Elder care model prediction system, method, device and storage medium
Data is collected through the health assessment module and neural network models are used to predict the length of time for elderly people to live in different nursing levels in the future, which solves the problem of inconvenient resource allocation in nursing homes and achieves more accurate resource planning and management.
Patent Information
- Application Number
- CN201911214705.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2019-12-02
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2039-12-02
AI Technical Summary
In the prior art, it is difficult for nursing homes to accurately predict the length of time the elderly will live in different nursing levels in the future, resulting in the inability to allocate resources reasonably, which brings inconvenience to the management of nursing homes.
The health assessment module is used to collect user data, and based on the neural network model, predict the length of the elderly's future living in different nursing levels, and push the prediction results to users to help nursing homes arrange resources reasonably.
Through intelligent prediction of neural network models, the accuracy and efficiency of elderly care model prediction are improved, helping elderly care institutions to allocate resources reasonably, and reducing resource waste and management pressure.
Smart Images

Figure CN112992348B_ABST
Abstract
Description
Background Art
[0002] With the increase of the elderly population in my country and the continuous extension of the average life expectancy of the people, the problem of elderly care is becoming increasingly serious. In response to social problems such as active population mobility and heavy burden of elderly care for only children, the elderly care model has evolved from the traditional single family elderly care to institutional elderly care, community elderly care and other models. Institutional elderly care is a model in which elderly care agencies such as nursing homes, welfare homes, and elderly apartments provide systematic and standardized elderly care services to the elderly. Compared with family elderly care, institutional elderly care is a model of elderly care that obtains daily care services through payment. However, as the resources of elderly care institutions are becoming increasingly scarce, how to reasonably arrange the various resources of elderly care institutions has become a highly concerned issue.
[0003] At present, in order to provide better services to the elderly, nursing homes arrange the elderly to live in different care areas according to the evaluation results of the elderly, and provide different nursing care services. The elderly care resources allocated to different care areas are different. At present, in order to estimate the length of time the elderly will stay in nursing homes, there are two types of solutions: one is to set the residence years of different nursing levels according to custom settings, and the other is to calculate the residence years of each nursing level according to a certain ratio. These two methods have large deviations in both the calculation of the elderly's admission fees and the demand for admission resources, and have no great guiding value for the planning of residential resources, resulting in the inability to reasonably allocate the resources of nursing homes, which has brought certain inconveniences to the management of nursing homes. Summary of the invention
[0004] In response to the problems in the prior art, the purpose of the present invention is to provide a pension model prediction system, method, device and storage medium, which automatically obtain user evaluation data, and based on the user evaluation data and neural network, automatically predict the future length of time the elderly will live in different levels of care and push them to the user, which is conducive to the rational allocation of pension resources.
[0005] The embodiment of the present invention provides a pension mode prediction system, including:
[0006] A health assessment module, used to provide a health assessment interface and collect assessment data input by the user in the health assessment interface;
[0007] The pension mode prediction module is used to integrate the evaluation data into evaluation features in a preset model input format, input the evaluation features into a trained pension mode prediction model, and obtain output pension mode data, wherein the pension mode prediction model is a model built based on a neural network, and the pension mode data includes duration data of each nursing level;
[0008] The display module is used to push the pension model data to the user.
[0009] Optionally, the health assessment interface provided by the health assessment module includes assessment questionnaires of multiple assessment categories, and the health assessment module collects assessment data entered by the user in the health assessment interface, including collecting answer data entered by the user in the health assessment interface, and calculating health assessment scores of multiple assessment categories of the user according to the answer data;
[0010] The assessment feature includes a plurality of assessment feature values, each of which corresponds to a health assessment score;
[0011] The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, including integrating the health evaluation scores in the order of evaluation feature values in the preset model input format to obtain the evaluation features.
[0012] Optionally, the system further includes a prediction model training module, which is used to construct the pension mode prediction model by adopting the following steps:
[0013] The prediction model training module constructs a pension mode prediction model based on an adaptive radial basis neural network;
[0014] The prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level and adds them into a training set;
[0015] The prediction model training module uses the evaluation characteristics of each user and the duration data of each nursing level as a training sample, and uses the training set to train the elderly care mode prediction model.
[0016] Optionally, the prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level, including:
[0017] The prediction model training module obtains historical user data from the historical data management system of the nursing home;
[0018] The prediction model training module selects historical data of users who have left the nursing home from the historical user data, wherein the historical data includes evaluation data of the users and duration data of the users corresponding to each nursing level;
[0019] The prediction model training module integrates the user's evaluation data into evaluation features in a preset model input format.
