CN118356169B - Automatic monitoring system for medical care - Google Patents
Automatic monitoring system for medical care Download PDFInfo
- Publication number
- CN118356169B CN118356169B CN202410789344.9A CN202410789344A CN118356169B CN 118356169 B CN118356169 B CN 118356169B CN 202410789344 A CN202410789344 A CN 202410789344A CN 118356169 B CN118356169 B CN 118356169B
- Authority
- CN
- China
- Prior art keywords
- dimension
- data
- nursing
- dimensions
- nursing data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000012544 monitoring process Methods 0.000 title claims abstract description 31
- 230000000474 nursing effect Effects 0.000 claims abstract description 116
- 239000013598 vector Substances 0.000 claims abstract description 101
- 238000000034 method Methods 0.000 claims abstract description 34
- 230000002159 abnormal effect Effects 0.000 claims abstract description 22
- 238000004458 analytical method Methods 0.000 claims abstract description 6
- 238000003062 neural network model Methods 0.000 claims description 37
- 238000012549 training Methods 0.000 claims description 21
- 210000002569 neuron Anatomy 0.000 claims description 7
- 238000000513 principal component analysis Methods 0.000 claims description 5
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000036772 blood pressure Effects 0.000 claims description 2
- 238000010606 normalization Methods 0.000 claims description 2
- 238000007405 data analysis Methods 0.000 abstract description 2
- 238000013528 artificial neural network Methods 0.000 description 8
- 238000011156 evaluation Methods 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000758 substrate Substances 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Landscapes
- Measuring And Recording Apparatus For Diagnosis (AREA)
Abstract
The invention relates to the technical field of data analysis, in particular to an automatic medical care monitoring system, which comprises: the data acquisition module is used for acquiring nursing data of all dimensions of a patient at each moment; the multidimensional analysis module is used for acquiring the similarity degree of the trend of each dimension and the direction vector of the main component according to the nursing data of all dimensions of the patient at each moment; the single-dimension module is used for acquiring the characteristic expression degree of each dimension according to the amplitude value and the acquisition time of the nursing data of each dimension and combining the similarity degree of the trend of each dimension and the direction vector of the main component; and the monitoring module is used for acquiring the core dimension according to the characteristic expression of each dimension, and judging the nursing data acquired in real time by utilizing the nursing data in the core dimension. According to the method, the dimension of the multidimensional nursing data is reduced, so that the dimension disaster is avoided, and whether abnormal data exist in the nursing data acquired in real time or not is accurately judged.
Description
Technical Field
The invention relates to the technical field of data analysis, in particular to an automatic monitoring system for medical care.
Background
The medical hospital performs clinical evaluation, treatment plan making, necessary examination and inspection on the patient, and combines the cooperative work of doctors and nursing staff to ensure that the patient obtains optimal medical care, namely, the nursing data acquired in real time is required to be monitored, and the traditional method for automatically monitoring the nursing data of the patient mainly judges whether the nursing data of the patient is abnormal or not through a preset threshold value; however, the method is limited by the value of a preset threshold, and the accuracy of monitoring whether the nursing data is abnormal or not is low in a threshold judgment mode due to the difference of high body diagnosis of patients; therefore, the application provides a method for judging whether nursing data is abnormal based on the SOM neural network algorithm, but because the high-dimensional data increases the distance between samples due to excessive dimensionality of the nursing data in the process of training the SOM neural network, the structure and the similarity between the data are difficult to capture, namely, the dimensionality disaster is caused, and whether the abnormal data exists in the nursing data which are acquired in real time cannot be accurately judged directly through the SOM neural network algorithm.
Disclosure of Invention
The invention provides an automatic medical care monitoring system, which aims to solve the existing problems: by setting the threshold and the direct SOM neural network algorithm, whether abnormal data exist in the nursing data acquired in real time cannot be accurately judged.
The invention relates to an automatic medical care monitoring system, which adopts the following technical scheme:
the method comprises the following modules:
the data acquisition module is used for acquiring nursing data of all dimensions of a patient at each moment;
The multidimensional analysis module is used for acquiring a plurality of main components of the nursing data of all dimensions, a direction vector of each main component and a variance contribution rate of each main component according to the nursing data of all dimensions of the patient at each moment; constructing a sample space according to nursing data of each dimension of a patient, and acquiring the similarity degree of the trend of each dimension and the direction vector of the main component according to the position of the data point in the sample space and combining the direction vector of each main component and the variance contribution rate of each main component;
the single-dimension module is used for acquiring the increment of each piece of nursing data in each dimension according to the amplitude value and the acquisition time of the nursing data in each dimension, classifying each piece of nursing data in each dimension, and obtaining the classification result of each piece of nursing data in each dimension; according to the increment and classification result of each nursing data in each dimension, combining the similarity degree of the trend of each dimension and the direction vector of the main component, and obtaining the feature expression degree of each dimension;
The monitoring module is used for acquiring a core dimension according to the characteristic expression of each dimension and training a neural network model by using nursing data in the core dimension; and monitoring the nursing data acquired in real time according to the neural network model.
