Clinical nursing abnormity monitoring system based on artificial intelligence

Abstract

The invention discloses a clinical nursing abnormity monitoring system based on artificial intelligence. The clinical nursing abnormity monitoring system comprises a data acquisition module, a clinical nursing abnormity monitoring model building module, a model optimization module and an abnormity monitoring module. The invention relates to the technical field of medical monitoring, in particular to a clinical nursing abnormity monitoring system based on artificial intelligence, which is characterized in that by designing an activation function, a typing adjustment layer, a bidirectional memory adjustment layer, a global correction layer and a loss function, the dynamic change of state data of a patient is more comprehensively understood, and the monitoring accuracy is improved. The global information integration capability, the monitoring precision and the abnormal state prediction capability of the model are improved; the model is optimized by defining the parameter search speed, the global leading search and the depth following search, the search process is guided to be carried out in a more potential area, the search process can cover the global range, careful search can be carried out in a key area, and the search efficiency and the global search capability are improved.

Description

Technical Field

[0001] The present invention relates to the field of medical monitoring technology, and in particular to an artificial intelligence-based clinical nursing abnormality monitoring system. Background Art

[0002] An artificial intelligence-based clinical nursing abnormality monitoring system uses artificial intelligence technology and big data processing technology to analyze and monitor patients in real time, and quickly and accurately identifies potential abnormalities through powerful algorithms and models. However, the traditional clinical nursing abnormality monitoring model has problems such as insufficient monitoring accuracy, incomplete capture of dynamic changes in patient status data, and limited model generalization ability; the traditional model parameter optimization method has problems such as low search efficiency, easy to fall into local optimal solutions, and difficulty in balancing global search and local refinement search. 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 clinical nursing abnormality monitoring system based on artificial intelligence. In view of the problems that the traditional clinical nursing abnormality monitoring model has insufficient monitoring accuracy, incomplete capture of the dynamic changes of patient status data, and limited model generalization ability, this scheme improves the model's fine-grained processing and global information integration capabilities of patient status data by designing activation functions, typing adjustment layers, bidirectional memory adjustment layers, global correction layers and loss functions, thereby improving the monitoring accuracy, being able to more comprehensively understand the dynamic changes of patient status data, and improving the ability to predict abnormal states; in view of the problems that the traditional model parameter optimization method has low search efficiency, is easy to fall into local optimal solutions, and is difficult to balance global search and local refinement search, this scheme optimizes the model by preliminarily optimizing parameter settings, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization, and guiding the search process to more potential areas, so that the search process can cover the global range and conduct detailed searches in key areas, avoiding blind searches and repeated searches, improving search efficiency and global search capabilities, and thus optimizing the model.

[0004] The technical solution adopted by the present invention is as follows: A clinical nursing abnormality monitoring system based on artificial intelligence provided by the present invention includes a data acquisition module, a module for building a clinical nursing abnormality monitoring model, a model optimization module and an abnormality monitoring module;

[0005] The data collection module collects historical clinical nursing patients’ vital signs data, laboratory test data, nursing record data, living habit data and patient status data;

[0006] The clinical nursing abnormality monitoring model construction module constructs the clinical nursing abnormality monitoring model by designing the model structure, designing the activation function, designing the typing adjustment layer, designing the bidirectional memory adjustment layer, designing the global correction layer, designing the output layer and loss calculation;

[0007] The model optimization module optimizes the model by preliminary optimization parameter setting, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization;

[0008] The abnormality monitoring module collects clinical nursing patient data in real time and inputs it into the clinical nursing abnormality monitoring model. The model outputs the patient status. When the patient status is abnormal, the system issues an alarm.

[0009] Furthermore, the data acquisition module collects historical clinical nursing patients' vital signs data, laboratory test data, nursing record data, living habit data and patient status data; the vital signs data include the patient's body temperature, heart rate, respiratory rate, blood pressure and blood oxygen saturation; the laboratory test data refers to the patient's biochemical index detection data; the nursing record data refers to the nurse's nursing operation record; the living habit data includes the patient's eating habits, exercise habits, smoking and drinking conditions; the patient status data includes normal state and abnormal state.