[0020] Optionally, the prediction model training module constructs a pension mode prediction model based on an adaptive radial basis function neural network, including:
[0021] The prediction model training module determines the number of neurons in the input layer, hidden layer and output layer of the pension mode prediction model, the number of neurons in the input layer is the number M of evaluation feature values in the evaluation feature, the number of neurons in the hidden layer is K, and the number of neurons in the output layer is the number of nursing levels;
[0022] The prediction model training module uses the following formula to construct the connection between neurons in the input layer, hidden layer and output layer of the pension model prediction model:
[0023]
[0024] Among them, t represents the time, w k (t) represents the connection weight between the kth neuron in the hidden layer and the output layer, k∈(1,K), x(t)=[x1(t),x2(t),x3(t),…,x M (t)] represents the input data of the input layer, y(t) represents the actual output data of the pension model prediction model, φ k (x(t)) is the output data of the kth neuron in the hidden layer;
[0025] Among them, the output data φ of the kth neuron in the hidden layer is calculated using the following formula: k (x(t)):
[0026]
[0027] Among them, μ k represents the central value of the kth neuron in the hidden layer, σ k Represents the central width of the kth neuron in the hidden layer.
[0028] Optionally, the prediction model training module uses the training set to train the pension mode prediction model, including sequentially using the following steps to train the pension mode prediction model for each training sample x(t) in the training set:
[0029] Represent the parameters of the radial basis neural network as particles in a particle swarm, and initialize the position and velocity of each particle;
[0030] Input the training sample x(t) into the pension mode prediction model to calculate the fitness value of each particle;
[0031] Calculating the inertia weight of each particle and the global optimal position of the particle according to the fitness value of the particle;
[0032] Update the position and velocity of each particle according to the inertia weight of each particle, the global optimal position of the particle, and the historical optimal position of the particle;
[0033] The optimal network structure is found according to the global optimal position of the particle, and the number of neurons in the hidden layer of the neural network corresponding to each particle is updated according to the difference between the number of neurons of each particle and the number of neurons of the optimal particle.
[0034] Optionally, when the prediction model training module trains the pension mode prediction model, the position and velocity of each particle are initialized using the following formula:
[0035]
[0036]
[0037] Among them, a i represents the position of the i-th particle, i∈(1,s), s represents the total number of particles, μ i,k , σ i,k and w i,k They represent the center value, center width and connection weight of the kth hidden layer neuron in the ith particle, respectively. i represents the number of neurons in the hidden layer represented by the i-th particle, v i represents the velocity of the ith particle, D i represents the dimension of the i-th particle;
[0038] For the input x(t) of the neural network, the fitness value of each particle is calculated using the following formula:
[0039]
[0040] Among them, T represents the number of training samples input to the neural network, y d (t) represents the expected output of the pension model prediction model;
[0041] The prediction model training module uses the following formula to calculate the inertia weight of each particle:
[0042] ω i (t) = S(t) t A i (t)
[0043] Among them, S(t) and A i (t) is calculated using the following formula:
[0044] S(t)=f min (a(t)) / f max (a(t))
[0045] A i (t) = f(g(t)) / f(a i (t))
[0046] Among them, fmin (a(t)),f max (a(t)) are the minimum fitness value and the maximum fitness value at the current moment respectively;
[0047] The prediction model training module uses the following formula to calculate the historical optimal position p of the particle itself i (t):
[0048]
[0049] The prediction model training module uses the following formula to calculate the global optimal position g(t) of the particle:
[0050]
[0051] The prediction model training module uses the following formula to update the position and velocity of each particle:
[0052] v i (t+1)=ω i (t)v i (t)+c1r1(p i (t)-a i (t))+c2r2(g(t)-a i (t))
[0053] a i (t) = a i (t-1)+αv i (t)
[0054] Among them, c1 and c2 are the particle swarm acceleration constants, α is the particle swarm balance weight, r1 and r2 are the particle historical optimal coefficient and the global optimal position coefficient respectively;
[0055] The prediction model training module finds the optimal network structure according to the global optimal position g(t). The number of neurons in the hidden layer of the optimal neural network is K. best , and according to the difference between the number of neurons of each particle and the optimal number of neurons of the particle, the number of neurons in the hidden layer corresponding to each particle is updated as follows:
[0056]
[0057] Optionally, the system further comprises a cost prediction module, which is used to obtain the nursing costs per unit time corresponding to each nursing level from the cost management system, calculate the predicted nursing costs of the user in each nursing level according to the predicted duration of each nursing level, and obtain the total predicted nursing costs of the user by statistics;
[0058] The display module is also used to push the total predicted nursing cost to the user.
[0059] The embodiment of the present invention further provides a method for predicting a pension mode, comprising the following steps:
[0060] The health assessment module provides a health assessment interface and collects assessment data input by the user in the health assessment interface;
[0061] The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, inputs the evaluation features into a trained elderly care model prediction model, and obtains output elderly care model data, wherein the elderly care model prediction model is a model constructed based on a neural network, and the elderly care model data includes duration data of each nursing level;
[0062] The display module pushes the pension model data to the user.
[0063] The embodiment of the present invention further provides a pension mode prediction device, including:
[0064] processor;
[0065] a memory storing executable instructions of the processor;
[0066] Wherein, the processor is configured to perform the steps of the pension model prediction method by executing the executable instructions.