Preferably, the method for acquiring the plurality of principal components of the nursing data in all dimensions, the direction vector of each principal component and the variance contribution rate of each principal component according to the nursing data in all dimensions of the patient at each time includes the following specific steps:
and (3) inputting all-dimensional nursing data of the patient at each moment into a PCA principal component analysis algorithm to obtain a plurality of principal components of all-dimensional nursing data, a direction vector of each principal component and a variance contribution rate of each principal component.
Preferably, the method for constructing a sample space according to nursing data of each dimension of a patient and combining a direction vector of each principal component and a variance contribution rate of each principal component according to a position of a data point in the sample space to obtain a similarity degree between a trend of each dimension and the direction vector of the principal component includes the following specific steps:
acquiring the dimension number of nursing data of all dimensions of a patient at each moment, and recording the dimension number as Construct aA dimensional coordinate system for embedding all dimensional nursing data of the patient at each momentObtaining a sample space in a dimensional coordinate system;
For the first Dimension and the firstThe main component is obtainedThe direction vector of each principal component is at the firstProjection vectors in the dimensions are obtained, and projection vectors of the direction vector of each principal component in all the dimensions are obtained; acquiring projections of all data points in a sample space in all dimensions, recorded as projection points in all dimensions, for the firstThe first in the dimensionThe projection points are acquiredThe first in the dimensionA projection point at the firstThe direction vector of each principal component is at the firstProjection vectors on projection vectors in each dimension, denoted as the firstThe first in the dimensionThe first projection pointThe projection vectors are used for acquiring all projection vectors of all projection points in each dimension;
According to the direction vector of each principal component Projection vector in each dimension, the thAll projection vectors of all projection points in each dimension, variance contribution ratio of each principal component, and the thThe number of projection points in each dimension is obtainedThe trend of each dimension is similar to the direction vector of the principal component.
Preferably, the acquiring a firstThe similarity degree of the trend of each dimension and the direction vector of the principal component comprises the following specific calculation formulas:
wherein, Represent the firstThe degree of similarity of the trend of the individual dimensions to the direction vector of the principal component,Represent the firstThe number of proxels in the individual dimensions,Represent the firstThe variance contribution rate of the individual principal components,Represent the firstThe direction vector of each principal component is at the firstThe projection vector in the respective dimensions is calculated,Represent the firstThe first in the dimensionThe first projection pointThe projection vectors; Indicating the number of principal components.
Preferably, the step of acquiring the increment of each piece of care data in each dimension according to the amplitude value and the acquisition time of the care data in each dimension, and classifying each piece of care data in each dimension to obtain a classification result of each piece of care data in each dimension comprises the following specific steps:
For the first In the third dimensionNursing data collected at each moment, will beIn the third dimensionNursing data collected at each moment, subtracting the thIn the third dimensionThe difference obtained from the care data collected at each time is recorded as the firstIn the third dimensionIncrement of nursing data acquired at each moment, obtain the firstIncrement of all care data in each dimension, will beThe incremental average of all the care data in each dimension is recorded as a reference value, and the first dimension is recorded as theThe nursing data with increment smaller than the reference value in each dimension is recorded as first type data, and the first type data is recordedCare data in which the increment in each dimension is greater than the reference value is noted as second class data.
Preferably, the method for obtaining the feature expression degree of each dimension according to the increment and the classification result of each care data in each dimension and combining the similarity degree of the trend of each dimension and the direction vector of the main component includes the following specific steps:
For the first Dimension according to the firstIncrement of all nursing data in each dimension, obtain the firstKurtosis of all increments of care data in the individual dimensions, combined with the firstThe similarity degree of the trend of each dimension and the direction vector of the main component, the quantity of the first type data and the quantity of the second type data are obtained to obtain the first typeFeature representation in each dimension.
Preferably, the acquiring a firstThe feature expression degree of each dimension comprises the following specific calculation formulas:
In the method, in the process of the invention, Represent the firstFeature expression of individual dimensions; Represent the first The degree of similarity of the trend of the individual dimensions with the direction vector of the principal component; Represent the first Kurtosis of increments of all care data in the individual dimensions; representing the amount of the first type of data; representing the amount of the second type of data; Representing a normalization function; Representing an absolute value operation; an exponential function based on a natural constant is represented.