[0010] Furthermore, the construction of the clinical nursing abnormality monitoring model module specifically includes the following contents:

[0011] Design the model structure. The clinical nursing abnormality monitoring model includes a classification adjustment layer, a two-way memory adjustment layer, a global correction layer, and an output layer. At the same time, the patient status data is set as the label data of the model.

[0012] Design the activation function, expressed as follows:

[0013] ;

[0014] in, represents the input of the normal mixture activation function, represents the normal mixture activation function, represents taking the logarithm, Indicates taking the absolute value, represents the activation factor, e represents the natural constant, represents the noise intensity, represents a random number that follows a standard normal distribution. represents the sigmoid function;

[0015] Design the classification adjustment layer, as shown below:

[0016] ;

[0017] in, The output value of the table classification adjustment layer, and Respectively represent the weight and input value of the classification adjustment layer, and represents the bias of the classification adjustment layer, represents the hyperbolic tangent function;

[0018] A bidirectional memory adjustment layer is designed. The memory adjustment layer includes a forget adjustment unit, an input unit, a candidate state unit, an update state unit, and an output unit. The bidirectional memory adjustment layer introduces two independent memory adjustment layers, namely a forward memory adjustment layer and a reverse memory adjustment layer. The forward memory adjustment layer generates a forward state; the reverse memory adjustment layer generates a reverse state. Specifically, it includes the following contents:

[0019] Design the forgetting adjustment unit, which is expressed as follows:

[0020] ;

[0021] Where t represents the time index, represents the output value of the forgetting adjustment unit at time t, and Represent the weight and bias of the forgetting adjustment unit, represents the state of the memory adjustment layer at time t-1, represents the input value of the forgetting adjustment unit at time t, represents the state of the memory adjustment layer at time t-2, Indicates taking the L2 norm;

[0022] Design the input unit, expressed as follows:

[0023] ;

[0024] in, represents the output value of the input unit at time t, and denote the weight and bias of the input unit respectively, Represents the input value of the input unit at time t;

[0025] Design candidate state units, expressed as follows:

[0026] ;

[0027] in, represents the output value of the candidate state unit at time t, and Represent the weight and bias of the candidate state unit respectively, Represents the input value of the candidate state unit at time t;

[0028] Design the update status unit, expressed as follows:

[0029] ;

[0030] in, Represents the output value of the updated state unit at time t, represents the element-by-element multiplication symbol, Represents the output value of the candidate state unit at time t-1;

[0031] Design the output unit, expressed as follows:

[0032] ;

[0033] ;

[0034] in, represents the output value of the output unit at time t, and denote the weight and bias of the output unit respectively, represents the input value of the output unit at time t, Represents the state of the memory adjustment layer at time t;

[0035] State merging, combining the forward state and the reverse state to get the complete state, is expressed as follows:

[0036] ;

[0037] in, represents the complete state of the bidirectional memory adjustment layer at time t after the state is merged. and denote the merge weight and merge bias respectively, and Represent the forward state and the reverse state respectively;

[0038] Design a global correction layer, expressed as follows:

[0039] ;

[0040] in, represents the output of the global correction layer, and Represent the weight and bias of the global correction layer respectively, Represents the input value of the global correction layer;

[0041] Design the output layer, expressed as follows:

[0042] ;

[0043] in, represents the output result of the output layer, represents the normalized exponential function, , and denote the weight, input and bias of the output layer respectively, Indicates taking the result with the highest probability value;

[0044] The loss function is designed as follows:

[0045] ;

[0046] in, represents the loss value of the model, j represents the index of the data sample used to train the model, and N represents the total number of data samples used to train the model. represents the true value of the jth data sample, represents the model's predicted value for the jth data sample, represents the loss coordination factor, Indicates square root.