[0067] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, which implements the steps of the pension model prediction method when executed.
[0068] The embodiment of the present invention further provides a pension mode prediction device, including:
[0069] processor;
[0070] a memory storing executable instructions of the processor;
[0071] Wherein, the processor is configured to perform the steps of the pension model prediction method by executing the executable instructions.
[0072] An embodiment of the present invention also provides a computer-readable storage medium for storing a program, which implements the steps of the pension model prediction method when executed.
[0073] It is to be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present disclosure.
[0074] By adopting the present invention, user evaluation data is automatically collected, and a trained neural network model is used to predict the length of time the elderly will live in different levels of care in the future, and the prediction results are pushed to the user, thereby solving the difficult problem of the difficulty in predicting the length of time the elderly will live in different levels of care in the future, and being able to well assist the elderly care community in planning residential resources and planning the admission of different elderly people; the input data of the neural network model is the health assessment data of the elderly, and therefore, the health level of the elderly can be comprehensively considered for intelligent prediction, and compared with the custom residence years method and the proportion allocation method in the prior art, the length of time the elderly will live in different care formats can be more accurately predicted. BRIEF DESCRIPTION OF THE DRAWINGS
[0075] Other features, objectives and advantages of the present invention will become more apparent from a reading of the detailed description of non-limiting embodiments made with reference to the following accompanying drawings.
[0076] Figure 1 is a structural block diagram of a pension model prediction system according to an embodiment of the present invention;
[0077] Figure 2 It is a structural diagram of a pension mode prediction system according to an embodiment of the present invention;
[0078] Figure 3 is a flow chart of a method for predicting a pension model according to an embodiment of the present invention;
[0079] Figure 4 is a structural schematic diagram of a pension mode prediction device according to an embodiment of the present invention;
[0080] Figure 5 It is a schematic diagram of the structure of a computer-readable storage medium according to an embodiment of the present invention. DETAILED DESCRIPTION
[0081] Example embodiments will now be described more fully with reference to the accompanying drawings. Example embodiments may, however, be embodied in many forms and should not be construed as limited to the examples set forth herein.
[0082] In addition, the accompanying drawings are only schematic illustrations of the present disclosure and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. These functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor devices and / or microcontroller devices.
[0083] In order to provide more reasonable check-in services for the elderly, the elderly can undergo various health assessments after check-in, such as functional family assessment, fall history assessment, stand-up and walk time assessment, self-selected walking speed assessment, audio-visual assessment, psychological state assessment, modified Barthel Index assessment, Montreal cognitive assessment, brief cognitive assessment, global decline assessment, etc., to comprehensively assess the health status of the elderly. In order to solve the technical problems in the prior art, an embodiment of the present invention provides a method for predicting a pension model. The predicted pension model here is the length of time the elderly live in areas of various nursing levels. The present invention combines various health assessments of the elderly with a neural network model to intelligently predict the length of time the elderly live in areas of various nursing levels, thereby improving the efficiency and accuracy of pension model prediction and facilitating the rational allocation of resources by pension institutions.
[0084] like Figure 1 and Figure 2 As shown, an embodiment of the present invention provides a pension model prediction system, including:
[0085] The health assessment module M100 is used to provide a health assessment interface and collect assessment data input by the user in the health assessment interface;
[0086] The elderly care model prediction module M200 is used to integrate the evaluation data into evaluation features in a preset model input format, input the evaluation features into a trained elderly care model prediction model, and obtain output elderly care model data, wherein the elderly care model prediction model is a model constructed based on a neural network, and the elderly care model data includes duration data of each nursing level;
[0087] The display module M300 is used to push the elderly care model data to the user. The push here can be to provide an elderly care model data viewing interface for the user to directly click to view, or to send it to the user's terminal for the user to view.
[0088] For each user, the combination of the predicted duration corresponding to each nursing level is the predicted elderly care mode of the user. The nursing level here can be obtained by the nursing level classification method preset in the elderly care institution, and the number of nursing levels can be selected as needed.
[0089] The present invention first automatically collects the user's assessment data through the health assessment module M100, and predicts the nursing level duration in combination with the user's health assessment data, thereby improving the accuracy of the prediction. Furthermore, the present invention uses a trained neural network model through the elderly care mode prediction module M200 to predict the duration of the elderly living in different nursing levels in the future, thereby solving the problem that the duration of the elderly living in different nursing levels in the future is difficult to predict, and improving the efficiency of the elderly care mode prediction. For elderly care institutions, the number of elderly people who will live in different nursing level areas in the future can be known in advance, which can not only reasonably arrange bed resources, but also help elderly care institutions to reasonably arrange the elderly to stay, thereby improving the efficiency of the use of elderly care institutions. The present invention pushes the predicted elderly care mode data to the user through the display module M300, and the user can conveniently view the elderly care mode data.
