A medical big data intelligent analysis method and system based on deep learning

Abstract

The application relates to the technical field of medical big data intelligent analysis, and discloses a medical big data intelligent analysis method and system based on deep learning, which comprises the following steps: acquiring patient image data, case text data and wearable device monitoring data, carrying out pretreatment, and standardizing and encoding various modal data. The dynamic weight of each modal data is calculated, and a multi-level fusion weight is synthesized. A modal value-clinical emergency degree double-factor resource decision model is constructed by combining the fusion weight and the patient clinical state parameter, and a priority score is obtained. According to the priority score, the patient is divided into different emergency treatment grades, and different treatment strategies are proposed. The model can maintain high prediction ability in various clinical scenes, improves the accuracy of clinical decision, and makes the resource allocation of the hospital more scientific and reasonable. In the emergency environment, real-time priority assessment of the patient can significantly improve the efficiency of emergency treatment, reduce the waiting time, and improve the survival rate and treatment effect of the patient.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent analysis of medical big data, and specifically to a method and system for intelligent analysis of medical big data based on deep learning. Background Art

[0002] In recent years, with the widespread adoption of medical imaging, electronic medical record (EMR) systems, and wearable devices, the application of medical big data has seen significant growth. The healthcare industry is gradually moving towards an era of data-driven, intelligent diagnosis and treatment. Multimodal data sources such as imaging data, case text data, and wearable device monitoring data provide clinicians with a richer array of patient information. The introduction of deep learning technology, in particular, has significantly advanced medical data analysis. Deep learning can automatically extract effective features from large-scale medical data, enabling efficient prediction and diagnosis, and has demonstrated significant advantages in areas such as disease prediction, image processing, and natural language processing. Furthermore, with the continuous advancement of cloud computing and big data technologies, the storage, processing, and analysis of medical data are no longer bottlenecks. The deployment and inference capabilities of deep learning models are also maturing, significantly promoting the development of intelligent medical big data analysis systems.

[0003] Despite significant progress in intelligent analysis methods for medical big data, existing technologies still face significant shortcomings in multimodal data fusion, personalized diagnosis and treatment, and clinical urgency assessment. Most existing methods for processing medical data rely on single-modal data, such as diagnosing diseases based solely on imaging data or inferring symptoms based solely on case text, lacking comprehensive analysis of multimodal data. This single data source limits diagnostic accuracy, especially in complex disease diagnoses where single-modal information cannot fully reflect a patient's health status. Furthermore, while deep learning technology has found numerous applications in medical image analysis and text processing, effectively integrating multiple sources of information—such as images, text, and wearable device monitoring data—and generating comprehensive diagnostic conclusions through deep learning models remains an unresolved challenge. Existing fusion methods often lack effective weighting mechanisms, resulting in poor multimodal data fusion performance and even information loss or data conflicts, compromising model accuracy and reliability.

[0004] Furthermore, despite the excellent performance of deep learning models in medical diagnosis, existing models often struggle to accurately assess individual patient conditions due to the diversity and complexity of training data. In clinical decision-making processes such as emergency care, personalized treatment recommendations, and patient prioritization, existing deep learning models lack sufficient flexibility and accuracy to dynamically optimize based on real-time patient data and changes in their condition. This prevents existing systems from providing comprehensive, real-time, and personalized medical support, and they also inadequately assess clinical urgency, leading to uneven resource allocation and delays in optimal treatment opportunities for patients.

[0005] Our invention incorporates deep learning technology to efficiently fuse and analyze multimodal data in real time. Using Bayesian optimization and weighted least squares methods for dynamic weight adjustment, it enables precise patient priority assessment. Compared to existing technologies, our invention not only improves the fusion of multimodal data but also dynamically adjusts treatment strategies based on patient condition changes in real time, significantly improving the accuracy and timeliness of clinical decision-making. Summary of the Invention

[0006] In view of the above-mentioned problems, the present invention is proposed.

[0007] Therefore, the technical problems solved by the present invention are: the existing medical data analysis methods have poor multimodal data fusion effects, lack of accuracy in personalized diagnosis and treatment and urgency assessment, and the inability to dynamically adjust resource allocation and treatment strategies according to patients' real-time data; as well as how to improve the accuracy of multimodal data fusion, optimize treatment plans in real time, and perform accurate clinical priority assessments.