[0047] Furthermore, the model optimization module specifically includes the following contents:

[0048] Preliminary optimization parameter settings were performed to create a parameter space. The parameters to be optimized included activation factor, noise intensity, and loss coordination factor. The inverse of the loss function value of the clinical nursing abnormality monitoring model was set as the parameter optimization efficiency.

[0049] Generate initial parameter points, expressed as follows:

[0050] ;

[0051] Where k represents the index of the generated initial parameter point, represents the position of the kth initial parameter point generated, is a random number ranging from 0.5 to 2. represents the position halfway between the upper and lower bounds of the parameter space, Represents a random number between 0 and 1. represents the upper bound of the parameter space, represents the lower bound of the parameter space;

[0052] Define the parameter search speed, expressed as follows:

[0053] ;

[0054] in, Indicates the current parameter search times. Indicates Parameter search speed during parameter search, Indicates the basic parameter search speed, Indicates the maximum number of parameter searches, represents a random number between -1 and 1, and V represents the expansion factor;

[0055] The global leading search is expressed as follows:

[0056] ;

[0057] in, Indicates The global leading search results in the parameter position. Indicates the parameter position with the highest parameter optimization efficiency, Indicates The global leading search results in the parameter position. Indicates the parameter position with the lowest parameter optimization efficiency;

[0058] Depth follow search, expressed as follows:

[0059] ;

[0060] in, Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a random number that follows a standard normal distribution. Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a symbolic function;

[0061] Collaborative optimization, set the parameter optimization efficiency threshold and the maximum number of parameter searches, generate initial parameter points, calculate the parameter optimization efficiency of the initial parameter points, select the parameter positions with the highest and lowest parameter optimization efficiency, perform global leading search and deep following search in turn, calculate the parameter optimization efficiency of the parameter positions obtained by the search, update the parameter positions with the highest and lowest parameter optimization efficiency, and increase the number of searches by one; if during the search process more than 30% of the parameter positions have a parameter optimization efficiency greater than the parameter optimization efficiency threshold, stop the search and set the parameter with the highest parameter optimization efficiency as the final parameter of the model; if the number of searches is greater than or equal to the maximum number of parameter searches, search again; otherwise, perform the next search.

[0062] Furthermore, the abnormality monitoring module collects the vital signs data, laboratory test data, nursing record data and living habit data of clinical care patients in real time, and inputs the data into the clinical care abnormality monitoring model. The model outputs the patient status at this time. When the patient status is abnormal, the system issues an alarm.

[0063] The beneficial effects achieved by the present invention using the above scheme are as follows:

[0064] (1) In view of the problems of insufficient monitoring accuracy, incomplete capture of dynamic changes in patient status data, and limited model generalization ability of traditional clinical nursing abnormality monitoring models, this scheme improves the model's fine-grained processing and global information integration capabilities of patient status data by designing activation functions, classification adjustment layers, bidirectional memory adjustment layers, global correction layers, and loss functions, thereby improving monitoring accuracy, enabling a more comprehensive understanding of the dynamic changes in patient status data, and improving the ability to predict abnormal conditions.

[0065] (2) In view of the problems of low search efficiency, easy to fall into local optimal solutions and difficulty in balancing global search and local refined search in traditional model parameter optimization methods, this scheme optimizes the model by preliminarily optimizing parameter settings, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization, guiding the search process to more promising areas, so that the search process can cover the global range and conduct detailed searches in key areas, avoiding blind search and repeated search, improving search efficiency and global search capabilities, and thus optimizing the model. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 A schematic diagram of an artificial intelligence-based clinical nursing abnormality monitoring system provided by the present invention;

[0067] Figure 2 A schematic diagram of the module for constructing a clinical nursing abnormality monitoring model;

[0068] Figure 3 Schematic diagram for designing a bidirectional memory adjustment layer;

[0069] Figure 4 Schematic diagram of the model optimization module;

[0070] Figure 5 Schematic diagram of collaborative optimization.

[0071] 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

[0072] 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.

[0073] 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.