Preferably, the obtaining the core dimension according to the feature expression of each dimension includes the following specific methods:
presetting a expressive threshold value For the firstFeature expression level of each dimension, whenThe feature expressivity of each dimension is larger thanThen (1)And each dimension is a core dimension, and all the core dimensions are acquired.
Preferably, the training the neural network model by using the nursing data in the core dimension comprises the following specific methods:
taking nursing data in a core dimension as an input layer and an output layer of the SOM neural network model as one Square topology network of (2), doTraining for the second time to obtain an SOM neural network model; the saidAnd (3) withRespectively the preset square side length and training times.
Preferably, the monitoring of the nursing data collected in real time according to the neural network model includes the following specific steps:
after the SOM neural network model is obtained, the nursing data of each dimension collected in real time are input into the SOM neural network model, when the output of the SOM neural network model is abnormal neurons, the nursing data collected in real time are abnormal data, and when the output of the SOM neural network model is normal neurons, the nursing data collected in real time are normal data.
The technical scheme of the invention has the beneficial effects that: the invention collects the nursing data of all dimensions of the patient at each moment; according to nursing data of all dimensions of a patient at each moment, acquiring the similarity degree of the trend of each dimension and the direction vector of the main component, wherein the closer the trend of a single dimension is to the direction vector of the main component, the more representative the feature presented by the sign data of the patient is shown, and preparing data for the subsequent acquisition of the core dimension;
According to the amplitude of the nursing data in each dimension and the collection time, the feature expression degree of each dimension is obtained by combining the similarity degree of the trend of each dimension and the direction vector of the main component, and the higher the feature expression degree of the dimension is, the more the dimension is used as the dimension for training the SOM neural network model, so that the core dimension can be obtained according to the feature expression degree of each dimension, the nursing data collected in real time can be judged by utilizing the nursing data in the core dimension, and finally whether abnormal data exists in the nursing data collected in real time can be accurately judged.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a block diagram of an automated medical care monitoring system according to the present invention;
fig. 2 is a flowchart for determining whether abnormal data exists in nursing data collected in real time according to the present invention.
Detailed Description
In order to further describe the technical means and effects adopted by the invention to achieve the preset aim, the following detailed description refers to the specific implementation, structure, characteristics and effects of an automatic medical care monitoring system according to the invention with reference to the accompanying drawings and the preferred embodiments. In the following description, different "one embodiment" or "another embodiment" means that the embodiments are not necessarily the same. Furthermore, the particular features, structures, or characteristics of one or more embodiments may be combined in any suitable manner.
Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.
The following specifically describes a specific scheme of the automatic medical care monitoring system provided by the invention with reference to the accompanying drawings.
Referring to fig. 1, a block diagram of a medical care automatic monitoring system according to an embodiment of the present invention is shown, where the system includes the following modules:
the data acquisition module 101 is used for acquiring nursing data of all dimensions of a patient at each moment.
It should be noted that, by performing clinical evaluations, planning treatment, and performing necessary examinations and checks on patients, medical hospitals combine the cooperation of doctors and caregivers to ensure that patients are optimally medical-care. The present embodiment is a medical care automatic monitoring system, and the final purpose is to perform comprehensive evaluation on the care data of the patient by analyzing the care data of the patient, so that the care data of the patient needs to be collected first.
Specifically, through each item of equipment such as electronic blood pressure measuring instrument, oximeter, dynamic electrocardiograph, etc., each item of nursing data under each moment of patient is gathered to take care data of different items as the nursing data of different dimensionalities, wherein sampling time all sets up to 1 second.
So far, nursing data of all dimensions of the patient at each time point are obtained.
The multidimensional analysis module 102 is configured to obtain, according to nursing data of all dimensions of a patient at each moment, a plurality of principal components of the nursing data of all dimensions, a direction vector of each principal component, and a variance contribution rate of each principal component; and constructing a sample space according to nursing data of each dimension of the patient, and acquiring the similarity degree of the trend of each dimension and the direction vector of the main component according to the position of the data point in the sample space and combining the direction vector of each main component and the variance contribution rate of each main component.