[0090] In this embodiment, the assessment data includes health assessment scores of multiple assessment categories of the user, and the assessment features include multiple assessment feature values, each of which corresponds to a health assessment score. The assessment categories can be selected from one or more of functional family assessment, fall history assessment, stand-up and walk time assessment, self-selected walking speed assessment, audio-visual assessment, psychological state assessment, modified Barthel index assessment, Montreal cognitive assessment, concise cognitive assessment, and comprehensive decline assessment as needed. The correspondence between the assessment feature value and the health assessment score can be to use the health assessment score directly as the assessment feature value, or to divide the health assessment score into multiple intervals, each interval corresponding to an assessment feature value, for example, the assessment feature value of an assessment category is set to 0, 1, and 2. For a health assessment score of 90 points or more, the corresponding assessment feature value is 0, for a health assessment score of 60 to 90, the corresponding assessment feature value is 1, and for a health assessment score of less than 60, the corresponding assessment feature value is 2, which is more convenient for classifying and counting users.
[0091] The health assessment interface provided by the health assessment module M100 includes assessment questionnaires of multiple assessment categories. The health assessment module M100 collects assessment data entered by the user in the health assessment interface, including collecting answer data entered by the user in the health assessment interface, and calculates the user's health assessment scores of multiple assessment categories based on the answer data.
[0092] The elderly care model prediction module M200 integrates the evaluation data into the evaluation features of the preset model input format, including: integrating the health assessment score in the order of the evaluation feature values in the preset model input format to obtain the evaluation features. For example, the preset model input format includes four evaluation categories: functional family assessment, fall history assessment, stand-up and walk timing assessment, and self-selected walking speed assessment. The evaluation feature values of these four evaluation categories need to be arranged in the sorting order in the preset model input format to form a feature vector of the health assessment score.
[0093] like Figure 2 As shown, an embodiment of the present invention further provides a method for predicting a pension mode, the method comprising:
[0094] S100: The health assessment module provides a health assessment interface and collects assessment data input by the user in the health assessment interface;
[0095] S200: The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, inputs the evaluation features into a trained elderly care model prediction model, and obtains output elderly care model data, wherein the elderly care model prediction model is a model constructed based on a neural network, and the elderly care model data includes duration data of each nursing level;
[0096] S300: The display module pushes the elderly care model data to the user.
[0097] Each step in the elderly care model prediction method of the present invention can be implemented by adopting the functional implementation method of each module in the elderly care model prediction system as mentioned above. For example, step S100 can be implemented by adopting the functional implementation method of the health assessment module M100, step S200 can be implemented by adopting the functional implementation method of the elderly care model prediction module M200, and step S300 can be implemented by adopting the functional implementation method of the display module M300, that is, providing an elderly care model data display interface or pushing the elderly care model data to the user terminal, which will not be elaborated here.
[0098] In this embodiment, the elderly care model prediction system further includes a prediction model training module, and the prediction model training module is used to perform the elderly care model prediction model training step. The elderly care model prediction method further includes S400: the prediction model training module trains the elderly care model prediction model step. Specifically, the elderly care model prediction model training includes the following steps:
[0099] S410: The prediction model training module constructs a pension mode prediction model based on an adaptive radial basis function neural network;
[0100] Radial basis neural network is a feedforward network with good performance, which has the performance of optimal approximation, simple training, fast learning convergence speed and overcoming local minimum problem. It has been proved that radial basis neural network can approximate any continuous function with arbitrary precision. Therefore, it has been widely used in the fields of pattern recognition, nonlinear control and image processing. The basic idea of radial basis neural network is to use radial basis function (RBF) as the "basis" of hidden unit to form the space of hidden layer. The hidden layer transforms the input vector and converts the low-dimensional pattern input data into high-order space, so that the linear inseparable in low-dimensional space is converted into linear separable in high-dimensional space. The present invention can be applied to different populations and different age groups by adopting adaptive radial basis neural network model, which greatly improves the applicability of prediction model.
[0101] S420: the prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level and adds them into a training set, and uses the evaluation features of each user and the duration data of each nursing level as a training sample;
[0102] S430: The prediction model training module uses the training set to train the pension mode prediction model.
[0103] In this embodiment, in step S420, the prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level, including the following steps:
[0104] S421: The prediction model training module obtains historical user data from the historical data management system of the nursing home;
[0105] S422: the prediction model training module selects historical data of users who have left the nursing home from the historical user data, wherein the historical data includes evaluation data of the users and duration data of the users corresponding to each nursing level;
[0106] S423: The prediction model training module integrates the user's evaluation data into evaluation features in a preset model input format.
[0107] Therefore, when training the elderly care model prediction model, the present invention can analyze based on the historical data of users in the elderly care institution, which can more realistically reflect the selection of elderly care models by users of the elderly care institution. If the amount of historical user data stored in the current elderly care institution is insufficient, it can also be further expanded on this basis, for example, by adjusting one or two health assessment feature values in a user's historical data, and correspondingly adjusting the duration of each nursing level in the elderly care model to generate new training samples. Alternatively, it is also possible to collect data of users living in the current elderly care institution, obtain the duration and evaluation characteristics of the existing nursing levels of the users, fill in the duration of the nursing levels that the users are missing, and obtain new training samples.
