladder residual urine volume; the user behavior data includes urination habits, water intake and exercise volume; the health status data includes heart rate and respiratory rate; the higher the risk assessment level, the greater the risk of urinary incontinence for the user, and the risk assessment level is used as a data label.

7. The urinary incontinence early warning system based on smart chip technology according to claim 1, characterized in that: The data preprocessing module performs data cleaning, data conversion, data normalization and data set construction on the collected data; data cleaning includes processing missing values, outliers and duplicate values; Data conversion is to convert data into vector form; Data normalization is to unify the data range based on the maximum and minimum normalization method; Building a data set is to build a risk assessment data set based on the processed data, and transmit the risk assessment data set to a cloud server based on the smart chip.

Images