[0074] Example 1, see Figure 1 , the present invention provides an artificial intelligence-based clinical nursing abnormality monitoring system, including a data acquisition module, a clinical nursing abnormality monitoring model construction module, a model optimization module and an abnormality monitoring module;

[0075] The data acquisition module collects the vital signs data, laboratory test data, nursing record data, living habits data and patient status data of historical clinical nursing patients; the vital signs data include the patient's body temperature, heart rate, respiratory rate, blood pressure and blood oxygen saturation; the laboratory test data refers to the patient's biochemical index detection data; the nursing record data refers to the nurse's nursing operation record; the living habits data includes the patient's eating habits, exercise habits, smoking and drinking conditions; the patient status data includes normal state and abnormal state, and the data is sent to the module for building a clinical nursing abnormality monitoring model;

[0076] The clinical nursing abnormality monitoring model construction module constructs a clinical nursing abnormality monitoring model by designing a model structure, an activation function, a typing adjustment layer, a bidirectional memory adjustment layer, a global correction layer, an output layer and loss calculation, and sends data to a model optimization module and an abnormality monitoring module;

[0077] The model optimization module optimizes the model by preliminary optimization parameter setting, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization, and sends data to the anomaly monitoring module;

[0078] The abnormality monitoring module collects clinical nursing patient data in real time and inputs it into the clinical nursing abnormality monitoring model. The model outputs the patient status. When the patient status is abnormal, the system issues an alarm.

[0079] Example 2, see Figure 1 , Figure 2 and Figure 3 This embodiment is based on the above embodiment, and the construction of the clinical nursing abnormality monitoring model module specifically includes the following contents:

[0080] Design the model structure. The clinical nursing abnormality monitoring model includes a classification adjustment layer, a two-way memory adjustment layer, a global correction layer, and an output layer. At the same time, the patient status data is set as the label data of the model.

[0081] Design the activation function, expressed as follows:

[0082] ;

[0083] in, represents the input of the normal mixture activation function, represents the normal mixture activation function, represents taking the logarithm, Indicates taking the absolute value, represents the activation factor, e represents the natural constant, represents the noise intensity, represents a random number that follows a standard normal distribution. represents the sigmoid function;

[0084] Design the classification adjustment layer, as shown below:

[0085] ;

[0086] in, The output value of the table classification adjustment layer, and Respectively represent the weight and input value of the classification adjustment layer, and represents the bias of the classification adjustment layer, represents the hyperbolic tangent function;

[0087] A bidirectional memory adjustment layer is designed. The memory adjustment layer includes a forget adjustment unit, an input unit, a candidate state unit, an update state unit, and an output unit. The bidirectional memory adjustment layer introduces two independent memory adjustment layers, namely a forward memory adjustment layer and a reverse memory adjustment layer. The forward memory adjustment layer generates a forward state; the reverse memory adjustment layer generates a reverse state. Specifically, it includes the following contents:

[0088] Design the forgetting adjustment unit, which is expressed as follows:

[0089] ;

[0090] Where t represents the time index, represents the output value of the forgetting adjustment unit at time t, and Represent the weight and bias of the forgetting adjustment unit, represents the state of the memory adjustment layer at time t-1, represents the input value of the forgetting adjustment unit at time t, represents the state of the memory adjustment layer at time t-2, Indicates taking the L2 norm;

[0091] Design the input unit, expressed as follows:

[0092] ;

[0093] in, represents the output value of the input unit at time t, and denote the weight and bias of the input unit respectively, Represents the input value of the input unit at time t;

[0094] Design candidate state units, expressed as follows:

[0095] ;

[0096] in, represents the output value of the candidate state unit at time t, and Represent the weight and bias of the candidate state unit respectively, Represents the input value of the candidate state unit at time t;

[0097] Design the update status unit, expressed as follows:

[0098] ;

[0099] in, Represents the output value of the updated state unit at time t, represents the element-by-element multiplication symbol, Represents the output value of the candidate state unit at time t-1;

[0100] Design the output unit, expressed as follows:

[0101] ;

[0102] ;

[0103] in, represents the output value of the output unit at time t, and denote the weight and bias of the output unit respectively, represents the input value of the output unit at time t, Represents the state of the memory adjustment layer at time t;

