Low-body-temperature infection monitoring method and system based on artificial intelligence and storage medium

Abstract

The invention discloses a low-body-temperature infection monitoring method and system based on artificial intelligence and a storage medium. The method comprises the following steps: extracting the category and occurrence time of an infection event from comprehensive medical information based on a body temperature interval, extracting first sample data from all the comprehensive medical information, and generating second sample data corresponding to the first sample data based on a time resolution; preprocessing the second sample data based on the data type, extracting data features, and training and generating a monitoring model based on the data features; collecting monitoring data of a patient to be monitored in real time based on the data label, preprocessing the monitoring data, inputting the preprocessed monitoring data into the monitoring model, and outputting a prediction event and prediction occurrence time of the prediction event; and summarizing all prediction events and prediction occurrence time, generating a monitoring report, and outputting feedback measures based on the monitoring report. According to the invention, the timeliness and accuracy of low-body-temperature infection monitoring are improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent monitoring of medical data, and in particular to an artificial intelligence-based hypothermia infection monitoring method, system and storage medium. Background Art

[0002] Hypothermia refers to a state in which the body's core temperature falls below 36°C. It is often experienced during surgery due to factors such as anesthesia, low operating room temperatures, prolonged surgery, and fluid loss. Hypothermia can adversely affect multiple physiological systems, including coagulation, immunity, and drug metabolism. Studies have shown a significant correlation between hypothermia and surgical wound infection. Specifically, hypothermia suppresses immune function and reduces white blood cell activity, thereby increasing the risk of surgical wound infection. Traditional thermometers typically only measure intermittently and are unable to monitor temperature changes in real time and continuously. They also lack the ability to comprehensively analyze temperature data, making it difficult to correlate temperature data with other physiological parameters (such as heart rate, blood pressure, and blood oxygen saturation). Consequently, it is difficult to fully assess the impact of hypothermia on surgical wound infection. Therefore, it is necessary to combine the advantages of artificial intelligence technology in data analysis and pattern recognition to overcome the limitations of traditional monitoring methods.

[0003] Similar prior art includes a Chinese patent application with publication number CN115331816A, which discloses an information decision support system for perioperative temperature management of children with congenital heart disease. The system includes an information acquisition module, an automatic risk identification module, a hypothermia protection module, and a nursing order generation module. The information acquisition module acquires patient data and transmits the relevant data to the automatic risk identification module. A calculation model is set up in the automatic risk identification module, and the relevant data is substituted into the calculation model for calculation. If the calculated value is ≥0.3387, the hypothermia protection module is automatically activated. The hypothermia protection module automatically pushes relevant nursing measures to medical staff to formulate a personalized nursing plan through the automatic push system for temperature management knowledge for high-risk groups. If the calculated value is <0.3387, the personalized nursing plan is terminated, the end time of the nursing measures is calculated, and the medical staff is reminded to evaluate the nursing outcomes, while recording the evaluation time. Finally, the nursing order generation module generates nursing documents. The beneficial effects of this invention are: good predictive efficiency and decision support functions. There is also a Chinese patent application with publication number CN117476213A, which discloses a hypothermia warning method and system based on non-invasive parameters, including obtaining non-invasive biological indicators of hypothermia patients and non-invasive biological indicators of the medical environment; preprocessing the non-invasive biological indicators of hypothermia patients to obtain a data set; inputting the data set into a real-time warning model to obtain a first probability value; inputting the non-invasive biological indicators of the medical environment into a data maintenance processing module to obtain a second probability value; calculating the first probability value and the second probability value to obtain a probability of rewarming treatment requirement; judging whether the hypothermia patient needs rewarming treatment based on the probability of rewarming treatment requirement to obtain a judgment result; and determining the warning result of the hypothermia patient based on the judgment result.

[0004] The shortcomings of existing technologies are mainly reflected in the fact that decision-making management is generated only after predicting the risk values ​​corresponding to hypothermia, and only early warning analysis is performed on non-invasive biological indicators of hypothermia patients. No prediction analysis is performed on the risk types of hypothermia. In actual situations, it is necessary to accurately predict and analyze hypothermia infection events and generate feedback measures in a timely manner. Summary of the Invention

[0005] The present application provides an artificial intelligence-based hypothermia infection monitoring method, system and storage medium for improving the timeliness and accuracy of hypothermia infection monitoring.