It should be noted that, when the nursing data for automatically monitoring the nursing data of the patient is real-time data, the traditional method for automatically monitoring the nursing data of the patient mainly includes that whether the nursing data of the patient exceeds a threshold value or not is judged by presetting the threshold value and monitoring whether the nursing data of the patient is abnormal or not; however, the method is limited by the value of a preset threshold, and the accuracy of monitoring whether the nursing data is abnormal or not is low in a threshold judgment mode due to the difference of high body diagnosis of patients; therefore, the embodiment provides a method for judging whether nursing data is abnormal based on the SOM neural network algorithm, but because the high-dimensional data increases the distance between samples due to excessive dimensionality of the nursing data in the process of training the SOM neural network, the structure and the similarity between the data are difficult to capture, namely, the dimensionality disaster is caused, and whether the abnormal data exists in the nursing data which are acquired in real time cannot be accurately judged directly through the SOM neural network algorithm.
It should be further noted that, in order to improve accuracy of the training model, in this embodiment, dimension reduction is performed on multidimensional data, so as to obtain a feature expression degree of each dimension compared with a feature difference of a main component, thereby obtaining a core dimension, and an SOM neural network model is built according to the core dimension, so as to avoid causing a dimension disaster, thereby accurately judging whether abnormal data exists in nursing data acquired in real time.
Specifically, nursing data of all dimensions of a patient at each moment is input into a PCA principal component analysis algorithm to obtain each principal component of the nursing data of all dimensions, a direction vector of each principal component and a variance contribution rate of each principal component; as the PCA principal component analysis algorithm is a well-known prior art, principal components, principal component vectors and variance contribution rates thereof can be obtained by the PCA principal component analysis algorithm, which is not described in detail in this embodiment;
then, the number of dimensions in which the care data of all dimensions at each time of the patient is acquired is recorded as Construct aA dimensional coordinate system for embedding all dimensional nursing data of the patient at each momentObtaining a sample space in a dimensional coordinate system;
Finally, for the first Dimension and the firstThe main component is obtainedThe direction vector of each principal component is at the firstProjection vectors in the dimensions are obtained, and projection vectors of the direction vector of each principal component in all the dimensions are obtained; acquiring projections of all data points in a sample space in all dimensions, recorded as projection points in all dimensions, for the firstThe first in the dimensionThe projection points are acquiredThe first in the dimensionA projection point at the firstThe direction vector of each principal component is at the firstProjection vectors on projection vectors in each dimension, denoted as the firstThe first in the dimensionThe first projection pointAnd obtaining all projection vectors of all projection points in each dimension.
In the sample space formed by the nursing data in multiple dimensions, the overall characteristic of the overall trend of each data in all dimensions in the dimensions is the direction vector of the main component, so in the multidimensional data, the projection size of each data in the direction vector of the main component determines the contribution relation of the data value to the direction vector of the main component, the larger the projection is, the more the data approaches to the direction vector of the main component, namely the overall trend of the data, therefore, when the monitored data in the medical care process is required to be evaluated according to the monitoring, the closer the trend of a single dimension is to the direction vector of the main component, the more the characteristic of the physical sign data of a patient is represented, and the model trained in the dimensions can represent the characteristic of the normal training, thereby the characteristic of the abnormal data is represented in isolation. It is therefore necessary to acquire the degree of similarity of the trend of each dimension to the direction vector of the principal component.
Specifically, for the firstA plurality of dimensions, the direction vector of each principal component being at the firstProjection vector in each dimension, the thAll projection vectors of all projection points in each dimension, variance contribution ratio of each principal component, and the thThe number of projection points in each dimension is obtainedThe degree of similarity between the trend of each dimension and the direction vector of the principal component is as follows:
wherein, Represent the firstThe degree of similarity of the trend of the individual dimensions to the direction vector of the principal component,Represent the firstThe number of proxels in the individual dimensions,Represent the firstThe variance contribution rate of the individual principal components,Represent the firstThe direction vector of each principal component is at the firstThe projection vector in the respective dimensions is calculated,Represent the firstThe first in the dimensionThe first projection pointThe projection vectors; Indicating the number of principal components.
It should be noted that the number of the substrates,Representing the firstThe first in the dimensionThe first projection pointProjection vector and the firstThe direction vector of each principal component is at the firstThe ratio of projection vectors in the individual dimensions, thusThe greater the value of (2)The first in the dimensionThe more the data point in the sample space corresponding to the projection point is approaching the firstThe main component is thusThe greater the value of (2)The first main componentThe first in the dimensionThe higher the component content of the data point in the sample space corresponding to the projection point; i.e.The greater the value of (2)The more similar the trend of the individual dimensions is to the direction vector of the principal component.
So far, the similarity degree of the trend of each dimension and the direction vector of the principal component is obtained.
