Medical equipment monitoring system based on Internet of Things

Abstract

The invention discloses a medical equipment monitoring system based on Internet of Things, which belongs to the technical field of medical equipment monitoring and comprises a data acquisition module, a data processing module, a feature extraction module, a model training module, a monitoring alarm module, a data storage module and an Internet of Things management module. According to the invention, comprehensive, real-time and intelligent monitoring of the operation state of the medical equipment is realized, so that the efficiency and safety of medical equipment management are remarkably improved; the problem that it is difficult to carry out real-time and accurate operation risk early warning on the monitored medical equipment is solved.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical equipment monitoring, and in particular relates to a medical equipment monitoring system based on the Internet of Things. Background Art

[0002] Modern medical equipment is a fusion of high-sensitivity sensors, large-scale integrated circuits, precision machinery manufacturing and optical imaging. The application of these devices improves the diagnosis and treatment effects, but also increases the risk of equipment use. Based on the Internet of Things technology, real-time monitoring of the operating status of medical equipment can effectively ensure the smooth progress of medical work.

[0003] At present, there is still a large gap between medical equipment monitoring based on the Internet of Things and medical equipment management in hospitals. At present, the medical equipment operation management concept of most hospitals is still relatively backward, and the focus is still on medical equipment failure repair and equipment maintenance. The current medical equipment monitoring is mainly used for post-fault repair and maintenance of medical equipment. The problem is that after the equipment failure has been detected, passive repair is carried out. Passive fault monitoring cannot effectively prevent the risks brought to the diagnosis and treatment of patients due to equipment failure downtime. Summary of the invention

[0004] In view of the above-mentioned deficiencies in the prior art, the present invention provides a medical equipment monitoring system based on the Internet of Things, which solves the problem of difficulty in providing real-time and accurate operational risk warnings for monitored medical equipment.

[0005] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0006] The present invention provides a medical device monitoring system based on the Internet of Things, comprising:

[0007] A data acquisition module, used to obtain real-time operation data and positioning data of medical equipment;

[0008] A data processing module, used for preprocessing the historical operation data of several medical devices to obtain the preprocessed operation data;

[0009] A feature extraction module is used to extract features from the preprocessed operation data to obtain a feature set for predicting equipment operation risks;

[0010] A model training module is used to train a support vector machine based on a device operation risk prediction feature set to obtain a trained support vector machine;

[0011] The monitoring and alarm module is used to use the trained support vector machine to conduct medical equipment operation risk warning for the real-time collected operation data of the medical equipment according to the preset risk level threshold, and obtain the real-time operation risk warning information of the medical equipment;

[0012] A data storage module, used to store the operation data, operation risk warning information and positioning data of the medical equipment in the cloud according to a preset time interval;

[0013] The Internet of Things management module is used to build an Internet of Things data management platform to visualize the real-time operating data, operating risk warning information and positioning data of different medical devices.

[0014] The beneficial effects of the present invention are as follows: the present invention provides a medical equipment monitoring system based on the Internet of Things, which comprehensively collects the operation data of medical equipment, obtains the high-importance medical equipment operation risk features based on data preprocessing and feature extraction, and constitutes the equipment operation risk prediction feature set, and trains the multi-classification support vector machine based on the equipment operation risk prediction feature set to obtain the trained support vector machine that can accurately predict the operation risk of medical equipment from multiple aspects, so that the medical equipment operation risk warning can be realized for the operation data of the medical equipment collected in real time, and the real-time operation risk warning information of the medical equipment can be obtained; the present invention also stores the operation data, operation risk warning information and positioning data of the medical equipment in the cloud according to the preset time interval, which is convenient for fault analysis and subsequent system update, so as to improve the prediction accuracy of the operation risk of the medical equipment; the present invention can also visualize the real-time operation data, operation risk warning information and positioning data of different medical equipment through the Internet of Things data management platform, so as to facilitate the timely and rapid discovery of the operation risk and fault of the medical equipment; the present invention realizes the comprehensive, real-time and intelligent monitoring of the operation status of the medical equipment, which can significantly improve the efficiency and safety of medical equipment management.

[0015] Furthermore, the operation data of the medical device includes operation parameter data, usage record data, fault record data, maintenance record data, working environment data and consumables usage data.