[0006] In a first aspect, the present application provides an artificial intelligence-based hypothermia infection monitoring method, the artificial intelligence-based hypothermia infection monitoring method comprising: Collecting comprehensive medical information of surgical patients, setting temperature intervals, extracting the type and occurrence time of infection events from the comprehensive medical information based on the temperature intervals, extracting first sample data from all the comprehensive medical information based on the type and occurrence time, setting a time resolution, and generating second sample data corresponding to the first sample data based on the time resolution; Obtaining a data type and a data label of the second sample data, preprocessing the second sample data based on the data type to extract data features, and training and generating a monitoring model based on the data features; Collecting monitoring data of the patient to be monitored in real time based on the data tags, pre-processing the monitoring data and inputting it into the monitoring model, and outputting a predicted event and a predicted occurrence time of the predicted event; Summarize all the predicted events and the predicted occurrence times, generate a monitoring report, and output feedback measures based on the monitoring report.

[0007] In combination with the first aspect, generating second sample data corresponding to the first sample data based on the time resolution includes: Acquire basic information of the surgical patient, and based on the basic information, acquire subsample data of the same category and having a time resolution higher than a day from all the first sample data; Extracting medical data from the subsample data, obtaining specific changes in any of the medical data, marking time information corresponding to the specific changes, and generating estimated events and estimated occurrence times corresponding to the estimated events based on the time information, wherein the specific changes include changes in data values, changes in collection frequency, and changes in data items; The estimated event and the estimated occurrence time are added to the first sample data in correspondence to generate the second sample data.

[0008] In combination with the first aspect, generating an estimated event and an estimated occurrence time corresponding to the estimated event based on the time information includes: Establishing a medical event database, obtaining supplementary events associated with the infection event in the medical event database, extracting associated events from the supplementary events based on the specific change, and setting the associated events as the estimated events; A change timestamp is output based on a change type of the specific change, a lag time is output based on the change type, and the change timestamp and the lag time are combined to set as the estimated occurrence time.

[0009] In combination with the first aspect, the step of generating a monitoring model based on the data feature training includes: Marking category labels for the data features based on the data labels, setting the category labels corresponding to the infection events as target variables, encoding all the data features based on label encoding, and converting them into numerical sample sets; Construct a neural network model, input the numerical sample set into the neural network model after standardization, set a training set and a validation set based on the target variable to train the neural network model, adjust the hyperparameters of the neural network model based on a grid search method, and set the adjusted neural network model as the monitoring model.

[0010] In combination with the first aspect, setting a training set and a validation set based on the target variable to train the neural network model includes: The neural network model includes an input layer, an intermediate layer, and an output layer, wherein the first neurons included in the input layer and part of the intermediate layer are set as a first training unit, and the second neurons included in the output layer and the remaining intermediate layer are set as a second training unit; Dividing the numerical sample set into the training set, the validation set, and the test set, obtaining training data of the category label and the label code pair in the training set, inputting the training data into the first training unit to generate a first analysis result, and the second training unit receiving the first analysis result and generating a second analysis result; The second analysis result is compared with the category label based on the target variable, and a model error is calculated based on a loss function. This step is repeated until the model error is less than a first preset value.

[0011] In combination with the first aspect, the real-time collection of monitoring data of the patient to be monitored based on the data tag includes: Based on the data labels, a sensor device for collecting vital sign data of the patient to be monitored is deployed. Based on the operating status of the sensor device, it is determined whether all the vital sign data are valid. If they are all valid, the vital sign data are set as the monitoring data. Otherwise, all the vital sign data are divided into a valid data set and an invalid data set. The encoder and decoder of the automatic encoder are trained using the valid data set, and the encoder is trained using the invalid data set. The encoder maps the invalid data set to the latent space of the valid data set. The trained encoder encodes and decodes the invalid data in the invalid data set, restores the invalid data and sets it as the monitoring data.

[0012] In conjunction with the first aspect, generating a monitoring report includes: The monitoring model obtains a feature value of the monitoring data, performs a prediction analysis on the feature value based on the data label, and outputs the predicted event and the predicted occurrence time of the predicted event; All the predicted events are arranged in sequence based on the predicted occurrence time, and the monitoring report is generated.