[0108] In this embodiment, the step S410: the prediction model training module constructs a pension mode prediction model based on an adaptive radial basis function neural network, including the following steps:
[0109] S411: The prediction model training module determines the input and output variables of the neural network model: M groups of historical data are used as the input of the model, and the M+1 to M+6 are used as the output variables, which are the residence time of nursing level 0, nursing level 1, nursing level 2, nursing level 3, nursing level 4, and nursing level 5, respectively. In the current practical application, 10 evaluations are taken as input, that is, the model has 10 inputs and 6 outputs. However, the present invention is not limited to this.
[0110] S412: The prediction model training module determines the number of neurons in the input layer, hidden layer and output layer of the elderly care model prediction model, and determines the connection mode of the neural network MK-6. The number of neurons in the input layer is the number M of evaluation feature values in the evaluation feature, the number of neurons in the hidden layer is K, and the number of neurons in the output layer is the number of nursing levels, which is 6 in this embodiment.
[0111] S413: The prediction model training module assigns values to the parameters of the neural network; assuming that there are T training samples in total, the input of the neural network at time t is x(t) = [x1(t), x2(t), x3(t), x4(t), x5(t), x6(t), x7(t), x8(t), x9(t), x 10 (t)], the expected output of the neural network is denoted as y d (t), the actual output is expressed as y(t), and the following formula is used to construct the connection between the neurons in the input layer, hidden layer and output layer of the pension model prediction model:
[0112]
[0113] Among them, w k (t) represents the connection weight between the kth neuron in the hidden layer and the output layer, k∈(1,K), φk (x(t)) is the output data of the kth neuron in the hidden layer;
[0114] Among them, the output data φ of the kth neuron in the hidden layer is calculated using the following formula: k (x(t)):
[0115]
[0116] Among them, μ k represents the central value of the kth neuron in the hidden layer, σ k Represents the central width of the kth neuron in the hidden layer.
[0117] In this embodiment, the step S430: using the training set to train the pension mode prediction model includes sequentially using the following steps to train the pension mode prediction model for each training sample x(t) in the training set:
[0118] S431: The prediction model training module represents the parameters of the radial basis neural network as particles in a particle swarm, and initializes the position and velocity of each particle;
[0119] S432: the prediction model training module inputs the training sample x(t) into the pension mode prediction model to calculate the fitness value of each particle;
[0120] S433: The prediction model training module calculates the inertia weight of each particle and the global optimal position of the particle according to the fitness value of the particle;
[0121] S434: The prediction model training module updates the position and speed of each particle according to the inertia weight of each particle, the global optimal position of the particle, and the historical optimal position of the particle;
[0122] S435: The prediction model training module finds the optimal network structure according to the global optimal position of the particle, and updates the number of neurons in the hidden layer of the neural network corresponding to each particle according to the difference between the number of neurons of each particle and the number of neurons of the optimal particle.
[0123] Specifically, the step S431 includes: the prediction model training module initializes the particle swarm acceleration constants c1 and c2, c1∈(0,1), c2∈(0,1); sets the particle swarm balance weight α∈[0,1], represents the parameters of the radial basis neural network as particles in the particle swarm, and uses the following formula to initialize the position and speed of each particle:
[0124]
[0125]
[0126] Among them, a i represents the position of the i-th particle, i∈(1,s), s represents the total number of particles, μ i,k , σ i,k and w i,k They represent the center value, center width and connection weight of the kth hidden layer neuron in the ith particle, respectively. i represents the number of neurons in the hidden layer represented by the i-th particle, μ i,k , σ i,k and w i,k The initial value of K is any number between (0,1). i The initial value of v is any positive integer. i represents the velocity of the ith particle, D i represents the dimension of the i-th particle;
[0127] In step S432, for the input x(t) of the neural network, the prediction model training module calculates the fitness value of each particle using the following formula:
[0128]
[0129] Among them, T represents the number of training samples input to the neural network, y d (t) represents the expected output of the pension model prediction model;
[0130] In step S433, the prediction model training module calculates the inertia weight of each particle using the following formula:
[0131] ω i (t) = S(t) t A i (t) (6)
[0132] Among them, S(t) and A i (t) is calculated using the following formula:
[0133] S(t)=f min (a(t)) / f max (a(t)) (7)
[0134] A i (t) = f(g(t)) / f(a i (t)) (8)
[0135] Among them, f min (a(t)),f max (a(t)) are the minimum fitness value and the maximum fitness value at the current moment respectively;
[0136] The prediction model training module uses the following formula to calculate the historical optimal position p of the particle itselfi (t):
[0137]
[0138] The prediction model training module uses the following formula to calculate the global optimal position g(t) of the particle:
[0139]
[0140] In step S434, the prediction model training module updates the position and velocity of each particle using the following formula:
[0141] v i (t+1)=ω i (t)v i (t)+c1r1(p i (t)-a i (t))+c2r2(g(t)-a i (t)) (11)
[0142] a i (t) = a i (t-1)+αv i (t) (12)
[0143] Among them, c1 and c2 are the particle swarm acceleration constants, α is the particle swarm balance weight, r1 and r2 are the particle historical optimal coefficient and the global optimal position coefficient, r1 and r2 are arbitrary numbers between [0,1];
[0144] The step S435 includes: the prediction model training module finds the optimal network structure according to the global optimal position g(t), and the number of neurons in the hidden layer of the optimal neural network is K best , and according to the difference between the number of neurons of each particle and the optimal number of neurons of the particle, the number of neurons in the hidden layer corresponding to each particle is updated as follows:
[0145]
[0146] After the prediction model training module completes the above steps S431 to S435 for a training sample x(t), the next training sample x(t+1) is inputted until all training samples are inputted into the neural network model.