[0016] The beneficial effect of adopting the above further scheme is: the present invention conducts all-round operation status monitoring of medical equipment based on the operation data of different types of medical equipment, and can further ensure the accuracy of predicting the operation risk of medical equipment based on the relationship between different key types of operation data.

[0017] Furthermore, the operating parameter data, usage record data, fault record data, maintenance record data, working environment data, consumable usage data and positioning data all have time tags on the data of each dimension.

[0018] The beneficial effect of adopting the above further scheme is: the present invention adds time tags to the location data of various types of operation data sets, which is convenient for ensuring data consistency when visualizing data through the Internet of Things data management platform, and is convenient for quickly and accurately identifying operation risks and faults.

[0019] Furthermore, the calculation expression of the operation data of the medical device is as follows:

[0020] {x k (n)}={x j (i),1≤j≤k,1≤i≤n},

[0021] Among them, x k (n) represents the medical equipment operation data, x j (i) represents the i-th dimension data corresponding to the j-th type of medical equipment operation data, k represents the total number of types of medical equipment operation data, n represents the total number of dimensions of medical equipment operation data, i = 1, 2,…, t,…, n.

[0022] The beneficial effect of adopting the above further scheme is: the present invention provides a method for constructing operating data of medical equipment, which can comprehensively reflect the operating conditions of medical equipment from different types and dimensions, thereby providing a basis for accurately predicting various aspects of operating risks of medical equipment.

[0023] Furthermore, the data processing module includes:

[0024] A first preprocessing submodule, used to filter out random noise from the operation data of the medical device to obtain first preprocessed operation data;

[0025] The calculation expression of the first preprocessing operation data is as follows:

[0026]

[0027] Among them, x 1 (t) represents the first preprocessed operating data, and x(i) represents the i-th dimension data in the operating data of the medical device;

[0028] A second preprocessing submodule is used for standardizing the first preprocessing operation data to obtain second preprocessing operation data;

[0029] The calculation expression of the second preprocessing operation data is as follows:

[0030]

[0031] Among them, x 2 (t) represents the second pre-processing operation data, x min represents the maximum data value in the second preprocessing run data, x maxrepresents the minimum data value in the second preprocessing operation data;

[0032] A third preprocessing submodule is used to perform equal-width discretization processing on the second preprocessing running data based on a discrete model to obtain third preprocessing running data, and to set a category label for each category interval in the third preprocessing running data;

[0033] The calculation expression of the discrete model is as follows:

[0034]

[0035] Wherein, τ represents the category label of the τth category interval in the third preprocessing operation data, and ω represents the category interval width;

[0036] A fourth preprocessing module, used for performing missing value supplementation processing on the third preprocessing operation data based on a missing value model to obtain fourth preprocessing operation data;

[0037] The calculation expression of the missing value model is as follows:

[0038]

[0039] in, indicates missing data in the i-th dimension, Represents the previous adjacent dimension data of the missing data of the i-th dimension, Represents the data of the adjacent dimension after the missing data of the i-th dimension;

[0040] a fifth preprocessing module, configured to perform outlier processing on the fourth preprocessed operating data based on an outlier processing model to obtain preprocessed operating data;

[0041] The calculation expression of the outlier processing model is as follows:

[0042]

[0043] Among them, z i (x 4 ) represents the function of performing Z-score abnormal monitoring on the fourth preprocessing operation data, x 4 (i) represents the fourth preprocessing operation data, μ represents the mean of the fourth preprocessing operation data, and δ represents the standard deviation of the fourth preprocessing operation data.

[0044] The beneficial effect of adopting the above further scheme is: the present invention performs comprehensive and systematic preprocessing of the operating data of medical equipment through noise filtering, data standardization, discretization processing, missing value supplementation and outlier processing, which effectively improves the quality and availability of the data and provides a basis for improving feature extraction and model training effects.