[0104] State merging, combining the forward state and the reverse state to get the complete state, is expressed as follows:

[0105] ;

[0106] in, represents the complete state of the bidirectional memory adjustment layer at time t after the state is merged. and denote the merge weight and merge bias respectively, and Represent the forward state and the reverse state respectively;

[0107] Design a global correction layer, expressed as follows:

[0108] ;

[0109] in, represents the output of the global correction layer, and Represent the weight and bias of the global correction layer respectively, Represents the input value of the global correction layer;

[0110] Design the output layer, expressed as follows:

[0111] ;

[0112] in, represents the output result of the output layer, represents the normalized exponential function, , and denote the weight, input and bias of the output layer respectively, Indicates taking the result with the highest probability value;

[0113] The loss function is designed as follows:

[0114] ;

[0115] in, represents the loss value of the model, j represents the index of the data sample used to train the model, and N represents the total number of data samples used to train the model. represents the true value of the jth data sample, represents the model's predicted value for the jth data sample, represents the loss coordination factor, Indicates square root.

[0116] By performing the above operations, in order to address the problems of insufficient monitoring accuracy, incomplete capture of dynamic changes in patient status data, and limited model generalization ability in traditional clinical nursing abnormality monitoring models, this solution improves the model's fine-grained processing and global information integration capabilities of patient status data by designing activation functions, classification adjustment layers, bidirectional memory adjustment layers, global correction layers, and loss functions, thereby improving monitoring accuracy, being able to more comprehensively understand the dynamic changes in patient status data, and improving the ability to predict abnormal conditions.

[0117] Example 3, see Figure 1 , Figure 4 and Figure 5 This embodiment is based on the above embodiment, and the model optimization module specifically includes the following contents:

[0118] Preliminary optimization parameter settings were performed to create a parameter space. The parameters to be optimized included activation factor, noise intensity, and loss coordination factor. The inverse of the loss function value of the clinical nursing abnormality monitoring model was set as the parameter optimization efficiency.

[0119] Generate initial parameter points, expressed as follows:

[0120] ;

[0121] Where k represents the index of the generated initial parameter point, represents the position of the kth initial parameter point generated, is a random number ranging from 0.5 to 2. represents the position halfway between the upper and lower bounds of the parameter space, Represents a random number between 0 and 1. represents the upper bound of the parameter space, represents the lower bound of the parameter space;

[0122] Define the parameter search speed, expressed as follows:

[0123] ;

[0124] in, Indicates the current parameter search times. Indicates Parameter search speed during parameter search, Indicates the basic parameter search speed, Indicates the maximum number of parameter searches, represents a random number between -1 and 1, and V represents the expansion factor;

[0125] The global leading search is expressed as follows:

[0126] ;

[0127] in, Indicates The global leading search results in the parameter position. Indicates the parameter position with the highest parameter optimization efficiency, Indicates The global leading search results in the parameter position. Indicates the parameter position with the lowest parameter optimization efficiency;

[0128] Depth follow search, expressed as follows:

[0129] ;

[0130] in, Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a random number that follows a standard normal distribution. Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a symbolic function;

[0131] Collaborative optimization, set the parameter optimization efficiency threshold and the maximum number of parameter searches, generate initial parameter points, calculate the parameter optimization efficiency of the initial parameter points, select the parameter positions with the highest and lowest parameter optimization efficiency, perform global leading search and deep following search in turn, calculate the parameter optimization efficiency of the parameter positions obtained by the search, update the parameter positions with the highest and lowest parameter optimization efficiency, and increase the number of searches by one; if during the search process more than 30% of the parameter positions have a parameter optimization efficiency greater than the parameter optimization efficiency threshold, stop the search and set the parameter with the highest parameter optimization efficiency as the final parameter of the model; if the number of searches is greater than or equal to the maximum number of parameter searches, search again; otherwise, perform the next search.