[0013] In combination with the first aspect, a medical knowledge base is constructed, and the medical method corresponding to the monitoring report is retrieved from the medical knowledge base based on a natural language algorithm, and the feedback measure is generated based on the medical method.

[0014] In a second aspect, the present application provides an artificial intelligence-based hypothermia infection monitoring system, the artificial intelligence-based hypothermia infection monitoring system comprising: a collection module for collecting comprehensive medical information of surgical patients, setting a temperature range, extracting a type and occurrence time of an infection event from the comprehensive medical information based on the temperature range, extracting first sample data from all the comprehensive medical information based on the type and occurrence time, setting a time resolution, and generating second sample data corresponding to the first sample data based on the time resolution; a training module, configured to obtain a data type and a data label of the second sample data, pre-process the second sample data based on the data type, extract data features, and train and generate a monitoring model based on the data features; A prediction module, configured to collect monitoring data of the patient to be monitored in real time according to the data tags, pre-process the monitoring data and input it into the monitoring model, and output a predicted event and a predicted occurrence time of the predicted event; The feedback module is used to summarize all the predicted events and the predicted occurrence time, generate a monitoring report, and output feedback measures based on the monitoring report.

[0015] A third aspect of the present application provides a computer-readable storage medium, which stores instructions that, when executed on a computer, enable the computer to execute the above-mentioned artificial intelligence-based hypothermia infection monitoring method.

[0016] In the technical solution provided by the present application, first, by collecting comprehensive medical information to screen out the first sample data, and supplementing and expanding the first sample data based on the time resolution to generate the second sample data, the integrity and accuracy of the sample data can be improved. Then, a neural network model is trained based on the second sample data, and a second analysis result is generated based on the first analysis result until the model error is less than the first preset value, and a monitoring model is generated, which can improve the generalization ability and prediction accuracy of the monitoring model, and at the same time ensure the security of the second sample data. Finally, the monitoring data of the monitored patient is predicted and analyzed by the monitoring model, the predicted events are output, and a monitoring report and corresponding feedback measures are generated, which can improve the accuracy of hypothermia infection of the monitored patient and improve the efficiency and reliability of hypothermia infection monitoring.

[0017] This application also uses an automatic encoder to perform validity analysis on vital sign data and generate monitoring data, which can improve the integrity and accuracy of the monitoring data. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] 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 description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0019] Figure 1 This is a schematic diagram of an embodiment of a hypothermia infection monitoring method based on artificial intelligence in an embodiment of the present application; Figure 2 This is a schematic diagram of an embodiment of the neural network model training structure in the embodiment of the present application; Figure 3 This is a schematic diagram of an embodiment of a detailed process of an automatic encoder processing invalid data in an embodiment of the present application; Figure 4 This is a schematic diagram of an embodiment of an artificial intelligence-based hypothermia infection monitoring system in an embodiment of the present application. DETAILED DESCRIPTION

[0020] The embodiments of the present application provide a method, system and storage medium for monitoring hypothermia infection based on artificial intelligence. The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0021] For ease of understanding, the specific process of the embodiment of the present application is described below. Figure 1 In the embodiments of the present application, an embodiment of the artificial intelligence-based hypothermia infection monitoring method includes: Step S101: Collect comprehensive medical information of surgical patients, set body temperature ranges, extract the type and occurrence time of infection events from the comprehensive medical information based on the body temperature ranges, extract first sample data from all comprehensive medical information based on the type and occurrence time, set time resolution, and generate second sample data corresponding to the first sample data based on the time resolution.

[0022] It is understandable that the execution subject of this application can be an artificial intelligence-based hypothermia infection monitoring device, or a terminal or server, which is not limited here. The embodiment of this application is described by taking the server as the execution subject as an example.