The single-dimension analysis module 103 is configured to obtain an increment of each piece of care data in each dimension according to the amplitude value and the acquisition time of the care data in each dimension, and classify each piece of care data in each dimension to obtain a classification result of each piece of care data in each dimension; and according to the increment and classification result of each nursing data in each dimension, combining the similarity degree of the trend of each dimension and the direction vector of the main component, and acquiring the characteristic expression degree of each dimension.
It should be noted that, in this embodiment, whether the nursing data of the patient is abnormal is determined based on the SOM neural network algorithm, and when the SOM neural network model is trained, the selected dimensions should be mostly stable and normal, and a few have fluctuation or abnormality, when the model is trained by the data of these dimensions, the more easily the stable data is distributed into one or a plurality of neurons gathered together, when all the dimensional data is optimized, the dimensions with the same direction trend as the principal component are further optimized, and the dimensions in which the difference can be obviously exhibited are screened.
Specifically, for the firstIn the third dimensionNursing data collected at each moment, will beIn the third dimensionNursing data collected at each moment, subtracting the thIn the third dimensionThe difference obtained from the care data collected at each time is recorded as the firstIn the third dimensionIncrement of nursing data collected at each moment and make the firstThe increment of the care data acquired at the first moment in the dimension is equal to the firstThe increment of the nursing data acquired at the second moment in the dimension is acquiredIncrement of all care data in each dimension, will beThe incremental average of all the care data in each dimension is recorded as a reference value, and the first dimension is recorded as theThe nursing data with increment smaller than the reference value in each dimension is recorded as first type data, and the first type data is recordedThe nursing data with the increment larger than the reference value in each dimension is recorded as second class data;
According to the first Increment of all nursing data in each dimension, obtain the firstThe kurtosis of all the increments of care data in each dimension, since the specific process of obtaining the kurtosis is known as a prior art, the description is omitted in this embodiment; according to the firstKurtosis, the first of the increments of all care data in each dimensionThe similarity degree of the trend of each dimension and the direction vector of the main component, the quantity of the first type data and the quantity of the second type data are obtained to obtain the first typeThe characteristic expression degree of each dimension is as follows:
In the method, in the process of the invention, Represent the firstFeature expression of individual dimensions; Represent the first The degree of similarity of the trend of the individual dimensions with the direction vector of the principal component; Represent the first Kurtosis of increments of all care data in the individual dimensions; representing the amount of the first type of data; representing the amount of the second type of data; representing a normalizing function, normalizing the object to all dimensions ；Representing an absolute value operation; representing an exponential function based on natural constants, the present embodiment employs The model is used to present the inverse proportional relationship,For model input, the practitioner can set an inverse proportion function according to actual conditions.
It should be noted that the number of the substrates,Representing the firstKurtosis of the increments of all care data in each dimension,The larger the value of (2) is, the description of the (1)The more concentrated the increments of all care data in each dimension, the description of the firstThe overall variation of all care data in each dimension is within a stable range, i.e. the firstThe individual dimensions meet the smooth and normal characteristics, andThe greater the value of (2)The more dimensions should be used as dimensions for training the SOM neural network model; representing the ratio of the first type of data to the second type of data in number when The closer the value of (1) is, the closer the first type data and the second type data are in number, at this time the first type dataThe closer the overall upward and downward trends of the individual dimensions are, i.e. the firstThe individual dimensions satisfy the smooth and normal characteristics, whileRepresenting the firstThe similarity degree of the trend of each dimension and the direction vector of the main component, and the closer the trend of a single dimension and the direction vector of the main component are, the more the trend of the single dimension and the direction vector of the main component are, the more the dimension represents the characteristic presented by the physical sign data of a patient; thus (2)The greater the value of (2)The more dimensions should be as dimensions for training the SOM neural network model.
Thus far, feature expressivity of all dimensions.
The monitoring module 104 is configured to obtain a core dimension according to the feature expression level of each dimension, and train the neural network model by using the care data in the core dimension; and monitoring the nursing data acquired in real time according to the neural network model.
After the feature expressive degrees of all dimensions are obtained by the single-dimension analysis module 103, the dimensions of the training SOM neural network model can be obtained according to the feature expressive degrees of the dimensions.
Specifically, a performance threshold is preset，The specific value of (2) can be set by combining with the actual situation, the embodiment does not have hard requirement, in the embodiment, the method usesTo describe, for the firstFeature expression level of each dimension, whenThe feature expressivity of each dimension is larger thanThen (1)And each dimension is a core dimension, and all the core dimensions are acquired.
The core dimension is a dimension for training the SOM neural network model, and after the core dimension is obtained, the SOM neural network model can be obtained according to the core dimension training.