[0132] By performing the above operations, in view of the problems of low search efficiency, easy to fall into local optimal solutions and difficulty in balancing global search and local refined search in traditional model parameter optimization methods, this solution optimizes the model by preliminarily optimizing parameter settings, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization, guiding the search process to more potential areas, so that the search process can cover the global range and conduct detailed searches in key areas, avoiding blind searches and repeated searches, improving search efficiency and global search capabilities, and thus optimizing the model.

[0133] Example 4, see Figure 1 This embodiment is based on the above embodiment. The abnormal monitoring module collects the vital signs data, laboratory test data, nursing record data and living habit data of clinical care patients in real time, and inputs the data into the clinical care abnormality monitoring model. The model outputs the patient status at this time. When the patient status is abnormal, the system issues an alarm.

[0134] 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.

[0135] 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.

[0136] 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. A clinical nursing abnormality monitoring system based on artificial intelligence, characterized by: It includes data collection module, clinical nursing abnormality monitoring model building module, model optimization module and abnormality monitoring module; The data collection module collects historical clinical nursing patients’ vital signs data, laboratory test data, nursing record data, living habit data and patient status data; The clinical nursing abnormality monitoring model construction module constructs the clinical nursing abnormality monitoring model by designing the model structure, designing the activation function, designing the typing adjustment layer, designing the bidirectional memory adjustment layer, designing the global correction layer, designing the output layer and loss calculation; The model optimization module optimizes the model by preliminary optimization parameter setting, generating initial parameter points, defining parameter search speed, global leading search, deep following search and collaborative optimization; The abnormality monitoring module collects clinical nursing patient data in real time and inputs it into the clinical nursing abnormality monitoring model. The model outputs the patient status. When the patient status is abnormal, the system issues an alarm.

2. According to claim 1, a clinical nursing abnormality monitoring system based on artificial intelligence is characterized by: The construction of the clinical nursing abnormality monitoring model module specifically includes the following contents: Design the model structure. The clinical nursing abnormality monitoring model includes a classification adjustment layer, a two-way memory adjustment layer, a global correction layer, and an output layer. At the same time, the patient status data is set as the label data of the model. Design the activation function, expressed as follows: ； in, represents the input of the normal mixture activation function, represents the normal mixture activation function, represents taking the logarithm, Indicates taking the absolute value, represents the activation factor, e represents the natural constant, represents the noise intensity, represents a random number that follows a standard normal distribution. represents the sigmoid function; Design the classification adjustment layer, as shown below: ； in, The output value of the table classification adjustment layer, and Respectively represent the weight and input value of the classification adjustment layer, and represents the bias of the classification adjustment layer, represents the hyperbolic tangent function; A bidirectional memory adjustment layer is designed. The memory adjustment layer includes a forget adjustment unit, an input unit, a candidate state unit, an update state unit, and an output unit. The bidirectional memory adjustment layer introduces two independent memory adjustment layers, namely a forward memory adjustment layer and a reverse memory adjustment layer. The forward memory adjustment layer generates a forward state; the reverse memory adjustment layer generates a reverse state. Specifically, it includes the following contents: Design the forgetting adjustment unit, which is expressed as follows: ； Where t represents the time index, represents the output value of the forgetting adjustment unit at time t, and Represent the weight and bias of the forgetting adjustment unit, represents the state of the memory adjustment layer at time t-1, represents the input value of the forgetting adjustment unit at time t, represents the state of the memory adjustment layer at time t-2, Indicates taking the L2 norm; Design the input unit, expressed as follows: ； in, represents the output value of the input unit at time t, and denote the weight and bias of the input unit respectively, Represents the input value of the input unit at time t; Design candidate state units, expressed as follows: ； in, represents the output value of the candidate state unit at time t, and Represent the weight and bias of the candidate state unit respectively, Represents the input value of the candidate state unit at time t; Design the update status unit, expressed as follows: ； in, Represents the output value of the updated state unit at time t, represents the element-by-element multiplication symbol, Represents the output value of the candidate state unit at time t-1; Design the output unit, expressed as follows: ； ； in, represents the output value of the output unit at time t, and denote the weight and bias of the output unit respectively, represents the input value of the output unit at time t, Represents the state of the memory adjustment layer at time t; State merging, combining the forward state and the reverse state to get the complete state, is expressed as follows: ； in, represents the complete state of the bidirectional memory adjustment layer at time t after the state is merged. and denote the merge weight and merge bias respectively, and Represent the forward state and the reverse state respectively; Design a global correction layer, expressed as follows: ； in, represents the output of the global correction layer, and Represent the weight and bias of the global correction layer respectively, Represents the input value of the global correction layer; Design the output layer, expressed as follows: ； in, represents the output result of the output layer, represents the normalized exponential function, , and denote the weight, input and bias of the output layer respectively, Indicates taking the result with the highest probability value; The loss function is designed as follows: ； in, represents the loss value of the model, j represents the index of the data sample used to train the model, and N represents the total number of data samples used to train the model. represents the true value of the jth data sample, represents the model's predicted value for the jth data sample, represents the loss coordination factor, Indicates square root.