[0045] Furthermore, the feature extraction module includes:

[0046] The first feature extraction submodule is used to extract the operation risk warning features of several medical devices based on the long short-term memory network according to the preprocessed operation data;

[0047] The second feature extraction submodule is used to evaluate the importance of the operation risk warning features of each medical device through the random forest method to obtain the feature average importance value of the operation risk warning features;

[0048] The third feature extraction submodule is used to select, according to a resampling method, a number of operation risk warning features whose average feature importance values ​​are greater than a preset importance threshold as important operation risk warning features, and obtain a first operation risk feature set consisting of the important operation risk warning features;

[0049] The fourth feature extraction submodule is used to repeatedly eliminate several important operation risk warning features with the lowest importance in the first operation risk feature set according to the random Mori recursive feature elimination method for several times, and obtain a corresponding second operation risk feature set after each elimination;

[0050] a fifth feature extraction submodule, for performing a medical equipment operation risk early warning detection test using each second operation risk feature set by using a random forest method, and taking the explanation coefficient corresponding to each second operation risk feature set as its average accuracy;

[0051] The sixth feature extraction submodule is used to select the second operation risk feature set corresponding to the highest average accuracy as the equipment operation risk prediction feature set.

[0052] The beneficial effect of adopting the above further scheme is as follows: the present invention realizes efficient extraction and optimization of medical equipment operation risk warning features by extracting dependencies in preprocessed operation data, evaluating importance, screening out low-quality features, testing evaluation accuracy and selecting the second operation risk feature set with the highest accuracy, thereby improving the accuracy and stability of medical equipment operation risk warning.

[0053] Furthermore, the model training module includes:

[0054] The first model training submodule is used to input the equipment operation risk prediction feature set into the support vector machine, train the support vector machine, and obtain the optimal normal vector and intercept of the hyperplane;

[0055] The second model training submodule is used to obtain a trained support vector machine based on an optimal hyperplane normal vector and a hyperplane intercept.

[0056] Furthermore, the first model training submodule includes:

[0057] The first model training unit is used to construct the support vector machine for the medical equipment operation risk warning problem:

[0058]

[0059] in, represents the minimization function based on the hyperplane normal vector, hyperplane intercept and slack variables, W k′ represents the hyperplane normal vector corresponding to the k′th risk warning category, ||·|| 2 represents the square of the norm, C represents the first regularization parameter, ξ i′,k′ represents the slack variable related to the data of the corresponding dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set and the k′th risk warning category, st represents the slack variable that makes y i′,k′ Indicates the indicator variable associated with the feature and risk warning category, represents the transpose of the hyperplane normal vector corresponding to the k′th risk warning category, X i′ represents the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, b k′ represents the hyperplane intercept corresponding to the k′th risk warning category, represents any selection, where y i′ = k′, the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set belongs to the k′th risk warning category, y i′ ≠k′, the data of the dimension corresponding to the i′th important operation risk warning feature in the equipment operation risk prediction feature set does not belong to the k′th risk warning category;

[0060] The second model training unit is used to construct a decision boundary maximization interval based on the medical equipment operation risk warning problem, so as to identify the risk warning category to which the data of the corresponding dimension of each important operation risk warning feature belongs;

[0061] The third model training unit is used to construct the loss function of the support vector machine:

[0062]

[0063] Among them, L csvm represents the loss function of the support vector machine, y i′ represents the true category label of the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, max(·) represents the maximum value function, and s j represents the score of the j-th risk warning category, represents the score of the true category, Δ represents the positive interval parameter, λ represents the second regularization parameter, and R(W) represents the regularization term of the hyperplane normal vector;

[0064] The fourth model training unit is used to input important operation risk warning features in the equipment operation risk prediction feature set into the support vector machine with the goal of minimizing the loss function of the support vector machine, train the support vector machine, and obtain the optimal hyperplane normal vector and hyperplane intercept.

[0065] The beneficial effect of adopting the above-mentioned further scheme is as follows: the present invention constructs a support vector machine for the medical equipment operation risk warning problem, and trains a multi-classification support vector machine based on the equipment operation risk prediction feature set, thereby realizing early warning of different aspects of operation risks of the operation data of the medical equipment based on different optimal hyperplane normal vectors and hyperplane intercepts.

[0066] Furthermore, the IoT data management platform matches the real-time operating data, operating risk warning information and positioning data of each medical device based on time tags, and displays them visually.

[0067] The beneficial effect of adopting the above further scheme is: after matching through time tags, the time consistency of the information displayed by the Internet of Things data management platform can be guaranteed, and while ensuring the accuracy of the information, it is convenient to observe the medical equipment failure information and operation risk information at that moment.