[0008] To solve the above technical problems, the present invention provides the following technical solutions: a medical big data intelligent analysis method based on deep learning, comprising: obtaining patient imaging data, case text data and wearable device monitoring data, performing preprocessing, and standardizing the encoding of each modality data.

[0009] Calculate the dynamic weight of each modality data and synthesize the multi-level fusion weight.

[0010] Combining the fusion weights and the patient's clinical status parameters, a modality value-clinical urgency dual-factor resource decision-making model was constructed to obtain the priority score.

[0011] According to the priority score, patients are divided into different emergency treatment levels and different treatment strategies are proposed.

[0012] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the preprocessing includes collecting multimodal data of patients from hospital imaging systems, electronic medical records, and wearable devices, including imaging data, case text data, and wearable device data.

[0013] De-noise and normalize imaging data. Segment and vectorize case text data, using a word embedding model to convert text data into vector representations. Detect and correct outliers on wearable device data, and perform normalization.

[0014] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the standardized encoding of each modality data includes using a deep convolutional network to extract features from the image data and obtaining high-level feature vectors of the image through a pre-trained model.

[0015] The text data is encoded through the natural language processing model to obtain a vector representation of the text.

[0016] The monitoring data of wearable devices are modeled into time series through deep neural networks to obtain the time series feature vectors of the device monitoring data.

[0017] The obtained vector representations of each modal data are integrated to form a unified multimodal feature vector, which is expressed as:

[0018] ;

[0019] in, represents the final multimodal feature vector, Represents the encoding characteristics of image data, Represents the encoding characteristics of text data, Indicates the data encoding characteristics of wearable devices.

[0020] As a preferred solution of the medical big data intelligent analysis method based on deep learning described in the present invention, wherein: the calculation of the dynamic weight of each modality data includes initializing the weight of each modality , and iteratively update it using the Bayesian optimization algorithm, which is expressed as:

[0021] ;

[0022] in, Indicates the Mode in time The dynamic weight of represents the final multimodal feature vector, Indicates the The initial parameters of the modes, Indicates the current time.

[0023] In order to further improve the accuracy of the dynamic weight, the weighted least squares method is introduced to fine-tune the dynamic weight according to the quality of each modal data. The optimized weight is expressed as:

[0024] ;

[0025] in, represents the number of modes, Indicates the The standard deviation of a mode indicates the relative contribution of the mode to the final decision. Indicates the Patients at time The multimodal feature vector of Indicates the prediction model for the Mode in time The estimated feature vector of .

[0026] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the synthesized multi-level fusion weight includes normalizing the dynamic weights of each modality after obtaining them, and using a weighted average method to fuse the features of different modalities, which can be expressed as:

[0027] ;

[0028] in, Indicates the The normalized weights of the modes.

[0029] Based on the normalized weights, the features of each modality are weighted and fused to obtain the fused multimodal feature vector. The process of multimodal fusion is expressed as:

[0030] ;

[0031] in, represents the fused multimodal feature vector, Indicates the Mode in time The original vector.

[0032] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the patient's clinical status parameters are screened from the collected patient imaging data, case text data and wearable device monitoring data, including vital signs data, disease scores and clinical observation data.

[0033] Vital sign data include heart rate, respiratory rate, blood oxygen saturation and blood pressure, the condition score is the acute condition score, and clinical observation data include body temperature, weight and clinical symptom description.

[0034] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the construction of the modal value-clinical urgency dual-factor resource decision model includes merging the fused multimodal feature vector and the patient's clinical status parameters into a unified input vector. Assuming that there are m clinical status parameters, then Will contain fusion weights and patient clinical status parameters The splicing result is expressed as:

[0035] ;

[0036] A multilayer perceptron network model is used to learn the relationship between the fused features and clinical status parameters, including input layer, hidden layer and output layer.

[0037] The input layer receives input consisting of multimodal features and clinical status , the input vector The dimension is , which is the sum of the dimensions of the fusion features and the dimensions of the clinical status parameters.

[0038] The hidden layer contains multiple neurons, each of which receives weighted input from the input layer and is processed by a nonlinear activation function. hidden layer neurons The output is represented as:

[0039] ;

[0040] in, express activation function, Represents the connection input layer nodes and hidden layers The weight of the node, Represents the input layer The value of the element, Represents the hidden layer The bias term of each node.

[0041] The output layer uses the weighted sum of the outputs of the hidden layer and then maps the network output to the probability distribution of the three priority categories through the Softmax function, which is expressed as:

[0042] ;

[0043] in, Represents the connection of hidden layer The weights of the nodes and output layer, Represents the hidden layer The output of a neuron, represents the bias term of the output layer, represents the hidden layer dimension, express Function, expressed as:

[0044] ;

[0045] in, represents the weighted sum of each output neuron, Corresponding to low, medium and high priority, Represents the exponential sum of all categories, used to normalize the output.

[0046] Obtain the probability values ​​of the patient's three categories at time t: low priority, medium priority, and high priority, and select the category with the largest probability value as the patient's priority category.

[0047] As a preferred solution of the deep learning-based intelligent analysis method for medical big data described in the present invention, the construction of the modality value-clinical urgency dual-factor resource decision model also includes optimizing the network using a multi-class cross entropy loss function and training the deep learning model by minimizing the loss. The objective function is expressed as:

[0048] ;

[0049] in, Indicates patient In time The true category label of the patient Belong to category ,but , otherwise 0. Indicates patient In time Corresponding category The predicted probability of . represents the number of samples, Corresponding to low, medium, and high priority categories.

[0050] As a preferred embodiment of the deep learning-based intelligent analysis method for medical big data described in the present invention, patients are categorized into different emergency treatment levels based on priority scores, and different treatment strategies are proposed, including: for patients with the highest probability of belonging to the high-priority category, the hospital must immediately provide emergency treatment, and all relevant medical staff must promptly attend to the patient. The patient is assigned to a dedicated intensive care unit, and all relevant medical resources are immediately activated and provided with priority.

[0051] Patients will be monitored in real time 24 hours a day during the emergency process, and medical staff will check key physiological parameters every hour to ensure a rapid response to any changes in their condition.

[0052] If necessary, emergency treatment and surgery will be carried out immediately to ensure that the patient's condition is stabilized as soon as possible.

[0053] For patients whose probability values ​​belong to the medium priority category and whose conditions need to be treated in the short term, the hospital will arrange diagnosis and treatment within 48 hours and ensure that the patients are admitted to the ward as soon as possible.

[0054] Assess the patient's condition every 24 hours to check for worsening conditions and other emergencies. Transfer patients whose conditions suddenly worsen to high-priority treatment procedures.

[0055] Based on the results of clinical evaluation, patients are provided with necessary drug treatment to ensure that their condition remains stable.

[0056] For patients whose probability values ​​belong to the low priority category, the hospital will arrange for them to be diagnosed and treated through regular outpatient clinics.

[0057] For low-priority patients, the hospital will prioritize the allocation of routine resources, including general wards and routine examination equipment.

[0058] Patients will be scheduled for follow-up visits within an appropriate timeframe to ensure their condition has not deteriorated further. If any changes occur, their priority will be reassessed based on the latest assessment of their condition. We prioritize the patient's quality of life and provide symptom relief and rehabilitation based on their condition.

[0059] A medical big data intelligent analysis system based on deep learning, characterized by: including,

[0060] The preprocessing module obtains patient imaging data, case text data and wearable device monitoring data, performs preprocessing, and standardizes the encoding of each modality data.

[0061] The calculation module calculates the dynamic weights of each modality data and synthesizes multi-level fusion weights. Combining the fusion weights with the patient's clinical status parameters, a dual-factor resource decision model based on modality value and clinical urgency is constructed to obtain a priority score.

[0062] The classification module divides patients into different emergency treatment levels according to priority scores and proposes different treatment strategies.

[0063] A computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0064] A computer-readable storage medium stores a computer program, which implements the steps of the method described above when executed by a processor.

[0065] The beneficial effects of the present invention are as follows: by collecting, preprocessing and standardizing imaging data, case text data and wearable device monitoring data, combined with deep learning technology, feature extraction and fusion of data from different modalities are performed. The implementation of this step enables the system to integrate data from different sources, improving the comprehensiveness and accuracy of data analysis. Compared with traditional single data analysis methods, multimodal data fusion can overcome the limitations of each single data source and improve the predictive accuracy and reliability of the analysis model. This method reduces the probability of misdiagnosis and missed diagnosis by integrating information, effectively supporting more accurate personalized diagnosis and treatment.

[0066] During the model optimization process, a Bayesian optimization algorithm was introduced to tune the model's hyperparameters, and weighted least squares was combined to adjust the model's weighted data. Through this dynamic optimization process, the system can more accurately address data heterogeneity and improve prediction results. The introduction of this optimization step not only enables efficient training of the model with less sample data, but also allows dynamic adjustment based on individual patient differences, enabling the model to maintain efficient predictive capabilities in a variety of clinical scenarios and improving the accuracy of clinical decision-making.

[0067] The fused data is processed using a multi-layer perceptron (MLP) network model, combining the patient's health status with clinical urgency parameters (such as vital signs and condition scores) to generate a priority score for each patient. This score is categorized as high priority, medium priority, and low priority based on the output of the Softmax function, ensuring that the urgency level of each patient can be accurately determined. Unlike traditional deep learning-based priority assessment methods, this invention dynamically assesses the patient's urgency through a deep learning model, avoiding the subjectivity and errors of human judgment and making hospital resource allocation more scientific and reasonable. In emergency settings, in particular, real-time patient priority assessment can significantly improve the efficiency of emergency treatment, reduce waiting times, and enhance patient survival rates and treatment outcomes. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0069] Figure 1 The first embodiment of the present invention provides an overall flow chart of a medical big data intelligent analysis method and system based on deep learning. DETAILED DESCRIPTION

[0070] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that 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 persons in this field without creative work should fall within the scope of protection of the present invention.

[0071] Example 1, reference Figure 1 , as one embodiment of the present invention, provides a medical big data intelligent analysis method based on deep learning, comprising:

[0072] S1: Obtain patient imaging data, case text data, and wearable device monitoring data, perform preprocessing, and standardize the encoding of each modality data.

[0073] Collect patients' multimodal data from hospital imaging systems, electronic medical records, and wearable devices, including imaging data, case text data, and wearable device data.

[0074] It should be noted that medical imaging equipment uses CT scanners, MRI (magnetic resonance imaging) equipment, X-ray equipment, and ultrasound scanners. The acquisition process is as follows:

[0075] In the hospital's imaging department, radiologists select appropriate imaging equipment for examination based on the patient's symptoms and clinical needs. The scan results generated by the imaging equipment are automatically recorded in the hospital's imaging system, and all imaging data (stored in DICOM format and transmitted to the hospital's data management system) are automatically recorded in the hospital's imaging system.

[0076] The process of collecting case text data is as follows:

[0077] When a patient visits the hospital, the doctor will record the patient's medical history, chief complaints, physical signs, diagnosis, and treatment plan, and this information will be entered into the hospital's electronic medical record system.

[0078] The electronic medical record system automatically generates and archives the patient's medical record files. These text data contain detailed medical information, diagnosis descriptions, treatment records, etc.

[0079] Wearable device data collection devices include smart bracelets, smart watches, and wearable physiological monitoring devices (such as heart rate monitors, blood oximeters, sleep monitoring devices, etc.). The collection process is as follows:

[0080] After the patient wears the wearable device, the device will monitor and record physiological data (such as heart rate, body temperature, respiratory rate, blood oxygen saturation, etc.) through sensors.

[0081] The data is transmitted to the hospital data platform and synchronized to the hospital data management system in real time.

[0082] De-noise and normalize imaging data. Segment and vectorize case text data, using a word embedding model to convert text data into vector representations. Detect and correct outliers on wearable device data, and perform normalization.

[0083] It should be noted that image denoising is performed using the U-Net model within a convolutional neural network (CNN). U-Net is widely used for denoising and segmentation tasks in medical imaging and has powerful feature extraction capabilities. Normalization normalizes the image data pixel values ​​to the [0, 1] range, facilitating subsequent model training.

[0084] Furthermore, case text data processing includes the following:

[0085] Word segmentation: spaCy (for English text word segmentation) and Jieba (for Chinese text word segmentation) are used to segment the case text data. The segmented data can be more easily processed by deep learning models.

[0086] Word Embedding Model: Use the BERT model to vectorize text data. The BERT model is particularly suitable for processing medical case texts and can obtain contextual semantic information for each word.

[0087] Wearable device data processing includes the following:

[0088] Outlier detection and correction: Use the isolation forest algorithm for outlier detection. These algorithms can effectively identify abnormal fluctuations in monitoring data and correct data points that do not conform to the regular pattern.