[0023] Specifically, comprehensive medical information refers to surgical patient data obtained from the hospital's electronic medical record (EMR), surgical record system, anesthesia record system, and nursing record system. This includes, but is not limited to, basic patient information (age, gender, weight, height, medical history, allergies, etc.), surgical-related information (type of surgery, duration of surgery, anesthesia method, intraoperative medications, etc.), intraoperative monitoring data (temperature, heart rate, blood pressure, blood oxygen saturation, respiratory rate, etc.), and postoperative nursing records (temperature changes, wound condition, drainage status, laboratory test results, etc.). Based on clinical experience and literature, a temperature range for hypothermia is determined. For example, a temperature below 36.0°C is defined as hypothermia, with mild hypothermia defined as between 34°C and 36°C, moderate hypothermia as between 32°C and 34°C, and severe hypothermia as below 32°C. Intraoperative and postoperative temperature data are extracted from comprehensive medical information and categorized according to the defined temperature ranges. An infection event refers to an event with infection-related symptoms occurring between the start and end of surgery. Clinical diagnostic criteria are used to determine the type and category of infection events, such as surgical site infection (SSI), pneumonia, and urinary tract infection. The onset time of infection events is extracted from comprehensive medical information and matched with temperature data. Both the infection event and onset time can be extracted using a semantic recognition algorithm. The first sample data refers to data information containing hypothermia and infection events, extracted based on temperature ranges and infection events. Therefore, the first sample data consists of multiple data. Temporal resolution refers to the time scale, such as hourly, every 30 minutes, or every 15 minutes. Because the first sample data may contain data with a higher temporal resolution, the accuracy of the infection event's onset time can be affected. For example, if temperature data is recorded daily, there may be debate about whether the infection event is related to the hypothermia data. Therefore, the temporal resolution needs to be reduced and data corresponding to the time scale in the first sample data needs to be added to generate the second sample data. This second sample data supplements and expands the first sample data.

[0024] Step S102: Obtain the data type and data label of the second sample data, pre-process the second sample data based on the data type, extract data features, and generate a monitoring model based on data feature training.

[0025] Specifically, the data type of the second sample data is identified, such as numeric, categorical, or time. A data label refers to the type of data information in the second sample data, such as body temperature, heart rate, or blood pressure. Preprocessing includes handling missing values, outliers, and duplicate values. For example, missing values ​​can be filled with the mean, outliers can be removed, and numeric data can be normalized using methods such as Z-score or Min-Max normalization. Feature extraction is performed based on the data type to generate data features, i.e., data characteristics, such as time series characteristics (mean, variance, trend, etc.), statistical characteristics (maximum, minimum, median, etc.), and physiological indicator characteristics (heart rate, blood pressure, blood oxygen saturation, etc.). An appropriate machine learning model is selected, such as logistic regression, decision tree, random forest, support vector machine (SVM), or neural network. The preprocessed second sample data is input into the model for training to generate a monitoring model.

[0026] Step S103: Based on the data tags, the monitoring data of the patient to be monitored is collected in real time, the monitoring data is pre-processed and input into the monitoring model, and the predicted event and the predicted occurrence time of the predicted event are output.

[0027] Specifically, based on data labels, real-time monitoring data, such as temperature, heart rate, and blood pressure, is collected from the patient being monitored. This real-time data is preprocessed, such as through standardization and normalization. This preprocessed data is then fed into a trained monitoring model, which then outputs a predicted event (e.g., whether an infection has occurred) and the predicted time of occurrence.

[0028] Step S104: Summarize all predicted events and predicted occurrence times, generate a monitoring report, and output feedback measures based on the monitoring report.

[0029] Specifically, all predicted events and predicted occurrence times are summarized to generate a monitoring report. This report includes patient information, monitoring data, prediction results, and predicted occurrence times. Based on the prediction results and clinical guidelines, appropriate feedback measures are formulated, such as adjusting treatment plans, strengthening monitoring, and using preventive medications. These feedback measures are provided to medical staff to guide clinical decision-making.

[0030] In a specific embodiment, generating second sample data corresponding to the first sample data based on the time resolution includes: (1) Obtain basic information of surgical patients, and based on the basic information, obtain sub-sample data with the same category and a time resolution higher than the daily unit from all first sample data.

[0031] (2) Extract medical data from the subsample data, obtain specific changes in any medical data, mark the time information corresponding to the specific changes, and generate estimated events and estimated occurrence times corresponding to the estimated events based on the time information, where specific changes include changes in data values, changes in collection frequency, and changes in data items.

[0032] (3) Add the estimated event and the estimated occurrence time to the first sample data to generate the second sample data.

[0033] Specifically, basic information about surgical patients, particularly their age, is extracted from the first sample data. Based on clinical experience and research, patient age is categorized into different age ranges, such as children (0-12 years), adolescents (13-18 years), young adults (19-40 years), middle-aged adults (41-65 years), and the elderly (66 years and older). Subsample data is then filtered from all first sample data to include data of the same category as the current patient, such as surgical site infection (SSI), pneumonia, and urinary tract infection. Data with a temporal resolution higher than daily, such as hourly, 30-minute, or 15-minute data, is selected.