[0068] Other advantages of the present invention will be analyzed in more detail in subsequent embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0070] Figure 1 The block diagram of a medical device monitoring system based on the Internet of Things in an embodiment of the present invention. DETAILED DESCRIPTION

[0071] 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 a part of the embodiments of the present invention, rather than all the embodiments. The components of the embodiments of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the claimed invention, but merely represents the selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0072] like Figure 1 As shown, in one embodiment of the present invention, the present invention provides a medical device monitoring system based on the Internet of Things, comprising:

[0073] A data acquisition module, used to obtain real-time operation data and positioning data of medical equipment;

[0074] The operation data of the medical equipment includes operation parameter data, usage record data, fault record data, maintenance record data, working environment data and consumables usage data.

[0075] In this embodiment, the operating parameter data includes voltage data, current data, power data, pressure data, revolution data, etc. of the medical device; the usage record data includes usage time data, usage frequency data, usage times data, etc. of the medical device; the fault record data includes fault occurrence time data, fault type code data, fault duration data, fault times data, fault frequency data, etc. of the medical device; the maintenance record data includes maintenance times data, maintenance duration data, maintenance content code data, maintenance result code data, etc. of the medical device; the working environment data includes temperature data, humidity data, air quality data, electromagnetic interference data, etc. under the working environment of the medical device; the consumable usage data includes consumable usage data, consumable consumption time trend data, consumable warranty time ratio data, consumable replacement frequency data, etc. of the medical device. The consumable warranty time ratio data is the ratio between the time difference between the current date and the consumable production date and the consumable warranty time.

[0076] The operating parameter data, usage record data, fault record data, maintenance record data, working environment data and consumables usage data all have a time tag on the data of each dimension.

[0077] Through sensor technology, voltage sensors, current sensors, power meters, pressure sensors, speed sensors and other monitoring are used to obtain the operating parameter data in the operating data of medical equipment; through the hospital information system (HIS) and the data recording system of the medical equipment itself, the usage record data, maintenance record data and consumables usage data in the operating data of the medical equipment are obtained; through temperature sensors, humidity sensors, air quality sensors and electromagnetic sensors and other monitoring, the working environment data is obtained.

[0078] The calculation expression of the operation data of the medical equipment is as follows:

[0079] {x k (n)}={x j (i),1≤j≤k,1≤i≤n},

[0080] Among them, x k (n) represents the medical equipment operation data, x j (i) represents the i-th dimension data corresponding to the j-th type of medical equipment operation data, k represents the total number of types of medical equipment operation data, n represents the total number of dimensions of medical equipment operation data, i = 1, 2,…, t,…, n.

[0081] A data processing module, used for preprocessing the historical operation data of several medical devices to obtain the preprocessed operation data;

[0082] The data processing module comprises:

[0083] A first preprocessing submodule, used to filter out random noise from the operation data of the medical device to obtain first preprocessed operation data;

[0084] The calculation expression of the first preprocessing operation data is as follows:

[0085]

[0086] Among them, x 1 (t) represents the first preprocessed operating data, and x(i) represents the first dimension data in the operating data of the medical device;

[0087] A second preprocessing submodule is used for standardizing the first preprocessing operation data to obtain second preprocessing operation data;

[0088] The calculation expression of the second preprocessing operation data is as follows:

[0089]

[0090] Among them, x 2 (t) represents the second pre-processing operation data, x minrepresents the maximum data value in the second preprocessing run data, x max represents the minimum data value in the second preprocessing operation data;

[0091] A third preprocessing submodule is used to perform equal-width discretization processing on the second preprocessing running data based on a discrete model to obtain third preprocessing running data, and to set a category label for each category interval in the third preprocessing running data;

[0092] The calculation expression of the discrete model is as follows:

[0093]

[0094] Wherein, τ represents the category label of the τth category interval in the third preprocessing operation data, and ω represents the category interval width;

[0095] A fourth preprocessing module, used for performing missing value supplementation processing on the third preprocessing operation data based on a missing value model to obtain fourth preprocessing operation data;

[0096] The calculation expression of the missing value model is as follows:

[0097]

[0098] in, indicates missing data in the i-th dimension, Represents the previous adjacent dimension data of the missing data of the i-th dimension, Represents the data of the adjacent dimension after the missing data of the i-th dimension;

[0099] a fifth preprocessing module, configured to perform outlier processing on the fourth preprocessed operating data based on an outlier processing model to obtain preprocessed operating data;

[0100] The calculation expression of the outlier processing model is as follows:

[0101]

[0102] Among them, z i (x 4 ) represents the function of performing Z-score abnormal monitoring on the fourth preprocessing operation data, x 4 (i) represents the fourth preprocessing operation data, μ represents the mean of the fourth preprocessing operation data, and δ represents the standard deviation of the fourth preprocessing operation data.