Specifically, the nursing data in the core dimension is used as an input layer and an output layer of the SOM neural network model are oneSquare topology network of (2), doTraining for the second time to obtain an SOM neural network model; the saidAnd (3) withRespectively a preset square side length and training times,And (3) withThe specific value of (2) can be set by combining with the actual situation, the embodiment does not have hard requirement, in the embodiment, the method uses、The specific training process of the SOM neural network model is described as a well-known prior art, so that the detailed description is omitted in this embodiment;
After the SOM neural network model is obtained, the nursing data of each dimension collected in real time are input into the SOM neural network model, when the output of the SOM neural network model is an abnormal neuron, the nursing data collected in real time are abnormal data, and when the output of the SOM neural network model is a normal neuron, the nursing data collected in real time are normal data, and finally whether the abnormal data exist in the nursing data collected in real time is judged.
In this embodiment, a flowchart for determining whether an abnormality exists in the care data collected in real time is shown in fig. 2.
This embodiment is completed.
The above description is only of the preferred embodiments of the present invention and is not intended to limit the invention, but any modifications, equivalent substitutions, improvements, etc. within the principles of the present invention should be included in the scope of the present invention.
Claims (4)
1. A medical care automatic monitoring system, characterized in that the system comprises the following modules:
The data acquisition module is used for acquiring nursing data of all dimensions of a patient at each moment; the method comprises the following steps: collecting various nursing data of a patient at each moment by various devices such as an electronic blood pressure measuring instrument, an oximeter, a dynamic electrocardiograph and the like, and recording different nursing data as nursing data with different dimensions;
The multidimensional analysis module is used for acquiring a plurality of main components of the nursing data of all dimensions, a direction vector of each main component and a variance contribution rate of each main component according to the nursing data of all dimensions of the patient at each moment; constructing a sample space according to nursing data of each dimension of a patient, and acquiring the similarity degree of the trend of each dimension and the direction vector of the main component according to the position of the data point in the sample space and combining the direction vector of each main component and the variance contribution rate of each main component;
the single-dimension module is used for acquiring the increment of each piece of nursing data in each dimension according to the amplitude value and the acquisition time of the nursing data in each dimension, classifying each piece of nursing data in each dimension, and obtaining the classification result of each piece of nursing data in each dimension; according to the increment and classification result of each nursing data in each dimension, combining the similarity degree of the trend of each dimension and the direction vector of the main component, and obtaining the feature expression degree of each dimension;
The monitoring module is used for acquiring a core dimension according to the characteristic expression of each dimension and training a neural network model by using nursing data in the core dimension; monitoring nursing data acquired in real time according to the neural network model;
The method for constructing a sample space according to nursing data of each dimension of a patient, combining a direction vector of each main component and a variance contribution rate of each main component according to positions of data points in the sample space, and acquiring the similarity degree of a trend of each dimension and the direction vector of the main component comprises the following specific steps:
acquiring the dimension number of nursing data of all dimensions of a patient at each moment, and recording the dimension number as Construct aA dimensional coordinate system for embedding all dimensional nursing data of the patient at each momentObtaining a sample space in a dimensional coordinate system;
For the first Dimension and the firstThe main component is obtainedThe direction vector of each principal component is at the firstProjection vectors in the dimensions are obtained, and projection vectors of the direction vector of each principal component in all the dimensions are obtained; acquiring projections of all data points in a sample space in all dimensions, recorded as projection points in all dimensions, for the firstThe first in the dimensionThe projection points are acquiredThe first in the dimensionA projection point at the firstThe direction vector of each principal component is at the firstProjection vectors on projection vectors in each dimension, denoted as the firstThe first in the dimensionThe first projection pointThe projection vectors are used for acquiring all projection vectors of all projection points in each dimension;
According to the direction vector of each principal component Projection vector in each dimension, the thAll projection vectors of all projection points in each dimension, variance contribution ratio of each principal component, and the thThe number of projection points in each dimension is obtainedThe degree of similarity of the trend of the individual dimensions with the direction vector of the principal component;