[0034] The filtered subsample data is stored in a structured format, ensuring that each data point includes a timestamp and corresponding medical data. Medical data includes surgical information, intraoperative monitoring data, and postoperative care records. Specific changes include changes in data values ​​(such as a sudden increase in body temperature, increased heart rate, or decreased blood pressure), changes in collection frequency (such as changes in vital sign monitoring frequency or medication administration frequency), and changes in data items (such as new laboratory tests or changes in imaging results). Statistical methods (such as Z-scores) or machine learning methods (such as anomaly detection models) can be used to identify data value changes. Comparing collection frequencies across different time periods can identify changes in collection frequency, and comparing data items at different time points can identify changes in new or changed data items. Identified specific changes are marked, and the time at which the change occurred is recorded. The time information of the specific change is used as the estimated time of occurrence of the event. For example, if the first sample data at timestamp T1 shows a body temperature of 36, heart rate of 85, blood pressure of 118 / 82, antibiotics administered, and a white blood cell count of 10,000, and a temperature drop is detected through feature change identification, the second sample data can be supplemented with the following: At timestamp T1, body temperature of 36, heart rate of 85, blood pressure of 118 / 82, antibiotics administered, and a white blood cell count of 10,000, indicating a temperature drop. Through these steps, the temporal resolution-based generation of the second sample data effectively captures and labels specific changes in medical data. These changes not only enrich the sample data but also provide more comprehensive information for subsequent monitoring model training, thereby improving the accuracy and reliability of hypothermia infection monitoring.

[0035] In a specific embodiment, generating an estimated event and an estimated occurrence time corresponding to the estimated event based on the time information includes: (1) Establish a medical event database, obtain supplementary events associated with infection events in the medical event database, extract associated events from the supplementary events based on specific changes, and set the associated events as estimated events.

[0036] (2) Output the change timestamp based on the change type of the specific change, output the lag time based on the change type, and combine the change timestamp and lag time to set the estimated occurrence time.

[0037] Specifically, the medical event database is a database that lists possible medical symptoms based on medical knowledge, including correlations between these symptoms. Supplementary events are other medical events in the medical event database that are strongly associated with infection events, identified based on correlations. Therefore, there are multiple supplementary events. Among all supplementary events, medical events that only occur due to specific changes are considered associated events, and therefore, these associated events are set as estimated events.

[0038] The change types of specific changes are the data value changes, acquisition frequency changes, and data item changes mentioned above. If it is a data value change, the first change curve of the data value change can be predicted, and the change timestamp is generated based on the predicted time of the first change curve. If it is an acquisition frequency change, the change timestamp is generated based on the time interval of the acquisition frequency change. If it is a data item change, the correlation between all data items can be obtained, and other data items corresponding to the maximum correlation with the data item change can be selected. The data item is predicted to generate a second change curve, and the change timestamp is generated based on the predicted time of the second change curve. Lag time refers to the time period of a medical event that is delayed after a specific change occurs. For example, when the body temperature drops, the time period required for the infection event "coagulation dysfunction" to occur is T2, and the lag time is T2. Therefore, it is necessary to combine the change timestamp with the lag time and set it as the estimated occurrence time.

[0039] In a specific embodiment, generating a monitoring model based on data feature training includes: (1) Based on the data label, the data features are labeled with category labels, the category labels corresponding to the infection events are set as the target variables, and all data features are encoded based on the label encoding and converted into a numerical sample set.

[0040] (2) Construct a neural network model, input the standardized numerical sample set into the neural network model, set the training set and validation set based on the target variable to train the neural network model, adjust the hyperparameters of the neural network model based on the grid search method, and set the adjusted neural network model as the monitoring model.

[0041] Specifically, category labels are used to label data features. The category labels corresponding to data features with infection events are set as target variables, thus representing multi-class label information. Label encoding is used to encode data features, ensuring that all features are numerical, thus generating a numerical sample set.