[0103] A feature extraction module is used to extract features from the preprocessed operation data to obtain a feature set for predicting equipment operation risks;

[0104] The feature extraction module comprises:

[0105] The first feature extraction submodule is used to extract the operation risk warning features of several medical devices based on the long short-term memory network according to the preprocessed operation data;

[0106] The second feature extraction submodule is used to evaluate the importance of the operation risk warning features of each medical device through the random forest method to obtain the feature average importance value of the operation risk warning features;

[0107] The third feature extraction submodule is used to select, according to a resampling method, a number of operation risk warning features whose average feature importance values ​​are greater than a preset importance threshold as important operation risk warning features, and obtain a first operation risk feature set consisting of the important operation risk warning features;

[0108] The fourth feature extraction submodule is used to repeatedly eliminate several important operation risk warning features with the lowest importance in the first operation risk feature set several times according to the random Mori recursive feature elimination method, and obtain a corresponding second operation risk feature set after each elimination;

[0109] A fifth feature extraction submodule is used to use the random forest method to perform a medical equipment operation risk early warning detection test using each second operation risk feature set, and use the explanation coefficient corresponding to each second operation risk feature set as its average accuracy;

[0110] The sixth feature extraction submodule is used to select the second operation risk feature set corresponding to the highest average accuracy as the equipment operation risk prediction feature set.

[0111] A model training module is used to train a support vector machine based on a device operation risk prediction feature set to obtain a trained support vector machine;

[0112] The model training module includes:

[0113] The first model training submodule is used to input the equipment operation risk prediction feature set into the support vector machine, train the support vector machine, and obtain the optimal normal vector and intercept of the hyperplane;

[0114] The first model training submodule includes:

[0115] The first model training unit is used to construct the support vector machine for the medical equipment operation risk warning problem:

[0116]

[0117]

[0118] in, represents the minimization function based on the hyperplane normal vector, hyperplane intercept and slack variables, W k′ represents the hyperplane normal vector corresponding to the k′th risk warning category, ||·|| 2 represents the square of the norm, C represents the first regularization parameter, ξ i′,k′ represents the slack variable related to the data of the corresponding dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set and the k′th risk warning category, st represents the slack variable that makes y i′,k′ Indicates the indicator variable associated with the feature and risk warning category, represents the transpose of the hyperplane normal vector corresponding to the k′th risk warning category, X i′ represents the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, b k′ represents the hyperplane intercept corresponding to the k′th risk warning category, represents any selection, where y i′ = k′, the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set belongs to the k′th risk warning category, y i′ ≠k′, the data of the dimension corresponding to the i′th important operation risk warning feature in the equipment operation risk prediction feature set does not belong to the k′th risk warning category;

[0119] The second model training unit is used to construct a decision boundary maximization interval based on the medical equipment operation risk warning problem, so as to identify the risk warning category to which the data of the corresponding dimension of each important operation risk warning feature belongs;

[0120] The third model training unit is used to construct the loss function of the support vector machine:

[0121]

[0122] Among them, L csvm represents the loss function of the support vector machine, y i′ represents the true category label of the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, max(·) represents the maximum value function, and s j represents the score of the j-th risk warning category, represents the score of the true category, Δ represents the positive interval parameter, λ represents the second regularization parameter, and R(W) represents the regularization term of the hyperplane normal vector; the positive interval parameter is used to control the minimum gap between the correct category score and the wrong category score;

[0123] The fourth model training unit is used to input important operation risk warning features in the equipment operation risk prediction feature set into the support vector machine with the goal of minimizing the loss function of the support vector machine, train the support vector machine, and obtain the optimal hyperplane normal vector and hyperplane intercept.

[0124] The second model training submodule is used to obtain a trained support vector machine based on an optimal hyperplane normal vector and a hyperplane intercept.

