Severe illness specialized ability assessment method and system based on machine learning

Through machine learning-based methods, the data of critically ill patients are analyzed using correlation rules mining technology, and the risk prediction model is constructed, which solves the subjectivity and inaccuracy of traditional evaluation methods, and realizes the accurate prediction of risks of critically ill patients and the comprehensive diagnosis and treatment capabilities of critically ill patients. Evaluation has improved the quality of medical services.

CN118919089AInactive Publication Date: 2024-11-08THE FIRST AFFILIATED HOSPITAL OF ANHUI MEDICAL UNIV +1

Patent Information

Application Number
CN202411404834.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-10
Publication Date
2024-11-08
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The traditional comprehensive assessment method for critical care specialists relies on the subjective judgment of experts, lacks objective and quantitative evaluation indicators, and fails to make full use of data mining technology, resulting in the assessment results being incomplete and accurate enough, difficult to update in real time, and unable to meet the rapid response of modern medical care. and lack of accurate predictions of potential patients' risks.

Method used

Using machine learning-based methods, the clinical data of critically ill patients is deeply analyzed through correlation rule mining technology, potential correlations between different factors are discovered, potential risk prediction models are constructed, and the diagnosis and treatment capabilities and service quality of critical care specialists are combined with actual diagnosis and treatment effects. Comprehensive assessment.

Benefits of technology

It realizes accurate prediction of the potential risks of critically ill patients, improves diagnosis and treatment effects, provides personalized preventive measures, improves the quality of life and satisfaction of patients, and provides scientific decision-making support to the hospital management, and promotes medical services. Continuous improvement.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118919089A_ABST
    Figure CN118919089A_ABST
Patent Text Reader

Abstract

The invention discloses a machine learning-based severely-ill specialized ability assessment method and system, and relates to the technical field of medical information, and the method comprises the following specific steps: data collection and preprocessing: collecting clinical data of a severely-ill patient from a hospital information system and an electronic medical record system, according to the method, deep analysis and feature selection are performed on clinical data of critical patients by combining medical knowledge and a data mining technology, potential association rules among different factors are found by adopting an association rule mining technology, and a potential risk prediction model is constructed through a machine learning algorithm based on the association rules, so that the risk prediction accuracy is improved. The possible complication risk of the critical patient is predicted in advance, and personalized prevention measures and suggestions are given. Compared with a traditional method, the method is more accurate and personalized, the diagnosis and treatment effect can be improved, complications are reduced, and the life quality and satisfaction degree of patients are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of medical information technology, and in particular to a critical care specialist capability assessment method and system based on machine learning. Background Art

[0002] In the field of medical health, with the rapid development of medical technology and the in-depth advancement of medical informatization, medical health data has shown an explosive growth trend. These data not only include patients' basic information, diagnosis results, and treatment process, but also cover a wealth of prognostic information and clinical feedback, providing valuable resources for the continuous improvement of medical services and personalized treatment. Especially in the field of critical care medicine, patients' clinical data are often more complex and changeable, which places higher requirements on the timeliness and accuracy of diagnosis and treatment.

[0003] Traditional comprehensive assessment methods for critical care specialist capabilities often rely on the subjective judgment and experience of experts, and lack objective and quantitative assessment indicators and systems. In particular, traditional methods fail to fully utilize the advantages of association rule mining, an advanced data mining technology. On the one hand, traditional methods ignore the potential correlation between different factors in the data, resulting in incomplete and inaccurate assessment results. On the other hand, traditional methods are inefficient when processing massive amounts of data, making it difficult to update assessment results in real time and unable to meet the rapid response needs of modern medical services. In addition, traditional assessment methods lack accurate predictions of potential risks to patients and cannot provide timely decision-making support for medical staff.

[0004] In response to the above problems, it is necessary to optimize the existing comprehensive assessment methods and systems for critical care specialist capabilities, by deeply mining the potential association rules in the clinical data, treatment process and prognosis information of critically ill patients, and using machine learning algorithms to achieve early prediction of patients' future risks. At the same time, combined with the actual diagnosis and treatment effects, a comprehensive assessment of the diagnosis and treatment capabilities and service quality of critical care specialists is conducted to improve the diagnosis and treatment effects. Therefore, it is of great significance to develop a critical care specialist capability assessment method and system based on machine learning that can comprehensively realize the above characteristics. Summary of the invention

[0005] The purpose of the present invention is to make up for the shortcomings of the existing technology and provide a critical care specialist capacity assessment method and system based on machine learning. It can deeply analyze the clinical data, treatment process and prognosis information of critically ill patients through association rule mining technology, discover the potential correlation between different factors, and establish a model through machine learning algorithm based on this to predict the possible risks of patients in advance. At the same time, the present invention also combines the actual diagnosis and treatment effects to comprehensively evaluate the diagnosis and treatment capabilities and service quality of critical care specialists, providing a scientific basis for hospitals to continuously improve and enhance the level of medical services.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: On the one hand, a critical care specialist capability assessment method based on machine learning, the method comprises the following specific steps:

[0007] Data collection and preprocessing: Collect clinical data of critically ill patients from hospital information systems and electronic medical record systems, including but not limited to basic patient information, symptom descriptions, examination results, treatment plans, medication records, and prognosis, and process the collected data;

[0008] Feature selection and conversion: Based on medical knowledge and data mining experience, select features that are important for risk assessment of critically ill patients, including symptom combinations, vital signs indicators, and treatment responses, and convert the original data into a data format suitable for association rule mining;

[0009] Association rule mining and potential risk prediction: Association rule mining technology is used to conduct in-depth mining of preprocessed data to discover potential association rules between different factors, and support and confidence thresholds are set to screen out association rules with high support and confidence. Based on the mined association rules, a potential risk prediction model is constructed to predict the future risks of critically ill patients;

[0010] Comprehensive assessment of critical care specialist capabilities: Based on potential risk prediction results and actual diagnosis and treatment effects, a comprehensive assessment of the critical care specialist's diagnosis and treatment capabilities and service quality is conducted.

[0011] Furthermore, in the feature selection and conversion step, based on medical literature and expert experience, features that are important for risk assessment of critically ill patients are selected, the Spearman rank correlation coefficient is used to screen the features, and the original data is converted into a data format suitable for association rule mining, including discretizing continuous variables into categorical variables, encoding categorical variables, and converting patient records into transaction data form.

[0012] Furthermore, in the feature selection and conversion step, based on medical literature and expert experience, features that are important for risk assessment of critically ill patients are selected, and the Spearman rank correlation coefficient is used to screen the features, and the algorithm formula is: ,in, represents the number of samples, that is, the number of samples of critically ill patients participating in feature selection, Represents the difference between the ranks of two variables, namely a potential risk assessment feature and the risk level of critically ill patients, is the Spearman rank correlation coefficient.

[0013] Furthermore, the association rule mining and potential risk prediction step selects an association rule mining algorithm according to the data scale and characteristics, and sets a support threshold and a confidence threshold. The data is mined by applying the algorithm and parameters to generate association rules, and meaningful association rules are screened out according to the support and confidence thresholds. At the same time, the medical significance of the rules is evaluated, and rules with no practical significance are excluded. Based on the mined association rules, a potential risk prediction model is constructed, and the prediction model is applied to newly admitted critically ill patients or patients undergoing treatment to evaluate their potential risks and output prediction results.

[0014] Furthermore, in the association rule mining and potential risk prediction step, the Apriori algorithm is selected to mine the data according to the data scale and characteristics, and the support threshold and confidence threshold are set. Specifically, all frequent 1-item sets in the data set, that is, single items whose support is greater than the minimum support threshold, are found, and candidate 2-item sets are generated using these frequent 1-item sets, and their support is calculated. The calculation formula for the support is: ,in, is the number of transactions (or records) in the dataset that contain itemset X, is the total number of transactions in the data set. Repeat this process to gradually generate candidate k-item sets until no more frequent item sets can be generated. For each frequent item set, calculate its support and the confidence of the association rules formed with other item sets. The calculation formula is: ,in, It is an item set and item sets The ratio of the number of transactions occurring simultaneously to the total number of transactions, It is an item set The ratio of the number of transactions to the total number of transactions is used to filter out association rules that meet the conditions based on the confidence threshold.

[0015] Furthermore, in the association rule mining and potential risk prediction step, a potential risk prediction model is constructed based on the mined association rules. Specifically, the potential risk prediction model is constructed using a logistic regression equation, and its algorithm formula is: ,in, Indicates that given the independent variable Under the condition of The probability of is a weight vector, which indicates the influence of each independent variable on the probability of risk occurrence. is the independent variable vector, is the bias term, is the base of the natural logarithm, and the weight vector of the model is determined by the training data set and the bias term , and use the test data set to evaluate the model. According to the evaluation results, the model parameters are adjusted through the algorithm to optimize the model performance. The data to be predicted is input into the trained model, the probability of occurrence of potential risks is calculated, and risk warning or decision support is carried out according to the probability value.

[0016] Furthermore, in the association rule mining and potential risk prediction steps, the model parameters are adjusted by the algorithm according to the evaluation results to optimize the model performance. The algorithm formula is: ,in, is the number of training samples, It is The true labels of samples, It is The feature vector of the samples, is the predicted value, Represents model parameters.

[0017] Furthermore, in the step of comprehensive assessment of the critical care specialist's capabilities, a comprehensive assessment of the critical care specialist's diagnostic and treatment capabilities and service quality is conducted based on the potential risk prediction results and actual diagnostic and treatment effects, and the algorithm formula is: ,in, is the potential risk prediction accuracy, is the success rate of complication prevention, is patient satisfaction, Efficiency of medical resource utilization, is the weight of each indicator.

[0018] On the other hand, a critical care specialist capability assessment system based on machine learning is characterized in that the system comprises the following components: a data collection and preprocessing module, a feature selection and conversion module, an association rule mining and potential risk prediction module, and a critical care specialist capability comprehensive assessment module;

[0019] The data collection and preprocessing module collects clinical data of critically ill patients from the hospital information system and electronic medical record system, including but not limited to basic information of patients, symptom descriptions, examination results, treatment plans, medication records and prognosis, and processes the collected data;

[0020] The feature selection and conversion module selects features that are important for risk assessment of critically ill patients, including symptom combinations, vital sign indicators, and treatment responses, based on medical knowledge and data mining experience, and converts raw data into a data format suitable for association rule mining;

[0021] The association rule mining and potential risk prediction module uses association rule mining technology to perform in-depth mining on preprocessed data, discover potential association rules between different factors, set support and confidence thresholds, screen out association rules with high support and confidence, and build a potential risk prediction model based on the mined association rules to predict the future risks of critically ill patients;

[0022] The comprehensive assessment module for critical care unit capabilities conducts a comprehensive assessment of the critical care unit's diagnosis and treatment capabilities and service quality based on potential risk prediction results and actual diagnosis and treatment effects.

[0023] Compared with the existing technology, this critical care specialist capability assessment method and system based on machine learning has the following beneficial effects:

[0024] 1. The present invention combines medical knowledge and data mining technology to conduct in-depth analysis and feature selection of the clinical data of critically ill patients, and uses association rule mining technology to discover potential association rules between different factors. Based on these association rules, a potential risk prediction model is constructed through a machine learning algorithm to predict the risk of complications that may occur in critically ill patients in advance, and give personalized preventive measures. Compared with traditional methods, it is more accurate and personalized, which helps to improve the diagnosis and treatment effects, reduce the occurrence of complications, and improve the quality of life and satisfaction of patients.

[0025] 2. The present invention comprehensively measures the service performance of specialists by setting evaluation indicators. This regular and systematic evaluation mechanism can timely discover problems and deficiencies in specialist services and provide decision-making support for hospital management. At the same time, by comparing the evaluation results of different time periods or different specialties, it can motivate medical staff to continuously improve their professional skills and service levels and promote continuous improvement of medical quality.

[0026] Other advantages, objectives and features of the present invention will be set forth in part in the following description and, in part, will be apparent to those skilled in the art based on an examination of the following or may be taught from the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0028] Figure 1 This is a flowchart of a critical care specialist competency assessment method based on machine learning;

[0029] Figure 2 This is a flowchart of association rule mining and potential risk prediction in a critical care specialist capacity assessment method based on machine learning. DETAILED DESCRIPTION

[0030] The technical solutions in the embodiments of the present invention are described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0031] Embodiment 1

[0032] This embodiment describes in detail the application process of a critical care specialist capacity assessment method based on machine learning. By deeply mining the clinical data of critically ill patients, potential association rules between different factors are discovered, and a potential risk prediction model is constructed based on these rules. Finally, combined with the actual diagnosis and treatment effects, a comprehensive assessment of the diagnosis and treatment capabilities of critical care specialists is conducted to improve medical quality and patient satisfaction.

[0033] In the data collection and preprocessing steps, clinical data of critically ill patients are collected from the hospital information system (HIS) and electronic medical record system (EMR). These data include but are not limited to basic information of patients (such as age, gender, medical history, etc.), symptom descriptions, test results (such as blood tests, imaging examinations, etc.), treatment plans, medication records and prognosis, etc. The collected data are cleaned to remove noise, outliers and duplicate values, and the cleaned data are converted into a format suitable for association rule mining, which usually includes discretizing continuous variables (such as age, blood pressure, etc.) into categorical variables (such as age group, blood pressure level), encoding categorical variables (such as using one-hot encoding or label encoding), and converting patient records into transaction data form. Each transaction contains all relevant features of the patient's visit or a period of time.

[0034] In the feature selection and transformation step, based on medical literature and expert experience, features that are important for risk assessment of critically ill patients are selected. These features may include specific symptom combinations, vital signs (such as heart rate, blood pressure, blood oxygen saturation, etc.) and treatment responses (such as drug sensitivity, symptom improvement after treatment, etc.). The Spearman rank correlation coefficient is used to further screen features and calculate the correlation between each potential feature and the risk level of critically ill patients. The algorithm formula is: ,in, represents the number of samples, that is, the number of samples of critically ill patients participating in feature selection, Represents the difference between the ranks of two variables, namely a potential risk assessment feature and the risk level of critically ill patients, is the Spearman rank correlation coefficient. Features with higher correlation are selected as model inputs, and the screened features are converted into a data format suitable for Apriori algorithm mining, which usually includes discretizing continuous variables into categorical variables (if not discretized yet) and encoding categorical variables.

[0035] In the association rule mining and potential risk prediction steps, the Apriori algorithm is used to mine association rules on the preprocessed data. First, all frequent 1-item sets in the data set are found, that is, single items whose support is greater than the minimum support threshold. Then, these frequent 1-item sets are used to generate candidate 2-item sets and their support is calculated. This process is repeated to gradually generate candidate k-item sets until no more frequent item sets can be generated. For each frequent item set, its support and the confidence of the association rules formed with other item sets are calculated. According to the set confidence threshold, the association rules that meet the conditions are screened out. These rules reveal the potential associations between different factors and are of great significance for the risk assessment of critically ill patients. Based on the mined association rules, a potential risk prediction model is constructed, the model parameters are determined through the training data set, and the model performance is evaluated using the test data set. The data to be predicted is input into the trained model, the probability of occurrence of potential risks is calculated, and risk warnings or decision support are performed based on the probability value, providing timely and accurate risk prompts for medical staff.

[0036] In the comprehensive assessment step of critical care specialist capabilities, a series of evaluation indicators are set to comprehensively evaluate the diagnosis and treatment capabilities of critical care specialists based on the potential risk prediction results and actual diagnosis and treatment effects. These indicators may include the accuracy of potential risk prediction, the success rate of complication prevention, patient satisfaction, and the efficiency of medical resource utilization. The weighted summation method is used to quantify the scores of various indicators and calculate the weighted comprehensive scores. The algorithm formula is as follows: ,in, is the potential risk prediction accuracy, is the success rate of complication prevention, is patient satisfaction, Efficiency of medical resource utilization, It is the weight of each indicator. According to the comprehensive score, the diagnosis and treatment capabilities of critical care specialists are ranked or graded, and the evaluation results are presented to the hospital management and medical staff in the form of a report. By comparing and analyzing the evaluation results of different time periods or different specialties, existing problems and deficiencies are identified, and improvement measures and suggestions are proposed. At the same time, the evaluation results are used as an important basis for hospital performance appraisal and resource allocation, to promote the continuous improvement and enhancement of medical quality.

[0037] In summary, this embodiment combines medical knowledge and data mining technology to conduct in-depth analysis and feature selection of the clinical data of critically ill patients, and adopts association rule mining technology to discover potential association rules between different factors. Based on these association rules, a potential risk prediction model is constructed through a machine learning algorithm to predict the risk of complications that may occur in critically ill patients in advance, and give personalized preventive measures. Compared with traditional methods, it is more accurate and personalized, which helps to improve the diagnosis and treatment effects, reduce the occurrence of complications, and improve the quality of life and satisfaction of patients.

[0038] Embodiment 2

[0039] Based on the first embodiment, this embodiment describes in detail the specific application process of association rule mining and potential risk prediction in a critical care specialist capability assessment method based on machine learning.

[0040] The Apriori algorithm is used to mine association rules on the preprocessed data. First, all frequent 1-item sets in the data set are found, that is, single items whose support is greater than the minimum support threshold. Then, these frequent 1-item sets are used to generate candidate 2-item sets and their support is calculated. The calculation formula for the support is: ,in, The data set contains the itemset The number of transactions (or records) is the total number of transactions in the data set. Repeat this process to gradually generate candidate k-item sets until no more frequent item sets can be generated.

[0041] For each frequent item set, calculate its support and the confidence of the association rules formed with other item sets. According to the set confidence threshold, the calculation formula is: ,in, It is an item set and item sets The ratio of the number of transactions occurring simultaneously to the total number of transactions, It is an item set The proportion of the number of transactions that occur to the total number of transactions is used to screen out association rules that meet the conditions based on the confidence threshold. These rules reveal the potential associations between different factors and are of great significance for risk assessment of critically ill patients.

[0042] Based on the mined association rules, a potential risk prediction model is constructed. Specifically, the potential risk prediction model is constructed using the logistic regression equation, and its algorithm formula is: ,in, Indicates that given the independent variable Under the condition of The probability of is a weight vector, which indicates the influence of each independent variable on the probability of risk occurrence. is the independent variable vector, is the bias term, is the base of the natural logarithm, and the weight vector of the model is determined by the training data set and the bias term , and use the test dataset to evaluate the model.

[0043] According to the evaluation results, the model parameters are adjusted through the algorithm to optimize the model performance. The algorithm formula is: ,in, is the number of training samples, It is The true labels of samples, is the feature vector of the i-th sample, is the predicted value, Represents model parameters, inputs the data to be predicted into the trained model, calculates the probability of occurrence of potential risks, and performs risk warning or decision support based on the probability value.

[0044] In summary, this embodiment screens out high-value association rules by setting reasonable support and confidence thresholds. These association rules not only reveal the inherent laws of disease development, but also provide a scientific basis for risk assessment. The risk prediction model constructed based on these association rules can predict the risk of complications that may occur in critically ill patients in advance, and provide strong support for medical staff to formulate personalized diagnosis and treatment plans.

[0045] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. A critical care specialist capability assessment method based on machine learning, characterized in that: The method comprises the following specific steps: Data collection and preprocessing: Collect clinical data of critically ill patients from hospital information systems and electronic medical record systems, including but not limited to basic patient information, symptom descriptions, examination results, treatment plans, medication records, and prognosis, and process the collected data; Feature selection and conversion: Based on medical knowledge and data mining experience, select features that are important for risk assessment of critically ill patients, including symptom combinations, vital signs indicators, and treatment responses, and convert the original data into a data format suitable for association rule mining; Association rule mining and potential risk prediction: Association rule mining technology is used to conduct in-depth mining of preprocessed data to discover potential association rules between different factors, and support and confidence thresholds are set to screen out association rules with high support and confidence. Based on the mined association rules, a potential risk prediction model is constructed to predict the future risks of critically ill patients; Comprehensive assessment of critical care specialist capabilities: Based on potential risk prediction results and actual diagnosis and treatment effects, a comprehensive assessment of the critical care specialist's diagnosis and treatment capabilities and service quality is conducted.

2. According to claim 1, a critical care specialist capability assessment method based on machine learning is characterized in that: In the feature selection and conversion step, features that are important for risk assessment of critically ill patients are selected based on medical literature and expert experience, the Spearman rank correlation coefficient is used to screen the features, and the original data is converted into a data format suitable for association rule mining, including discretizing continuous variables into categorical variables, encoding categorical variables, and converting patient records into transaction data form.

3. A critical care specialist capability assessment method based on machine learning according to claim 2, characterized in that: In the feature selection and conversion step, based on medical literature and expert experience, features that are important for risk assessment of critically ill patients are selected, and the Spearman rank correlation coefficient is used to screen the features. The algorithm formula is: ,in, represents the number of samples, that is, the number of samples of critically ill patients participating in feature selection, Represents the difference between the ranks of two variables, namely a potential risk assessment feature and the risk level of critically ill patients, is the Spearman rank correlation coefficient.

4. The critical care specialist capability assessment method based on machine learning according to claim 1, characterized in that: The association rule mining and potential risk prediction step selects an association rule mining algorithm according to the data scale and characteristics, and sets a support threshold and a confidence threshold. The data is mined by applying the algorithm and parameters to generate association rules. Meaningful association rules are screened out according to the support and confidence thresholds, and the medical significance of the rules is evaluated at the same time, and rules without practical significance are excluded. Based on the mined association rules, a potential risk prediction model is constructed, and the prediction model is applied to newly admitted critically ill patients or patients undergoing treatment to evaluate their potential risks and output prediction results.