[0147] In this embodiment, the system further includes a cost prediction module, which is used to perform nursing cost prediction using the following steps:
[0148] S510: Obtaining nursing expenses per unit time corresponding to each nursing level from the expense management system;
[0149] S520: Calculate the predicted nursing expenses of the user at each nursing level according to the predicted duration of each nursing level;
[0150] S530: Obtain the total predicted nursing expenses of the user through statistics.
[0151] The display module M300 is also used to push the total predicted nursing expenses predicted by the expense prediction module to the user, so that the user can conveniently view his total predicted nursing expenses, which is convenient for the user to make decisions.
[0152] Therefore, the present invention can not only predict the length of stay of the elderly at different nursing levels in advance, which is very beneficial for guiding institutions to plan bed resources at each nursing level in advance, but also allow the elderly to understand the stay fee in advance, which is of great significance for reasonable stay arrangements. For the elderly, it can not only predict the length of stay of the elderly at different nursing formats, helping the elderly to understand the future stay in advance, but also help the elderly to better calculate the cost and better improve the user experience.
[0153] An embodiment of the present invention also provides a pension pattern prediction device, comprising a processor; and a memory, in which executable instructions of the processor are stored; wherein the processor is configured to execute the steps of the pension pattern prediction method by executing the executable instructions.
[0154] It will be appreciated by those skilled in the art that various aspects of the present invention may be implemented as methods, systems or program products. Therefore, various aspects of the present invention may be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or a combination of hardware and software, which may be collectively referred to herein as "circuits", "modules" or "systems".
[0155] Refer to the following Figure 4 The electronic device 600 according to this embodiment of the present invention is described. Figure 4 The electronic device 600 shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.
[0156] like Figure 4 As shown, the electronic device 600 is in the form of a general computing device. The components of the electronic device 600 may include, but are not limited to: at least one processing unit 610, at least one storage unit 620, a bus 630 connecting different system components (including the storage unit 620 and the processing unit 610), a display unit 640, etc.
[0157] The storage unit stores program codes, which can be executed by the processing unit 610, so that the processing unit 610 executes the steps according to various exemplary embodiments of the present invention described in the electronic prescription circulation processing method section of this specification. For example, the processing unit 610 can execute the following steps: Figure 3 Follow the steps shown in .
[0158] The storage unit 620 may include a readable medium in the form of a volatile storage unit, such as a random access memory unit (RAM) 6201 and / or a cache memory unit 6202 , and may further include a read-only memory unit (ROM) 6203 .
[0159] The storage unit 620 may also include a program / utility 6204 having a set (at least one) of program modules 6205, such program modules 6205 including but not limited to: a processing system, one or more application programs, other program modules and program data, each of which or some combination may include the implementation of a network environment.
[0160] Bus 630 may represent one or more of several types of bus structures, including a memory unit bus or memory unit controller, a peripheral bus, an accelerated graphics port, a processing unit, or a local bus using any of a variety of bus architectures.
[0161] The electronic device 600 may also communicate with one or more external devices 700 (e.g., keyboards, pointing devices, Bluetooth devices, etc.), may also communicate with one or more devices that enable a user to interact with the electronic device 600, and / or communicate with any device that enables the electronic device 600 to communicate with one or more other computing devices (e.g., routers, modems, etc.). Such communication may be performed via an input / output (I / O) interface 650. Furthermore, the electronic device 600 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 660. The network adapter 660 may communicate with other modules of the electronic device 600 via a bus 630. It should be understood that, although not shown in the figure, other hardware and / or software modules may be used in conjunction with the electronic device 600, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.
[0162] The embodiment of the present invention also provides a computer-readable storage medium for storing a program, which implements the steps of the pension model prediction method when executed. In some possible implementations, various aspects of the present invention can also be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps of various exemplary embodiments of the present invention described in the above-mentioned electronic prescription circulation processing method section of this specification.