[0125] The monitoring and alarm module is used to issue medical equipment operation risk warnings for the real-time collected operation data of the medical equipment according to a preset risk level threshold using a trained support vector machine to obtain real-time operation risk warning information of the medical equipment; the time label corresponding to the real-time operation risk warning information of the medical equipment is consistent with the time label of the corresponding operation data.

[0126] According to the preset risk level threshold, the risks caused by different types and dimensions of data in the operation data of medical equipment can be accurately divided into different hyperplanes, thereby achieving comprehensive and different degrees of risk warning for medical equipment.

[0127] A data storage module, used to store the operation data, operation risk warning information and positioning data of the medical equipment in the cloud according to a preset time interval;

[0128] The IoT management module is used to build an IoT data management platform to visualize the real-time operation data, operation risk warning information and positioning data of different medical devices. The IoT data management platform matches the real-time operation data, operation risk warning information and positioning data of each medical device based on time tags and visualizes them.

[0129] The above description is only a specific implementation mode of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. A medical equipment monitoring system based on the Internet of Things, characterized in that: include: A data acquisition module, used to obtain real-time operation data and positioning data of medical equipment; A data processing module, used for preprocessing the historical operation data of several medical devices to obtain the preprocessed operation data; A feature extraction module is used to extract features from the preprocessed operation data to obtain a feature set for predicting equipment operation risks; A model training module is used to train a support vector machine based on a device operation risk prediction feature set to obtain a trained support vector machine; The monitoring and alarm module is used to use the trained support vector machine to conduct medical equipment operation risk warning for the real-time collected operation data of the medical equipment according to the preset risk level threshold, and obtain the real-time operation risk warning information of the medical equipment; A data storage module, used to store the operation data, operation risk warning information and positioning data of the medical equipment in the cloud according to a preset time interval; The Internet of Things management module is used to build an Internet of Things data management platform to visualize the real-time operating data, operating risk warning information and positioning data of different medical devices.

2. The medical device monitoring system based on the Internet of Things according to claim 1, characterized in that: The operation data of the medical equipment includes operation parameter data, usage record data, fault record data, maintenance record data, working environment data and consumables usage data.

3. The medical device monitoring system based on the Internet of Things according to claim 2 is characterized in that: The operating parameter data, usage record data, fault record data, maintenance record data, working environment data, consumable usage data and positioning data all have time tags on the data of each dimension.

4. The medical device monitoring system based on the Internet of Things according to claim 3 is characterized in that: The calculation expression of the operation data of the medical equipment is as follows: {x k (n)}={x j (i),1≤j≤k,1≤i≤n}, Among them, x k (n) represents the medical equipment operation data, x j (i) represents the i-th dimension data corresponding to the j-th type of medical equipment operation data, k represents the total number of types of medical equipment operation data, n represents the total number of dimensions of medical equipment operation data, i = 1, 2, ..., t, ..., n.

5. The medical device monitoring system based on the Internet of Things according to claim 4 is characterized in that: The data processing module comprises: A first preprocessing submodule, used to filter out random noise from the operation data of the medical device to obtain first preprocessed operation data; The calculation expression of the first preprocessing operation data is as follows: Among them, x 1 (t) represents the first preprocessed operating data, and x(i) represents the i-th dimension data in the operating data of the medical device; A second preprocessing submodule is used for standardizing the first preprocessing operation data to obtain second preprocessing operation data; The calculation expression of the second preprocessing operation data is as follows: Among them, x 2 (t) represents the second pre-processing operation data, x min represents the maximum data value in the second preprocessing run data, x max represents the minimum data value in the second preprocessing operation data; A third preprocessing submodule is used to perform equal-width discretization processing on the second preprocessing running data based on a discrete model to obtain third preprocessing running data, and to set a category label for each category interval in the third preprocessing running data; The calculation expression of the discrete model is as follows: Wherein, τ represents the category label of the τth category interval in the third preprocessing operation data, and ω represents the category interval width; A fourth preprocessing module, used for performing missing value supplementation processing on the third preprocessing operation data based on a missing value model to obtain fourth preprocessing operation data; The calculation expression of the missing value model is as follows: in, indicates missing data in the i-th dimension, Represents the previous adjacent dimension data of the missing data of the i-th dimension, Represents the data of the adjacent dimension after the missing data of the i-th dimension; a fifth preprocessing module, configured to perform outlier processing on the fourth preprocessed operating data based on an outlier processing model to obtain preprocessed operating data; The calculation expression of the outlier processing model is as follows: Among them, z i (x 4 ) represents the function of performing Z-score abnormal monitoring on the fourth preprocessing operation data, x 4 (i) represents the fourth preprocessing operation data, μ represents the mean of the fourth preprocessing operation data, and δ represents the standard deviation of the fourth preprocessing operation data.