[0042] The neural network model can be configured as a multilayer perceptron, effectively improving the accuracy of hypothermia infection monitoring. Hyperparameters refer to parameters such as the learning rate, batch size, number of hidden layers, and number of neurons used during neural network training. Using validation and test sets to determine the accuracy of the neural network model, a monitoring model can be generated.

[0043] In a specific embodiment, setting a training set and a validation set based on a target variable to train a neural network model includes: (1) The neural network model includes an input layer, an intermediate layer, and an output layer. The first neurons contained in the input layer and part of the intermediate layer are set as the first training unit, and the second neurons contained in the output layer and the remaining intermediate layer are set as the second training unit.

[0044] (2) Divide the numerical sample set into a training set, a validation set, and a test set, obtain training data paired with category labels and label codes in the training set, input the training data into the first training unit, generate a first analysis result, and the second training unit receives the first analysis result and generates a second analysis result.

[0045] (3) Compare the second analysis result with the category label based on the target variable, and calculate the model error based on the loss function. Repeat this step until the model error is less than the first preset value.

[0046] Specifically, if Figure 2 As shown, the input layer L1 and part of the intermediate layer L2 form the first training unit, and the output layer L4 and the remaining intermediate layer L3 form the second training unit. The neural network model includes the first and second training units. The first analysis result is output by part of the intermediate layer, and the second analysis result is output by the output layer. The second analysis result is the probability of the infection event category corresponding to the target variable.

[0047] By comparing the second analysis result with the category label, the difference between the predicted value and the true value can be calculated. Based on the calculated model error, the weights and bias of the model are adjusted using the backpropagation algorithm, and the training process of this step is repeated until the model error is less than the first preset value.

[0048] In a specific embodiment, real-time collection of monitoring data of a patient to be monitored based on data tags includes: Based on data labels, sensor equipment is deployed to collect vital sign data of patients to be monitored. Based on the operating status of the sensor equipment, it is determined whether all vital sign data are valid. If they are valid, the vital sign data are set as monitoring data. Otherwise, all vital sign data are divided into valid data sets and invalid data sets. The encoder and decoder of the automatic encoder are trained using the valid data set, and the encoder is trained using the invalid data set. The encoder maps the invalid data set to the latent space of the valid data set. The trained encoder encodes and decodes the invalid data in the invalid data set, restores the invalid data and sets it as monitoring data.

[0049] Specifically, the vital sign data corresponds to the data tag. The operating status refers to whether the sensor device is operating normally. If the operating status is abnormal, it means that there is invalid vital sign data. Otherwise, it means that all vital sign data are valid, and the vital sign data is set as monitoring data.

[0050] In order to improve the effectiveness and accuracy of monitoring data, invalid data needs to be recovered. The specific recovery process is as follows: Figure 3 As shown in the figure, an autoencoder is an artificial neural network used for unsupervised learning. It is primarily used for data compression, reconstruction, and dimensionality reduction. It consists of two main components: an encoder and a decoder. The process of collecting patient monitoring data in real time based on data labels and processing invalid data through the autoencoder effectively ensures the quality and integrity of the monitoring data. This approach combines sensor technology and data processing techniques to improve the accuracy and reliability of hypothermia infection monitoring.

[0051] In a specific embodiment, generating a monitoring report includes: (1) The monitoring model obtains the characteristic quantities of the monitoring data, performs predictive analysis on the characteristic quantities based on the data labels, and outputs the predicted events and the predicted occurrence time of the predicted events.

[0052] (2) Arrange all predicted events in sequence based on the predicted occurrence time and generate a monitoring report.

[0053] Specifically, the extracted features are input into the trained monitoring model, which outputs the predicted event and predicted time point based on the probability value, with the predicted event point set as the predicted occurrence time. The predicted event refers to the type of event that indicates the patient to be monitored will become infected.

[0054] All predicted events are arranged in order based on their predicted occurrence time. Typically, earlier events are listed first, while later events are listed last. Monitoring data and predicted events can be displayed using charts and graphs, and monitoring reports can be generated.

[0055] In a specific embodiment, a medical knowledge base is constructed, and the medical method corresponding to the monitoring report is retrieved from the medical knowledge base based on a natural language algorithm, and feedback measures are generated based on the medical method.