the acquisition of the first The similarity degree of the trend of each dimension and the direction vector of the principal component comprises the following specific calculation formulas:
wherein, Represent the firstThe degree of similarity of the trend of the individual dimensions to the direction vector of the principal component,Represent the firstThe number of proxels in the individual dimensions,Represent the firstThe variance contribution rate of the individual principal components,Represent the firstThe direction vector of each principal component is at the firstThe projection vector in the respective dimensions is calculated,Represent the firstThe first in the dimensionThe first projection pointThe projection vectors; Representing the number of principal components;
According to the amplitude value and the acquisition time of the nursing data in each dimension, the increment of each nursing data in each dimension is acquired, and each nursing data in each dimension is classified to obtain a classification result of each nursing data in each dimension, and the specific method comprises the following steps:
For the first In the third dimensionNursing data collected at each moment, will beIn the third dimensionNursing data collected at each moment, subtracting the thIn the third dimensionThe difference obtained from the care data collected at each time is recorded as the firstIn the third dimensionIncrement of nursing data acquired at each moment, obtain the firstIncrement of all care data in each dimension, will beThe incremental average of all the care data in each dimension is recorded as a reference value, and the first dimension is recorded as theThe nursing data with increment smaller than the reference value in each dimension is recorded as first type data, and the first type data is recordedThe nursing data with the increment larger than the reference value in each dimension is recorded as second class data;
according to the increment and classification result of each nursing data in each dimension, combining the similarity degree of the trend of each dimension and the direction vector of the main component, and obtaining the characteristic expression degree of each dimension, wherein the specific method comprises the following steps:
For the first Dimension according to the firstIncrement of all nursing data in each dimension, obtain the firstKurtosis of all increments of care data in the individual dimensions, combined with the firstThe similarity degree of the trend of each dimension and the direction vector of the main component, the quantity of the first type data and the quantity of the second type data are obtained to obtain the first typeFeature expression of individual dimensions;
the acquisition of the first The feature expression degree of each dimension comprises the following specific calculation formulas:
In the method, in the process of the invention, Represent the firstFeature expression of individual dimensions; Represent the first The degree of similarity of the trend of the individual dimensions with the direction vector of the principal component; Represent the first Kurtosis of increments of all care data in the individual dimensions; representing the amount of the first type of data; representing the amount of the second type of data; Representing a normalization function; Representing an absolute value operation; an exponential function based on a natural constant;
The core dimension is obtained according to the feature expression of each dimension, which comprises the following specific methods:
presetting a expressive threshold value For the firstFeature expression level of each dimension, whenThe feature expressivity of each dimension is larger thanThen (1)And each dimension is a core dimension, and all the core dimensions are acquired.
2. The automatic medical care monitoring system according to claim 1, wherein the specific method for acquiring the plurality of principal components of the care data of all dimensions, the direction vector of each principal component and the variance contribution rate of each principal component according to the care data of all dimensions of the patient at each time comprises the following steps:
and (3) inputting all-dimensional nursing data of the patient at each moment into a PCA principal component analysis algorithm to obtain a plurality of principal components of all-dimensional nursing data, a direction vector of each principal component and a variance contribution rate of each principal component.
3. The automatic medical care monitoring system according to claim 1, wherein the training of the neural network model using the care data in the core dimension comprises the following specific methods:
taking nursing data in a core dimension as an input layer and an output layer of the SOM neural network model as one Square topology network of (2), doTraining for the second time to obtain an SOM neural network model; the saidAnd (3) withRespectively the preset square side length and training times.
4. The automatic medical care monitoring system according to claim 1, wherein the method for monitoring the real-time collected care data according to the neural network model comprises the following specific steps:
after the SOM neural network model is obtained, the nursing data of each dimension collected in real time are input into the SOM neural network model, when the output of the SOM neural network model is abnormal neurons, the nursing data collected in real time are abnormal data, and when the output of the SOM neural network model is normal neurons, the nursing data collected in real time are normal data.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410789344.9A CN118356169B (en) | 2024-06-19 | 2024-06-19 | Automatic monitoring system for medical care |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410789344.9A CN118356169B (en) | 2024-06-19 | 2024-06-19 | Automatic monitoring system for medical care |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118356169A CN118356169A (en) | 2024-07-19 |