[0163] refer to Figure 5 As shown, a program product 800 for implementing the above method according to an embodiment of the present invention is described, which can adopt a portable compact disk read-only memory (CD-ROM) and include program code, and can be run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited thereto. In this document, a readable storage medium can be any tangible medium containing or storing a program, which can be used by or in combination with an instruction execution system, an apparatus or a device.
[0164] The program product may use any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
[0165] The computer readable storage medium may include a data signal propagated in a baseband or as part of a carrier wave, wherein a readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. The readable storage medium may also be any readable medium other than a readable storage medium, which may send, propagate, or transmit a program for use by an instruction execution system, an apparatus, or a device or used in combination with it. The program code contained on the readable storage medium may be transmitted with any appropriate medium, including but not limited to wireless, wired, optical cable, RF, etc., or any suitable combination of the above.
[0166] Program code for performing the processes of the present invention may be written in any combination of one or more programming languages, including object-oriented programming languages such as Java, C++, and the like, as well as conventional procedural programming languages such as "C" or similar programming languages. The program code may be executed entirely on the user computing device, partially on the user device, as a separate software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device may be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computing device (e.g., via the Internet using an Internet service provider).
[0167] The elderly care model prediction system, method, device and storage medium provided by the present invention have the following advantages:
[0168] By adopting the present invention, a trained neural network model is used to predict the length of time the elderly will live in different levels of care in the future, which solves the problem that the length of time the elderly will live in different levels of care in the future is difficult to predict, and can well assist the elderly care community in planning residential resources and planning the admission of different elderly people; the input data of the neural network model is the health assessment data of the elderly, so the health level of the elderly can be comprehensively considered for intelligent prediction, and compared with the custom residence years method and the proportion allocation method in the prior art, the length of time the elderly will live in different care formats can be more accurately predicted.
[0169] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the embodiments of the system, device, and computer storage medium, since they are basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0170] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.
Claims
1. A pension model prediction system, characterized in that: include: A health assessment module, used to provide a health assessment interface and collect assessment data input by the user in the health assessment interface; The pension mode prediction module is used to integrate the evaluation data into evaluation features in a preset model input format, input the evaluation features into a trained pension mode prediction model, and obtain output pension mode data, wherein the pension mode prediction model is a model built based on a neural network, and the pension mode data includes duration data of each nursing level; A display module, used to push the elderly care model data to users; The health assessment interface provided by the health assessment module includes assessment questionnaires of multiple assessment categories. The health assessment module collects assessment data entered by the user in the health assessment interface, including collecting answer data entered by the user in the health assessment interface, and calculating health assessment scores of multiple assessment categories of the user according to the answer data; The assessment feature includes a plurality of assessment feature values, each of which corresponds to a health assessment score; The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, including integrating the health evaluation scores in the order of evaluation feature values in the preset model input format to obtain the evaluation features.
2. The pension model prediction system according to claim 1, characterized in that: The system also includes a prediction model training module, which is used to construct the pension mode prediction model by adopting the following steps: The prediction model training module constructs a pension mode prediction model based on an adaptive radial basis neural network; The prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level and adds them into a training set; The prediction model training module uses the evaluation characteristics of each user and the duration data of each nursing level as a training sample, and uses the training set to train the elderly care mode prediction model.
3. The method for predicting the pension model according to claim 2, characterized in that: The prediction model training module collects evaluation features of multiple users and duration data of each user corresponding to each nursing level, including: The prediction model training module obtains historical user data from the historical data management system of the nursing home; The prediction model training module selects historical data of users who have left the nursing home from the historical user data, wherein the historical data includes evaluation data of the users and duration data of the users corresponding to each nursing level; The prediction model training module integrates the user's evaluation data into evaluation features in a preset model input format.
4. The pension model prediction system according to claim 2, characterized in that: The prediction model training module constructs a pension mode prediction model based on an adaptive radial basis neural network, including: The prediction model training module determines the number of neurons in the input layer, hidden layer and output layer of the pension mode prediction model, the number of neurons in the input layer is the number M of evaluation feature values in the evaluation feature, the number of neurons in the hidden layer is K, and the number of neurons in the output layer is the number of nursing levels; The prediction model training module uses the following formula to construct the connection between neurons in the input layer, hidden layer and output layer of the pension model prediction model: Among them, t represents the time, w k (t) represents the connection weight between the kth neuron in the hidden layer and the output layer, k∈(1,K), x(t)=[x1(t),x2(t),x3(t),…,x M (t)] represents the input data of the input layer, y(t) represents the actual output data of the pension model prediction model, φ k (x(t)) is the output data of the kth neuron in the hidden layer; Among them, the output data φ of the kth neuron in the hidden layer is calculated using the following formula: k (x(t)): Among them, μ k represents the central value of the kth neuron in the hidden layer, σ k Represents the central width of the kth neuron in the hidden layer.
5. The pension model prediction system according to claim 4, characterized in that: The prediction model training module uses the training set to train the pension mode prediction model, including sequentially using the following steps to train the pension mode prediction model for each training sample x(t) in the training set: Represent the parameters of the radial basis neural network as particles in a particle swarm, and initialize the position and velocity of each particle; Input the training sample x(t) into the pension mode prediction model to calculate the fitness value of each particle; Calculating the inertia weight of each particle and the global optimal position of the particle according to the fitness value of the particle; Update the position and velocity of each particle according to the inertia weight of each particle, the global optimal position of the particle, and the historical optimal position of the particle; The optimal network structure is found according to the global optimal position of the particle, and the number of neurons in the hidden layer of the neural network corresponding to each particle is updated according to the difference between the number of neurons of each particle and the number of neurons of the optimal particle.
6. The pension model prediction system according to claim 5, characterized in that: When the prediction model training module trains the pension mode prediction model, the position and speed of each particle are initialized using the following formula: Among them, a i represents the position of the i-th particle, i∈(1,s), s represents the total number of particles, μ i,k , σ i,k and w i,k They represent the center value, center width and connection weight of the kth hidden layer neuron in the ith particle, respectively. i represents the number of neurons in the hidden layer represented by the i-th particle, v i represents the velocity of the ith particle, D i represents the dimension of the i-th particle; For the input x(t) of the neural network, the fitness value of each particle is calculated using the following formula: Among them, T represents the number of training samples input to the neural network, y d (t) represents the expected output of the pension model prediction model; The prediction model training module uses the following formula to calculate the inertia weight of each particle: ω i (t)=S(t) t A i (t) Among them, S(t) and A i (t) is calculated using the following formula: S(t)=f min (a(t)) / f max (a(t)) A i (t)=f(g(t)) / f(a i (t)) Among them, f min (a(t)),f max (a(t)) are the minimum fitness value and the maximum fitness value at the current moment respectively; The prediction model training module uses the following formula to calculate the historical optimal position p of the particle itself i (t): The prediction model training module uses the following formula to calculate the global optimal position g(t) of the particle: The prediction model training module uses the following formula to update the position and velocity of each particle: v i (t+1)=ω i (t)v i (t)+c1r1(p i (t)-a i (t))+c2r2(g(t)-a i (t)) a i (t)=a i (t-1)+αv i (t) Among them, c1 and c2 are the particle swarm acceleration constants, α is the particle swarm balance weight, r1 and r2 are the particle historical optimal coefficient and the global optimal position coefficient respectively; The prediction model training module finds the optimal network structure according to the global optimal position g(t). The number of neurons in the hidden layer of the optimal neural network is K. best , and according to the difference between the number of neurons of each particle and the optimal number of neurons of the particle, the number of neurons in the hidden layer corresponding to each particle is updated as follows:
7. The pension model prediction system according to claim 1, characterized in that: The system also includes a cost prediction module, which is used to obtain the nursing costs per unit time corresponding to each nursing level from the cost management system, calculate the predicted nursing costs of the user in each nursing level according to the predicted duration of each nursing level, and obtain the total predicted nursing costs of the user; The display module is also used to push the total predicted nursing cost to the user.
8. A method for predicting a pension model, characterized in that: The steps include: The health assessment module provides a health assessment interface and collects assessment data input by the user in the health assessment interface; The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, inputs the evaluation features into a trained elderly care model prediction model, and obtains output elderly care model data, wherein the elderly care model prediction model is a model constructed based on a neural network, and the elderly care model data includes duration data of each nursing level; The display module pushes the pension model data to the user; The health assessment interface provided by the health assessment module includes assessment questionnaires of multiple assessment categories, and the health assessment module collects assessment data entered by the user in the health assessment interface, including collecting answer data entered by the user in the health assessment interface, and calculating the health assessment scores of the user in multiple assessment categories according to the answer data; The assessment feature includes a plurality of assessment feature values, each of which corresponds to a health assessment score; The elderly care model prediction module integrates the evaluation data into evaluation features in a preset model input format, including integrating the health evaluation scores in the order of evaluation feature values in the preset model input format to obtain the evaluation features.
9. A pension mode prediction device, characterized in that: include: processor; a memory storing executable instructions of the processor; Wherein, the processor is configured to perform the steps of the pension pattern prediction method described in claim 8 by executing the executable instructions.
10. A computer-readable storage medium for storing a program, characterized in that: When the program is executed, the steps of the pension model prediction method described in claim 8 are implemented.
Citation Information
Patent Citations
Management method, device and system of intelligent old-age service cloud platform
CN106169167A
Intelligent big data-based human health feature modeling measurement method and platform
CN107863153A
Medical care assessment conclusion generation method and device and storage medium
CN109726904A
An internet-of-things-based intelligent nursing service system for long-term care elderly people
CN109767371A
Ecological pension suitability quantitative comprehensive analysis system
CN110458440A