3. The artificial intelligence-based clinical nursing abnormality monitoring system according to claim 1, characterized in that: The model optimization module specifically includes the following contents: Preliminary optimization parameter settings were performed to create a parameter space. The parameters to be optimized included activation factor, noise intensity, and loss coordination factor. The inverse of the loss function value of the clinical nursing abnormality monitoring model was set as the parameter optimization efficiency. Generate initial parameter points, expressed as follows: ； Among them, k represents the index of the generated initial parameter point, represents the position of the kth initial parameter point generated, is a random number ranging from 0.5 to 2. represents the position halfway between the upper and lower bounds of the parameter space, Represents a random number between 0 and 1. represents the upper bound of the parameter space, represents the lower bound of the parameter space; Define the parameter search speed, expressed as follows: ； in, Indicates the current parameter search times. Indicates Parameter search speed during parameter search, Indicates the basic parameter search speed, Indicates the maximum number of parameter searches, represents a random number between -1 and 1, and V represents the expansion factor; The global leading search is expressed as follows: ； in, Indicates The global leading search results in the parameter position. Indicates the parameter position with the highest parameter optimization efficiency, Indicates The global leading search results in the parameter position. Indicates the parameter position with the lowest parameter optimization efficiency; Depth follow search, expressed as follows: ； in, Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a random number that follows a standard normal distribution. Indicates During the secondary parameter search, the depth follows the parameter position obtained by the search. represents a symbolic function; Collaborative optimization, set the parameter optimization efficiency threshold and the maximum number of parameter searches, generate initial parameter points, calculate the parameter optimization efficiency of the initial parameter points, select the parameter positions with the highest and lowest parameter optimization efficiency, perform global leading search and deep following search in turn, calculate the parameter optimization efficiency of the parameter positions obtained by the search, update the parameter positions with the highest and lowest parameter optimization efficiency, and increase the number of searches by one; if during the search process more than 30% of the parameter positions have a parameter optimization efficiency greater than the parameter optimization efficiency threshold, stop the search and set the parameter with the highest parameter optimization efficiency as the final parameter of the model; if the number of searches is greater than or equal to the maximum number of parameter searches, search again; otherwise, perform the next search.

4. The artificial intelligence-based clinical nursing abnormality monitoring system according to claim 1, characterized in that: The data acquisition module collects the vital signs data, laboratory test data, nursing record data, living habits data and patient status data of historical clinical nursing patients; the vital signs data include the patient's body temperature, heart rate, respiratory rate, blood pressure and blood oxygen saturation; the laboratory test data refers to the patient's biochemical index detection data; the nursing record data refers to the nurse's nursing operation record; the living habits data includes the patient's eating habits, exercise habits, smoking and drinking conditions; the patient status data includes normal status and abnormal status; The abnormality monitoring module collects the vital signs data, laboratory test data, nursing record data and living habit data of clinical nursing patients in real time, and inputs the data into the clinical nursing abnormality monitoring model. The model outputs the patient status at that time. When the patient status is abnormal, the system issues an alarm.

Images