6. The medical device monitoring system based on the Internet of Things according to claim 5, characterized in that: The feature extraction module comprises: The first feature extraction submodule is used to extract the operation risk warning features of several medical devices based on the long short-term memory network according to the preprocessed operation data; The second feature extraction submodule is used to evaluate the importance of the operation risk warning features of each medical device through the random forest method to obtain the feature average importance value of the operation risk warning features; The third feature extraction submodule is used to select, according to a resampling method, a number of operation risk warning features whose average feature importance values ​​are greater than a preset importance threshold as important operation risk warning features, and obtain a first operation risk feature set consisting of the important operation risk warning features; The fourth feature extraction submodule is used to repeatedly eliminate several important operation risk warning features with the lowest importance in the first operation risk feature set according to the random Mori recursive feature elimination method for several times, and obtain a corresponding second operation risk feature set after each elimination; a fifth feature extraction submodule, for performing a medical equipment operation risk early warning detection test using each second operation risk feature set by using a random forest method, and taking the explanation coefficient corresponding to each second operation risk feature set as its average accuracy; The sixth feature extraction submodule is used to select the second operation risk feature set corresponding to the highest average accuracy as the equipment operation risk prediction feature set.

7. The medical device monitoring system based on the Internet of Things according to claim 6, characterized in that: The model training module includes: The first model training submodule is used to input the equipment operation risk prediction feature set into the support vector machine, train the support vector machine, and obtain the optimal normal vector and intercept of the hyperplane; The second model training submodule is used to obtain a trained support vector machine based on an optimal hyperplane normal vector and a hyperplane intercept.

8. The medical device monitoring system based on the Internet of Things according to claim 7, characterized in that: The first model training submodule includes: The first model training unit is used to construct the support vector machine for the medical equipment operation risk warning problem: in, represents the minimization function based on the hyperplane normal vector, hyperplane intercept and slack variables, W k′ represents the hyperplane normal vector corresponding to the k′th risk warning category, ||·|| 2 represents the square of the norm, C represents the first regularization parameter, ξ i′,k′ represents the slack variable related to the data of the corresponding dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set and the k′th risk warning category, st represents the slack variable that makes y i′,k′ Indicates the indicator variable associated with the feature and risk warning category, represents the transpose of the hyperplane normal vector corresponding to the k′th risk warning category, X i′ represents the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, b k′ represents the hyperplane intercept corresponding to the k′th risk warning category, represents any selection, where y i′ = k′, the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set belongs to the k′th risk warning category, y i′ ≠k′, the data of the dimension corresponding to the i′th important operation risk warning feature in the equipment operation risk prediction feature set does not belong to the k′th risk warning category; The second model training unit is used to construct a decision boundary maximization interval based on the medical equipment operation risk warning problem, so as to identify the risk warning category to which the data of the corresponding dimension of each important operation risk warning feature belongs; The third model training unit is used to construct the loss function of the support vector machine: Among them, L csvm represents the loss function of the support vector machine, y i′ represents the true category label of the data corresponding to the dimension of the i′th important operation risk warning feature in the equipment operation risk prediction feature set, max(·) represents the maximum value function, and s j represents the score of the j-th risk warning category, represents the score of the true category, Δ represents the positive interval parameter, λ represents the second regularization parameter, and R(W) represents the regularization term of the hyperplane normal vector; The fourth model training unit is used to input important operation risk warning features in the equipment operation risk prediction feature set into the support vector machine with the goal of minimizing the loss function of the support vector machine, train the support vector machine, and obtain the optimal hyperplane normal vector and hyperplane intercept.

9. The medical device monitoring system based on the Internet of Things according to claim 8, characterized in that: The IoT data management platform matches the real-time operating data, operating risk warning information and positioning data of each medical device based on time tags, and displays them visually.

Images