| CN118356169B true CN118356169B (en) | 2024-08-27 |
Family
ID=91884302
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410789344.9A Active CN118356169B (en) | 2024-06-19 | 2024-06-19 | Automatic monitoring system for medical care |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118356169B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260602A (en) * | 2018-11-15 | 2020-06-09 | 天津大学青岛海洋技术研究院 | Ultrasound image analysis techniques for SSI |
| CN118013382A (en) * | 2024-01-31 | 2024-05-10 | 成都青羊殊德中西医门诊有限公司 | Intelligent medical management system based on neural network |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100183202A1 (en) * | 2006-11-06 | 2010-07-22 | Fujifilm Ri Pharma Co., Ltd. | Brain-image diagnosis supporting method, program, and recording medium |
| US8512240B1 (en) * | 2007-11-14 | 2013-08-20 | Medasense Biometrics Ltd. | System and method for pain monitoring using a multidimensional analysis of physiological signals |
| KR101854258B1 (en) * | 2017-08-29 | 2018-05-03 | 한밭대학교 산학협력단 | Memory optimization method for applying deep learning to pc |
| US11593606B1 (en) * | 2017-10-20 | 2023-02-28 | Brain Trust Innovations I, Llc | System, server and method for predicting adverse events |
| KR20170140127A (en) * | 2017-11-24 | 2017-12-20 | 조선대학교산학협력단 | A method for analyzing brain image using KPCA, LDA and Multi-kernel learning SVM |
| JP7023210B2 (en) * | 2018-10-04 | 2022-02-21 | 日本電信電話株式会社 | Multidimensional data visualization equipment, methods and programs |
| EP3903244A1 (en) * | 2018-12-28 | 2021-11-03 | Telefonaktiebolaget LM Ericsson (publ) | A wireless device, a network node and methods therein for training of a machine learning model |
| KR102154763B1 (en) * | 2020-04-09 | 2020-09-10 | 영산대학교산학협력단 | System and method for predicting the development of dementia using smart devices |
| US11809391B2 (en) * | 2020-09-04 | 2023-11-07 | The Mitre Corporation | Systems and methods for reducing data collection burden |
| WO2023108173A1 (en) * | 2021-12-12 | 2023-06-15 | The General Hospital Corporation | System and method for clinical disorder assessment |
| CN117150283B (en) * | 2023-10-31 | 2024-01-12 | 生态环境部华南环境科学研究所(生态环境部生态环境应急研究所) | Emergency environmental event safety early warning method based on big data analysis |
-
2024
- 2024-06-19 CN CN202410789344.9A patent/CN118356169B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111260602A (en) * | 2018-11-15 | 2020-06-09 | 天津大学青岛海洋技术研究院 | Ultrasound image analysis techniques for SSI |
| CN118013382A (en) * | 2024-01-31 | 2024-05-10 | 成都青羊殊德中西医门诊有限公司 | Intelligent medical management system based on neural network |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118356169A (en) | 2024-07-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10722179B2 (en) | Residual-based monitoring of human health | |
| CN109009017B (en) | Intelligent health monitoring system and data processing method thereof | |
| Chen et al. | A new deep learning framework based on blood pressure range constraint for continuous cuffless BP estimation | |
| CN111248859A (en) | Automatic sleep apnea detection method based on convolutional neural network | |
| CN104462744A (en) | Data quality control method suitable for cardiovascular remote monitoring system | |
| CN117958835B (en) | Electrocardiogram analysis method for postoperative rehabilitation of patient | |
| Guo et al. | DS-CNN: Dual-stream convolutional neural networks-based heart sound classification for wearable devices | |
| CN111081379A (en) | Disease probability decision method and system | |
| Liang et al. | Passive fetal movement recognition approaches using hyperparameter tuned lightgbm model and bayesian optimization | |
| CN110575178B (en) | Diagnosis and monitoring integrated medical system for judging motion state and judging method thereof | |
| CN116098595A (en) | System and method for monitoring and preventing sudden cardiac death and sudden cerebral death | |
| CN116019429B (en) | Health monitoring method, device, equipment and storage medium based on physiological index | |
| CN117393153B (en) | Shock real-time risk early warning and monitoring method and system based on medical internet of things time sequence data and deep learning algorithm | |
| Somathilake et al. | Assessment of fetal and maternal well-being during pregnancy using passive wearable inertial sensor | |
| CN114118386A (en) | Blood pressure estimation method based on attention mechanism and time domain convolution neural network | |
| CN118356169B (en) | Automatic monitoring system for medical care | |
| CN117116475A (en) | Method, system, terminal and storage medium for predicting risk of ischemic cerebral apoplexy | |
| CN116013517A (en) | Method for quantifying health score and health care of old people based on IOT (information of interest) evaluation data | |
| AU2018200093A1 (en) | Residual-based monitoring of human health | |
| CN113208609A (en) | Electrocardio information management system | |
| Ding et al. | A Computer‐Aided Heart Valve Disease Diagnosis System Based on Machine Learning | |
| CN112971791B (en) | Individuation physiological state monitoring and analyzing method and equipment | |
| CN118299075B (en) | Real-time physical sign monitoring system and method for comatose patient | |
| KR102371443B1 (en) | Automatic system of analyzing sleep steps by using artificial intelligence and operating method thereof | |
| Zhang et al. | An Integrated Framework for Assessing the Quality of Non-invasive Fetal Electrocardiography Signals |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |