38 results about "Diagnosis standards" patented technology

The invention discloses a skin cancer pathological image analysis system and method based on a Transform model. The system comprises an image acquisition and quality control module, a multi-scale preprocessing module, a hierarchical feature extraction module, an intelligent diagnostic reasoning module, a knowledge graph aided decision-making module, a result output and feedback module and a federal learning training module. According to the method, multi-spectral image acquisition and quality control are carried out, multi-modal enhancement and dyeing standardization preprocessing are carried out, pathological features are extracted in a layered manner by using an improved Transform model, subtype identification and malignancy degree evaluation are realized in combination with multi-task learning, uncertainty is quantified by means of Monte Carlo dropout, cases and guidelines are associated through a knowledge graph, privacy is protected through federal learning, and the model is optimized. According to the scheme, the diagnosis efficiency and accuracy are greatly improved, the model interpretability is enhanced, various clinical scenes are adapted, diagnosis standardization is promoted, and improvement of basic medical capacity is assisted.

The application discloses a diagnosis data processing method and device and a storage medium, relates to the technical field of data processing, and comprises the following steps: acquiring a heart-lung auscultation sound signal collected by a collection device and a traditional Chinese medicine (TCM) diagnosis standard description received based on an interactive interface; determining a western medicine quantitative diagnosis index according to the heart-lung auscultation sound signal, and correlating a TCM syndrome differentiation type according to the TCM diagnosis standard description and the western medicine quantitative diagnosis index; and generating and outputting a TCM medication recommendation scheme corresponding to a diagnosis result according to the western medicine quantitative diagnosis index and / or the TCM syndrome differentiation type. The application provides strong objective evidence and quantitative reference by differentiating the TCM syndrome differentiation type according to the auscultation sound. When diagnosing and treating diseases, the TCM highly depends on and integrates the auscultation result, and takes the auscultation result as one of key bases for differentiation, prescription selection, treatment effect evaluation and prognosis.

The invention discloses an AI-based schizophrenia clinical thinking training simulation system, which relates to the technical field of artificial intelligence, and is characterized by comprising the steps of constructing a dynamic knowledge base, receiving and preprocessing multi-modal input, performing standardized mapping by extracting keywords, and then performing judgment of a rule layer and a semantic layer. A student question method and a standard answer method in the dynamic knowledge base are sent into a model for response generation, then four-dimensional quantitative scoring and clinical operation multi-dimensional scoring are carried out, error codes and positioning sentences thereof are output, and finally personalized training courses are recommended to students; and hot-updating the dynamic knowledge base continuous optimization model. According to the method, four dimensions of the ICD-10 diagnostic standard are converted into a quantitative scoring algorithm capable of being automatically executed, quantitative assessment of clinical operation specifications is integrated, a diagnosis and operation double-track evaluation system is constructed, the problems that a traditional method is lack of evaluation standards and high in subjectivity are solved, and the evaluation efficiency is improved. Therefore, the evaluation of the clinical thinking ability of the medical students becomes objective, comprehensive and evidence-based.

The invention relates to an intelligent tongue picture analysis system and method. The system comprises a high-definition camera, an image processing unit, an image segmentation module, a color space conversion module, a feature extraction module, an analysis module, a classification module and a diagnosis matching module. The system shoots a tongue image through the high-definition camera, the image processing unit carries out image preprocessing, the image segmentation module separates a tongue area, the color space conversion module extracts color features, and the feature extraction module calculates color mean and variance and shape features. The analysis module analyzes tongue coating distribution, thickness and color, the classification module performs feature classification based on a training data set, and the diagnosis matching module matches the features with a diagnosis standard to obtain a conclusion. The system provides scientific and objective tongue picture analysis, improves the diagnosis efficiency, and promotes the modernization and intellectualization of traditional Chinese medicine diagnosis.

The application discloses an out-of-hospital hierarchical diagnosis method based on IBS diseases, and relates to the technical field of medical diagnosis, and comprises the following steps: S1: acquiring multi-dimensional information related to IBS diseases provided by a patient in an out-of-hospital scene; S2: performing standardization preprocessing on the out-of-hospital multi-dimensional information to obtain feature data that can be input into an algorithm model; and S3: inputting the standardization feature data into a preset IBS hierarchical diagnosis algorithm model to output an IBS disease hierarchical diagnosis result of the patient. The out-of-hospital hierarchical diagnosis method based on IBS diseases integrates multi-dimensional out-of-hospital information, eliminates data errors by combining standardization preprocessing, adopts multi-feature screening and model optimization, references the Rome IV diagnostic standard to train the model, simultaneously judges the diagnosis reliability through the confidence, reduces the misdiagnosis and missed diagnosis rate, the model iteration updating mechanism can continuously improve the diagnosis accuracy, and adapts to the change of clinical requirements.

Embodiments of the present application disclose a disease diagnosis standard library construction method, device and equipment and a storage medium. The method can obtain first historical medical data from a target disease sample library; first image feature data of the first historical medical data is extracted to obtain a first key feature training data set, first short-term memory parameters and first long-term memory parameters are obtained based on the first key feature training data set and a detection standard model; second historical medical data is obtained; second image feature data of the second historical medical data is extracted to obtain a second key feature training data set, second short-term memory parameters and second long-term memory parameters are obtained based on the second key feature training data set and the detection standard model; target disease diagnosis standard data in a time dimension is obtained based on the short-term memory parameters and the long-term memory parameters, and a disease diagnosis standard library is constructed based on the target disease diagnosis standard data. The scheme can improve the accuracy of disease detection.

The invention discloses a system for endometriosis diagnosis and a computer storage medium. Aiming at major adjustment of endometriosis diagnosis standards, biomarkers and ploidy information of circulating endometrial cells in easily available samples such as peripheral blood of a patient are detected, and comprehensive risk assessment is performed by utilizing an optimization algorithm model, so that the diagnosis accuracy of endometriosis is improved. The invention provides an objective and quantifiable noninvasive diagnostic tool adapted to the adjusted standard, and the method solves the problem of missed diagnosis caused by false negative in the existing surgical diagnostic method, and has important clinical application value, social benefit and market prospect for improving the overall diagnostic level of endometriosis.

The invention belongs to the technical field of transmission casing bearing fault diagnosis, and particularly relates to a typical small bearing weak fault comprehensive diagnosis standard construction method, which comprises the following steps: respectively carrying out inner ring, outer ring and rolling body bearing fault simulation tests; a filtering frequency band is selected based on a weighted kurtosis value improved Protugram algorithm, an envelope spectrum is extracted after filtering and noise reduction, and a bearing fault diagnosis standard threshold value is established based on an envelope spectrum characteristic value method; based on a preset fault test, comprehensively analyzing envelope spectrum characteristic values of different faults and different measuring points, and establishing a comprehensive early warning diagnosis standard; based on a failure fault test, envelope spectrum characteristic values of different faults and different measuring points are comprehensively analyzed, and a comprehensive failure diagnosis standard is established. And after judging that the bearing has no fault, respectively carrying out a preset fault test and a failure fault test, and establishing an early warning diagnosis threshold value and a failure standard threshold value for the fault characteristic sensitive measuring points on the basis of measuring point optimization analysis, thereby realizing early warning and failure judgment of the bearing fault.

The invention relates to the field of livestock breeding, and discloses an automatic dairy cow veterinary management and analysis system, which comprises a standard management module, a disease declaration module, a treatment execution module and a data analysis module, and the modules realize data interaction and instruction transmission through a unified data transmission protocol. The reference module is used for providing a standardized disease diagnosis and treatment scheme reference and supporting dynamic updating; the disease declaration module is used for realizing accurate input and case filing of cattle disease information; and the treatment execution module is used for matching a standard prescription according to the case information and executing a traceable treatment process. According to the system, a'diagnosis-treatment 'whole-process standardization system is constructed through the standard management module, a standard disease library defines disease diagnosis standards from three dimensions of'typical symptoms, detection indexes and identification key points', an attached symptom picture library assists visual judgment, and misdiagnosis and missed diagnosis caused by experience differences are reduced.

The invention discloses a sepsis risk multi-modal prediction method and system based on deep learning, and belongs to the technical field of intelligent analysis of medical data, and the method comprises the following steps: a multi-modal data collection step: collecting vital sign sequences, bedside instant inspection, laboratories, microbial culture, images and medical history data; in the time sequence feature coding step, a time sequence embedded vector is extracted through a multi-scale time sequence coding network; the cross-modal fusion prediction step is based on the dynamic attention mechanism fusion features and predicts the sepsis occurrence probability; the organ function dynamic evaluation step is used for estimating a missing test result and calculating an organ function score; the risk layering early warning step outputs four-level risk layering results and feeds back optimized coding parameters, early warning can be output 4-8 hours before the diagnosis standard is met, and compared with qSOFA, the AUC is improved by 0.15 or above.

The invention discloses a plaque psoriasis severity grading system and a plaque psoriasis severity grading method. The plaque psoriasis severity grading system comprises a case inclusion module used for selecting patients meeting plaque psoriasis western medicine diagnosis standards; the ultrasonic examination module is used for carrying out ultrasonic examination on the typical diseased part of the patient by an ultrasonic physician to obtain a high-frequency ultrasonic index; the scoring module evaluates the skin lesion of the patient and gives an improved PASI score of local lesion; the model building module takes the local improved PASI score as a dependent variable, takes a high-frequency ultrasonic index as an independent variable, adopts a stepping method to screen out factors having significant influence on the local improved PASI score, and builds a severity grading model; the verification module performs internal verification on the constructed severity grading model by using a neural network multilayer perceptron, and evaluates the prediction performance of the model; and the output module is used for visualization and output. The invention provides a new and effective technical means for grading the severity of the plaque psoriasis, and has important clinical application value and popularization prospect.

The invention discloses a medical examination report analysis system and method based on big data, and relates to the technical field of medical data processing, the system comprises a data acquisition module, a preprocessing module, a storage module, an index correlation analysis module, an abnormity early warning module, a result generation module and a user interaction module, and the method comprises corresponding data processing and analysis steps. Multi-source data such as inspection reports, historical diagnosis and treatment, clinical cases and the like are collected through a standardized interface, are subjected to cleaning, standardization and structural conversion, and are accessed by means of a distributed storage architecture; an Apriori association rule algorithm is adopted to mine index association, a random forest and a neural network model are combined to realize risk assessment, and a three-level grading early warning mechanism is matched to push accurate suggestions. According to the method, traditional manual interpretation is replaced, the analysis efficiency and accuracy are improved, the problems of insufficient associated information mining, non-uniform diagnosis standards and the like are solved, medical institution scenes at all levels are adapted, interactive optimization is supported, clinical precise diagnosis is assisted, the burden of doctors is relieved, and the method has important popularization value for intelligent application of medical data.

The application belongs to the technical field of biological medicine, and particularly relates to application of a reagent specifically binding to OARD1 protein in preparation of a breast cancer diagnosis and / or prognosis evaluation product. The product detects the expression level of OARD1 protein in a breast tissue sample by the reagent specifically binding to OARD1 protein and carries out H-score scoring, takes a preset H-score score as a breast cancer positive diagnosis threshold, and judges whether the sample is breast cancer positive. Moreover, the product can distinguish high / low expression of OARD1 protein by a preset H-score score, constructs a prognosis prediction model in combination with TNM staging of a subject, and realizes evaluation of a breast cancer patient disease progression risk. The method is simple and easy to implement, the diagnosis process is safe and effective, is easy to be accepted by patients, diagnosis standards are unified, and is less affected by personal subjective factors, which shows that OARD1 protein has important significance for diagnosis or prognosis judgment of breast cancer.

The application discloses a kind of intelligent diagnosis and decision support method of coronary heart disease angina TCM syndrome factor based on HIS, it is related to medical information and TCM diagnosis and treatment technical field, the method will coronary heart disease angina TCM syndrome factor diagnostic criteria be deconstructed into standardization diagnostic item, build machine-readable weight rule base containing score, threshold and severity grading rule, response syndrome differentiation operation generates electronic collection form, receives structured data storage to the TCM syndrome factor fact table associated with patient main index and medical record, after structured data is submitted, trigger intelligent computing engine verification essential item, total score is accumulated, determine the establishment of syndrome factor and severity, through TCM diagnosis and treatment knowledge graph push personalized treatment or nursing suggestion in clinical business node, and continuously update rule base and knowledge graph through clinical data feedback.The application integrates corresponding function module to realize and deeply fuse with hospital information system, improves the standardization and efficiency of diagnosis and treatment, and has strong practicality.

The invention provides a health / exercise management system, which is characterized in that the system is applied to cerebrovascular disease supervision, and comprises a health management module used for collecting and recording health information, comparing the health information with a preset diagnosis standard to obtain a comparison result, and outputting a corresponding prompt according to the comparison result; the health management module is also used for collecting related scale questionnaires and recording health conditions according to the related scale and the questionnaires; the health management module is also used for receiving the recorded related diagnosis and treatment results of the cerebrovascular diseases and managing the recorded related diagnosis and treatment results; and the exercise management module is used for carrying out score evaluation on facial paralysis and upper limb and lower limb exercises according to the exercise video to be evaluated, and obtaining a detection result according to the score evaluation. Recording and feedback can be combined, and the universality from health recording to secondary prevention and the compatibility from health follow-up visit to function evaluation are met. User motion function self-test is realized, and user health self-test is facilitated.

The present application relates to the field of medical informatization and artificial intelligence technology, and specifically provides a medical record connotation quality control method and device based on a large language model, which has the following steps: S1, medical record document data access, document content preprocessing; S2, clear quality control content, construct quality control rule library; S3, construct medical knowledge base: based on clinical guidelines and evidence-based medicine evidence base, establish diagnosis standard, drug, treatment project, inspection project knowledge base; S4, entity extraction: use a large language model to extract diagnosis, drug, treatment project and inspection project name; S5, model reasoning: query the relevant information of the extracted entity in the knowledge base, according to the quality control rules, combine the queried relevant information with the medical document and submit them to the large model for reasoning to give the quality control result; S6, problem collection, model optimization. Compared with the prior art, the present application can significantly improve the intelligent level of medical record connotation quality control, effectively assist clinical doctors and quality control personnel to improve the quality of medical record writing, ensure medical safety, and improve hospital management efficiency.

The invention discloses an intelligent diagnosis system and method for multiple myeloma based on multi-modal data fusion, and relates to the technical field of myeloma. Comprising a medical professional module, a digital cloned construction module, a continuous optimization module, an interactive service module, a service requirement module and a basic service module, wherein the medical professional module is responsible for providing professional knowledge support related to multiple myeloma, including clinical manifestations, laboratory examination, diagnostic standards, prognosis layering and therapeutic schedules. According to the intelligent diagnosis system and method for multiple myeloma based on multi-modal data fusion, multi-channel medical knowledge can be acquired and integrated in real time, standardized and structured professional knowledge support is provided through systematic knowledge graph construction and algorithm reasoning, meanwhile, multi-source heterogeneous medical data are comprehensively integrated, and the diagnosis efficiency is improved. And constructing a full-life-cycle data file of the patient so as to provide personalized diagnosis standards and treatment schemes in combination with personal information, medical history and multi-dimensional examination results of the patient.

The invention relates to the technical field of medical artificial intelligence and disease risk prediction, and discloses a pancreatic cancer risk prediction model optimization method based on metabolic syndrome data, and the method comprises the steps: a training stage: obtaining historical clinical data of five physiological abnormalities of a sample individual, carrying out the grading assignment of the historical clinical data based on a metabolic syndrome diagnosis standard, and carrying out the prediction of a pancreatic cancer risk prediction model; in the application stage, the clinical data of a target individual are subjected to same preprocessing and feature engineering and input into the risk prediction model, and the risk prediction model is obtained by combining a pancreatic cancer diagnosis tag and training a logistic regression model through a maximum likelihood estimation method. And calculating to obtain a pancreatic cancer risk assessment result of the target individual. According to the method, data modeling is carried out by innovatively integrating metabolic syndrome data and diabetes disease course information, accurate and early individualized prediction of the pancreatic cancer risk is realized, and an effective tool is provided for clinical screening and intervention.

The invention relates to the technical field of fault diagnosis, and discloses a cross-device power adapter fault diagnosis method and system, and the method comprises the steps: collecting an operation data set of multiple types of devices; extracting a universal feature vector; calculating a deviation value between the general feature vector and a preset normal operation vector, and if the deviation value exceeds a preset deviation tolerance threshold value, marking the general feature vector as an abnormal vector and performing classification to obtain a fault classification label; performing cross-device unified correction on the fault feature vectors matched with the tags, and extracting fault distribution features; in combination with the real-time feature vector, standardized fault modes are extracted and integrated, and an initial diagnosis standard is obtained; performing matching analysis on the real-time operation parameters and an initial diagnosis standard, and obtaining an optimized diagnosis standard according to an analysis result; and monitoring the operation state of each device and performing diagnosis by using the optimized diagnosis standard to obtain a final fault diagnosis result. The method can unify the feature data of different devices under the same type of faults, and flexibly deals with the difference between the devices.

The invention provides a hemodialysis catheter related blood flow infection risk screening system and method, and belongs to the technical field of medical management, and the system comprises an information input module which is used for inputting patient case information; the data acquisition module is used for acquiring multi-source data such as catheter indwelling operation records, catheter indwelling time and daily maintenance records; the risk screening module is used for screening catheter-related blood flow infection risks and determining risk levels; the evaluation module is used for evaluating the patient state and the catheter state; and the diagnosis judgment module is used for judging whether a catheter-related blood flow infection diagnosis standard is met or not and generating treatment intervention notification information. The method has the beneficial effects that by screening and determining the risk of blood flow infection related to the catheter, further evaluating the patient and catheter states, and combining an evaluation result, clinical manifestation and a microbiological examination result, judging whether an infection diagnosis standard is met and generating a treatment intervention notice, the screening efficiency is improved, and the infection occurrence rate is reduced; the treatment effect and prognosis of a patient are improved.

The invention discloses a traditional Chinese medicine diagnosis standardization system and method based on Yin and Yang balance of a twelve-meridian resistor, and relates to the technical field of diagnosis standardization, the traditional Chinese medicine diagnosis standardization system based on Yin and Yang balance of the twelve-meridian resistor comprises a hardware measurement module used for collecting resistance values of the left side and the right side of the twelve-meridian resistor of a human body; the data processing and analyzing module is connected with the hardware measuring module and used for receiving the resistance values collected by the hardware measuring module, processing and analyzing the resistance values and calculating the ratio of the resistance values of the left side and the right side of each right meridian; according to the method, the subjective limitation that traditional four diagnosis methods depend on experience is broken through, abstract concepts of yin-yang imbalance, viscera function strength and the like in traditional Chinese medicine are converted into computable resistance values and ratios, diagnosis results are converted into quantitative analysis from qualitative description, and diagnosis deviation caused by experience difference of doctors is avoided.

The application discloses a kind of equipment fault diagnosis dynamic modeling method, equipment and storage medium.The dynamic modeling method includes the following steps: constructing data interface table, for real-time acquisition and storage of data information to be processed, and the data information to be processed is obtained directly by connecting data interface to the diagnosed equipment;Construct a diagnosis model table, the diagnosis model table includes fault name, diagnosis standard and the data source information corresponding to the diagnosis standard.The application improves the flexibility of the equipment fault diagnosis system, while reducing the threshold for manufacturing enterprises to participate in intelligent manufacturing and equipment fault diagnosis, suitable for most discrete manufacturing enterprises.

The invention discloses a tea tree leaf mineral element deficiency diagnosis method based on hyperspectral imaging and DRIS diagnosis. The method comprises the following steps: S1, collecting a sample; s2, acquiring and preprocessing hyperspectral data; s3, screening characteristic wave bands; s4, measuring the contents of various mineral elements in the sample in the S1, and establishing a basic database; s5, based on the data in S3 and S4, utilizing a machine learning algorithm to respectively construct inversion models of the multiple mineral elements, performing performance evaluation on the inversion models, and screening to obtain an optimal inversion model of the multiple mineral elements; s6, hyperspectral data of tea tree leaves to be detected are input into the optimal inversion model obtained in the S5, the corresponding mineral element content is predicted, and in combination with the DRIS diagnosis standard, diagnosis on whether the mineral elements of the tea tree leaves are in a lack state or not is completed. The invention provides an innovative method which organically combines hyperspectral nondestructive testing, machine learning intelligent modeling and DRIS comprehensive diagnosis standards, and realizes rapid, nondestructive, accurate and comprehensive diagnosis of the mineral nutrition status of the tea trees.

The present application relates to the technical field of blood lipid detection, and particularly relates to a blood lipid rapid detection method based on spectrum analysis. The method comprises the following steps: S1, judging the history of blood lipid abnormality; S2, investigating the multi-source interference factors; S3, analyzing the detection indexes; and S4, adjusting the preset examination period. The present application obtains the preset blood lipid diagnosis standard, historical detection data, demographic and physical examination information and spectrum detection indexes, and realizes the multi-element judgment on the existing blood lipid abnormality and the risk of blood lipid increase by fitting the blood lipid prediction curve and the health curve in real time, combining the multi-dimensional interference factor and the comorbidity correlation degree evaluation, and evaluating and selecting the corresponding detection mode according to the peak area and the peak displacement. The present application can effectively distinguish the primary and secondary hyperlipidemia, and dynamically adjust the detection period by selecting the infrared or metal Raman spectrum detection mode, so as to improve the detection efficiency and the early warning accuracy, and provide multiple choices for the clinical individual intervention scheme.

The application discloses an Alzheimer and frontotemporal dementia diagnosis device based on electroencephalogram, wherein the device comprises: collecting electroencephalogram data of a patient to be diagnosed; converting the electroencephalogram data into CSD images and PSD images by using an MNE library; expanding the CSD images and the PSD images by adopting a bilinear interpolation method to obtain CSD and PSD images with consistent width and height; merging the expanded CSD images and the PSD images to form a three-dimensional tensor with eight channels; inputting the three-dimensional tensor into a diagnosis model, and performing prediction based on the diagnosis model to obtain a prediction result; and classifying and comparing the prediction result with diagnosis standards of Alzheimer and frontotemporal dementia to obtain the prediction result and diagnose the patient. The method can efficiently and accurately analyze and diagnose electroencephalogram signals by optimizing the structure and parameters of the model, so that more accurate, rapid and objective disease diagnosis is realized.

The invention discloses application of Der p sIgE and / or Der f sIgE in nasal secretion in preparation of a kit or a detection platform for diagnosing local allergic rhinitis, and relates to the technical field of in vitro diagnosis and immunoassay. The invention relates to application of Der p sIgE and / or Der f sIgE in nasal secretion in preparation of a kit or a detection platform for diagnosing local allergic rhinitis. The invention proposes and verifies that the dust mite sIgE in the nasal secretion can be used as an effective biomarker for diagnosing LAR for the first time, and solves the technical problem of difficulty in typing serum sIgE negative rhinitis patients clinically. Detection can be completed only by collecting nasal secretions, operation is easy and convenient, safety is high, and clinical popularization is easy. According to the method, diagnosis is objective and accurate, a diagnosis critical value is scientifically established through research on healthy people, a quantitative objective diagnosis standard is provided, it is verified through the cell factor level that the diagnosis result is highly consistent with the intrinsic biological mechanism of allergic inflammation, and the accuracy is high.

The invention discloses a saliva flow rate detection strip for auxiliary diagnosis of xerostomia and a detection method thereof. The detection strip comprises strip-shaped filter paper pretreated by fluorescein sodium, the filter paper is packaged in a food-grade plastic shell, one end of the filter paper extends out of the shell to form a 0.5 cm sampling part, and scale marks are arranged on the filter paper. During detection, the sampling part is placed under the tongue for 3 minutes, saliva infiltrates along the filter paper under the capillary action and dissolves the fluorescein sodium for color development, and the flow rate of the unstimulated saliva can be obtained through conversion by reading the infiltration length. The detection strip can effectively distinguish a control person from a sicca syndrome patient, a xerostomia patient and the like. The problems that a traditional method is long in consumed time, poor in experience and inconvenient to operate are solved, and a detection tool which is rapid, objective and standardized and is in seamless butt joint with an existing diagnostic standard is provided.

The method mainly comprises three core stages from S1 to S3. S1, a data set construction step; s2, a cold start supervision fine tuning step; s3, a progressive course reinforcement learning step; according to the method, the data is obtained from the multi-source medical field text data and is subjected to structured processing by using the knowledge distillation model, so that the private ophthalmology reasoning data set containing ophthalmology professional knowledge and reasoning logic is generated; the constructed ophthalmology model can accurately master professional contents such as complex anatomical structures, pathophysiological mechanisms, diagnostic standards and therapeutic schedules in the ophthalmology field, and absorption of the knowledge is further enhanced in the cold start supervision fine tuning stage, so that the model has a solid ophthalmology knowledge basis. Compared with a model pre-trained by directly using general corpora, the precision and timeliness of the model during ophthalmology professional problem processing are greatly improved through the targeted data set construction mode, understanding deviation and misleading information generation are reduced, and therefore the rigorous requirements of clinical application are met.

This invention discloses a model for predicting the risk of severe SA-AKI in ICU sepsis patients based on clinical variables. This study is based on a recent prospective, multicenter, large-sample cohort study conducted in China. Using the latest ADQI definition as the diagnostic criteria for severe SA-AKI, 1715 ICU sepsis patients were included. Clinical variables were collected within 24 hours after diagnosis of ICU sepsis patients, and six feature variables with the strongest association with severe SA-AKI were screened. Six machine learning algorithms were used to construct an early prediction model, among which SVM and Logistic Regression algorithms showed better predictive performance. The SHAP method was introduced to generate feature importance maps and individual prediction interpretation maps, which can intuitively show the contribution of each clinical variable to the risk prediction of a specific patient and explain the reasons for the occurrence of severe SA-AKI, so as to help doctors take targeted preventive measures.