[0056] Specifically, a medical knowledge base contains information about infection events and medical diagnostic procedures. Natural language processing (NLP) models are computer algorithms and technologies used to understand and process human language. The goal of NLP models is to enable computers to understand, interpret, generate, and manipulate human language. Therefore, a natural language algorithm is used to filter medical diagnostic procedures (i.e., medical methods) from the medical knowledge base, and all of these methods are set as feedback measures.

[0057] The above describes the artificial intelligence-based hypothermia infection monitoring method in the embodiment of the present application. The following describes the artificial intelligence-based hypothermia infection monitoring system in the embodiment of the present application. Figure 4 In the embodiments of the present application, an embodiment of the artificial intelligence-based hypothermia infection monitoring system includes: The collection module 201 is used to collect comprehensive medical information of surgical patients, set body temperature ranges, extract the type and occurrence time of infection events from the comprehensive medical information based on the body temperature ranges, extract first sample data from all comprehensive medical information based on the type and occurrence time, set time resolution, and generate second sample data corresponding to the first sample data based on the time resolution.

[0058] The training module 202 is used to obtain the data type and data label of the second sample data, pre-process the second sample data based on the data type and extract data features, and generate a monitoring model based on the data feature training.

[0059] The prediction module 203 is used to collect monitoring data of the patient to be monitored in real time according to the data tags, pre-process the monitoring data and input it into the monitoring model, and output the predicted event and the predicted occurrence time of the predicted event.

[0060] The feedback module 204 is used to summarize all predicted events and predicted occurrence times, generate a monitoring report, and output feedback measures based on the monitoring report.

[0061] Through the collaborative cooperation of the above-mentioned components, first, by collecting comprehensive medical information to screen out the first sample data, and supplementing and expanding the first sample data based on the time resolution to generate the second sample data, the integrity and accuracy of the sample data can be improved. Then, the neural network model is trained based on the second sample data, and the second analysis result is generated based on the first analysis result until the model error is less than the first preset value, and a monitoring model is generated. This can improve the generalization ability and prediction accuracy of the monitoring model, while ensuring the security of the second sample data. Finally, the monitoring model is used to perform predictive analysis on the monitoring data of the monitored patient, output predicted events, and generate monitoring reports and corresponding feedback measures. This can improve the accuracy of hypothermia infection in the monitored patient and improve the efficiency and reliability of hypothermia infection monitoring.

[0062] This application also uses an automatic encoder to perform validity analysis on vital sign data and generate monitoring data, which can improve the integrity and accuracy of the monitoring data.

[0063] The present application also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. The computer-readable storage medium stores instructions, and when the instructions are executed on a computer, the computer executes the steps of the artificial intelligence-based hypothermia infection monitoring method.

[0064] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described systems, systems and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0065] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk, and other media that can store program code.

[0066] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A hypothermia infection monitoring method based on artificial intelligence, characterized in that: The artificial intelligence-based hypothermia infection monitoring method includes: Collecting comprehensive medical information of surgical patients, setting temperature intervals, extracting the type and occurrence time of infection events from the comprehensive medical information based on the temperature intervals, extracting first sample data from all the comprehensive medical information based on the type and occurrence time, setting a time resolution, and generating second sample data corresponding to the first sample data based on the time resolution; Obtaining a data type and a data label of the second sample data, preprocessing the second sample data based on the data type to extract data features, and training and generating a monitoring model based on the data features; Collecting monitoring data of the patient to be monitored in real time based on the data tags, pre-processing the monitoring data and inputting it into the monitoring model, and outputting a predicted event and a predicted occurrence time of the predicted event; Summarize all the predicted events and the predicted occurrence times, generate a monitoring report, and output feedback measures based on the monitoring report.

2. The artificial intelligence-based hypothermia infection monitoring method according to claim 1, characterized in that: The generating second sample data corresponding to the first sample data based on the time resolution includes: Acquire basic information of the surgical patient, and based on the basic information, acquire subsample data of the same category and having a time resolution higher than a day from all the first sample data; Extracting medical data from the subsample data, obtaining specific changes in any of the medical data, marking time information corresponding to the specific changes, and generating estimated events and estimated occurrence times corresponding to the estimated events based on the time information, wherein the specific changes include changes in data values, changes in collection frequency, and changes in data items; The estimated event and the estimated occurrence time are added to the first sample data in correspondence to generate the second sample data.