5. The critical care specialist capability assessment method based on machine learning according to claim 4, characterized in that: In the association rule mining and potential risk prediction step, the Apriori algorithm is selected to mine the data according to the data scale and characteristics, and the support threshold and confidence threshold are set. Specifically, all frequent 1-item sets in the data set, that is, single items whose support is greater than the minimum support threshold, are found, and candidate 2-item sets are generated using these frequent 1-item sets, and their support is calculated. The calculation formula for the support is: ,in, is the number of transactions in the dataset that contain itemset X, is the total number of transactions in the data set. Repeat this process to gradually generate candidate k-item sets until no more frequent item sets can be generated. For each frequent item set, calculate its support and the confidence of the association rules formed with other item sets. The calculation formula is: ,in, It is an item set and item sets The ratio of the number of transactions occurring simultaneously to the total number of transactions, It is an item set The ratio of the number of transactions to the total number of transactions is used to filter out association rules that meet the conditions based on the confidence threshold.

6. The critical care specialist capability assessment method based on machine learning according to claim 4, characterized in that: In the association rule mining and potential risk prediction step, a potential risk prediction model is constructed based on the mined association rules. Specifically, the potential risk prediction model is constructed using a logistic regression equation, and its algorithm formula is: ,in, Indicates that given the independent variable Under the condition of The probability of is a weight vector, which indicates the influence of each independent variable on the probability of risk occurrence. is the independent variable vector, is the bias term, is the base of the natural logarithm, and the weight vector of the model is determined by the training data set and the bias term , and use the test data set to evaluate the model. According to the evaluation results, the model parameters are adjusted through the algorithm to optimize the model performance. The data to be predicted is input into the trained model, the probability of occurrence of potential risks is calculated, and risk warning or decision support is carried out according to the probability value.

7. The method for evaluating critical care specialist capabilities based on machine learning according to claim 6, characterized in that: In the association rule mining and potential risk prediction steps, the model parameters are adjusted by the algorithm according to the evaluation results to optimize the model performance. The algorithm formula is: ,in, is the number of training samples, It is The true labels of samples, It is The feature vector of the samples, is the predicted value, Represents model parameters.

8. The critical care specialist capability assessment method based on machine learning according to claim 1, characterized in that: In the step of comprehensive assessment of critical care specialist capabilities, a comprehensive assessment of the critical care specialist's diagnostic and treatment capabilities and service quality is conducted based on the potential risk prediction results and actual diagnostic and treatment effects. The algorithm formula is: ,in, is the potential risk prediction accuracy, is the success rate of complication prevention, is patient satisfaction, Efficiency of medical resource utilization, is the weight of each indicator.

9. A critical care specialist capability assessment system based on machine learning, characterized in that: The system comprises the following components: a data collection and preprocessing module, a feature selection and conversion module, an association rule mining and potential risk prediction module, and a comprehensive assessment module of critical care specialist capabilities; The data collection and preprocessing module collects clinical data of critically ill patients from the hospital information system and electronic medical record system, including but not limited to basic information of patients, symptom descriptions, examination results, treatment plans, medication records and prognosis, and processes the collected data; The feature selection and conversion module selects features that are important for risk assessment of critically ill patients, including symptom combinations, vital sign indicators, and treatment responses, based on medical knowledge and data mining experience, and converts raw data into a data format suitable for association rule mining; The association rule mining and potential risk prediction module uses association rule mining technology to perform in-depth mining on preprocessed data, discover potential association rules between different factors, set support and confidence thresholds, screen out association rules with high support and confidence, and build a potential risk prediction model based on the mined association rules to predict the future risks of critically ill patients; The comprehensive assessment module for critical care unit capabilities conducts a comprehensive assessment of the critical care unit's diagnosis and treatment capabilities and service quality based on potential risk prediction results and actual diagnosis and treatment effects.

Citation Information

Patent Citations

  • Risk identification method, risk identification device, electronic equipment and storage medium thereof

    CN116385130A

  • Risk assessment model and device capable of explaining death of severe children and establishment method

    CN116543902A

  • Hypertension risk assessment model construction method, diet therapy and health management system

    CN117012392A

  • Medical information analysis method and system based on big data

    CN118315066A

  • Disease specific ontology-guided rule engine and machine learning for enhanced critical care decision support

    US20190057774A1

Cited By

  • Metadata model-based hospital operation index prediction method, system and equipment

    CN119153054A

  • Multi-source data fusion-based multi-point triggering intelligent early warning method for infectious diseases

    CN119274820A

  • Structural causal model construction method and device for acute myocardial infarction monitoring

    CN119418949A

  • A structural causal model construction method and device for acute myocardial infarction monitoring

    CN119418949B

  • Quality evaluation system for preparing superconducting material and preparation process thereof

    CN119624247A