3. The artificial intelligence-based hypothermia infection monitoring method according to claim 2, characterized in that: The generating of the estimated event and the estimated occurrence time corresponding to the estimated event based on the time information includes: Establishing a medical event database, obtaining supplementary events associated with the infection event in the medical event database, extracting associated events from the supplementary events based on the specific change, and setting the associated events as the estimated events; A change timestamp is output based on a change type of the specific change, a lag time is output based on the change type, and the change timestamp and the lag time are combined to set as the estimated occurrence time.

4. The artificial intelligence-based hypothermia infection monitoring method according to claim 1, characterized in that: The generating of the monitoring model based on the data feature training includes: Marking category labels for the data features based on the data labels, setting the category labels corresponding to the infection events as target variables, encoding all the data features based on label encoding, and converting them into numerical sample sets; Construct a neural network model, input the numerical sample set into the neural network model after standardization, set a training set and a validation set based on the target variable to train the neural network model, adjust the hyperparameters of the neural network model based on a grid search method, and set the adjusted neural network model as the monitoring model.

5. The artificial intelligence-based hypothermia infection monitoring method according to claim 4, characterized in that: The step of setting a training set and a validation set based on the target variable to train the neural network model includes: The neural network model includes an input layer, an intermediate layer, and an output layer, wherein the first neurons included in the input layer and part of the intermediate layer are set as a first training unit, and the second neurons included in the output layer and the remaining intermediate layer are set as a second training unit; Dividing the numerical sample set into the training set, the validation set, and the test set, obtaining training data of the category label and the label code pair in the training set, inputting the training data into the first training unit to generate a first analysis result, and the second training unit receiving the first analysis result and generating a second analysis result; The second analysis result is compared with the category label based on the target variable, and a model error is calculated based on a loss function. This step is repeated until the model error is less than a first preset value.

6. The artificial intelligence-based hypothermia infection monitoring method according to claim 1, characterized in that: The real-time collection of monitoring data of the patient to be monitored based on the data tag includes: Based on the data labels, a sensor device for collecting vital sign data of the patient to be monitored is deployed. Based on the operating status of the sensor device, it is determined whether all the vital sign data are valid. If they are all valid, the vital sign data are set as the monitoring data. Otherwise, all the vital sign data are divided into a valid data set and an invalid data set. The encoder and decoder of the automatic encoder are trained using the valid data set, and the encoder is trained using the invalid data set. The encoder maps the invalid data set to the latent space of the valid data set. The trained encoder encodes and decodes the invalid data in the invalid data set, restores the invalid data and sets it as the monitoring data.

7. The artificial intelligence-based hypothermia infection monitoring method according to claim 1, characterized in that: The generating of the monitoring report includes: The monitoring model obtains a feature value of the monitoring data, performs a prediction analysis on the feature value based on the data label, and outputs the predicted event and the predicted occurrence time of the predicted event; All the predicted events are arranged in sequence based on the predicted occurrence time, and the monitoring report is generated.

8. The artificial intelligence-based hypothermia infection monitoring method according to claim 7, characterized in that: A medical knowledge base is constructed, and the medical method corresponding to the monitoring report is retrieved from the medical knowledge base based on a natural language algorithm, and the feedback measure is generated based on the medical method.

9. A hypothermia infection monitoring system based on artificial intelligence, characterized in that: The artificial intelligence-based hypothermia infection monitoring system includes: a collection module for collecting comprehensive medical information of surgical patients, setting a temperature range, extracting a type and occurrence time of an infection event from the comprehensive medical information based on the temperature range, extracting first sample data from all the comprehensive medical information based on the type and occurrence time, setting a time resolution, and generating second sample data corresponding to the first sample data based on the time resolution; a training module, configured to obtain a data type and a data label of the second sample data, pre-process the second sample data based on the data type, extract data features, and train and generate a monitoring model based on the data features; A prediction module, configured to collect monitoring data of the patient to be monitored in real time according to the data tags, pre-process the monitoring data and input it into the monitoring model, and output a predicted event and a predicted occurrence time of the predicted event; The feedback module is used to summarize all the predicted events and the predicted occurrence time, generate a monitoring report, and output feedback measures based on the monitoring report.

10. A computer-readable storage medium having instructions stored thereon, characterized in that: When the instructions are executed by the processor, the artificial intelligence-based hypothermia infection monitoring method as described in any one of claims 1 to 8 is implemented.

Images