[0089] Standardization: Use Z-score standardization to standardize the device data so that they have the same scale, which is convenient for subsequent model processing.

[0090] Use deep convolutional networks to extract features from image data and obtain high-level feature vectors of images through pre-training models.

[0091] The text data is encoded through the natural language processing model to obtain a vector representation of the text.

[0092] The monitoring data of wearable devices are modeled into time series through deep neural networks to obtain the time series feature vectors of the device monitoring data.

[0093] It should be noted that image feature extraction uses ResNet50, a deep convolutional neural network (CNN). A pre-trained model is fine-tuned to extract high-level feature vectors from the images. Text data feature extraction uses the BERT model to encode case text data and convert it into a fixed-dimensional text vector representation. Wearable device data feature extraction: LSTM is used for time series modeling to extract time series feature vectors from device monitoring data.

[0094] The obtained vector representations of each modal data are integrated to form a unified multimodal feature vector, which is expressed as:

[0095] ;

[0096] in, represents the final multimodal feature vector, Represents the encoding characteristics of image data, Represents the encoding characteristics of text data, Indicates the data encoding characteristics of wearable devices.

[0097] S2: Calculate the dynamic weights of each modality data and synthesize multi-level fusion weights.

[0098] Initialize the weights of each modality , and iteratively update it using the Bayesian optimization algorithm, which is expressed as:

[0099] ;

[0100] in, Indicates the Mode in time The dynamic weight of represents the final multimodal feature vector, Indicates the The initial parameters of the modes, Indicates the current time.

[0101] In order to further improve the accuracy of the dynamic weight, the weighted least squares method is introduced to fine-tune the dynamic weight according to the quality of each modal data. The optimized weight is expressed as:

[0102] ;

[0103] Where n represents the number of modes, Indicates the The standard deviation of a mode indicates the relative contribution of the mode to the final decision. Indicates the Patients at time The multimodal feature vector of Indicates the prediction model for the Mode in time The estimated feature vector of .

[0104] It should be noted that Bayesian optimization is a global optimization algorithm that uses surrogate models (such as Gaussian processes) to optimize the objective function. It can find the optimal hyperparameter configuration with a small number of evaluations, making it suitable for deep learning models that require hyperparameter tuning. Bayesian optimization allows for dynamic adjustment of model parameters during training, thereby improving model performance.

[0105] Furthermore, the introduction of weighted least squares optimizes by assigning weights to each data point. This is very effective for data imbalance. For medical data, certain patients' conditions may be more important. Giving these samples higher weights helps the model better learn their characteristics. Using weighted least squares can improve the model's prediction accuracy for important categories, which is particularly important in the medical field for detecting and classifying key conditions.

[0106] After obtaining the dynamic weights of each modality, normalization is performed and the weighted average method is used to fuse the features of different modalities, which can be expressed as:

[0107] ;

[0108] in, Indicates the The normalized weights of the modes.

[0109] Based on the normalized weights, the features of each modality are weighted and fused to obtain the fused multimodal feature vector. The process of multimodal fusion is expressed as:

[0110] ;

[0111] in, represents the fused multimodal feature vector, Indicates the Mode in time The original vector.

[0112] S3: Combining the fusion weights and the patient's clinical status parameters, a modality value-clinical urgency dual-factor resource decision-making model is constructed to obtain a priority score.

[0113] Patient clinical status parameters are screened from collected patient imaging data, case text data, and wearable device monitoring data, including vital signs data, disease scores, and clinical observation data.

[0114] Vital sign data include heart rate, respiratory rate, blood oxygen saturation and blood pressure, the condition score is the acute condition score, and clinical observation data include body temperature, weight and clinical symptom description.

[0115] Furthermore, the acute condition score is assigned by the doctor based on the patient's condition, combining multiple dimensions such as clinical diagnosis, pathological indicators, and vital signs to assess the severity of the patient's condition. This score is manually assigned by the doctor using a scoring table based on clinical experience and the patient's specific condition.

[0116] The fused multimodal feature vector and the patient's clinical status parameters are combined into a unified input vector. Assuming there are m clinical status parameters, then Will contain fusion weights and patient clinical status parameters The splicing result is expressed as:

[0117] ;

[0118] A multilayer perceptron network model is used to learn the relationship between the fused features and clinical status parameters, including input layer, hidden layer and output layer.

[0119] The input layer receives input consisting of multimodal features and clinical status , the input vector The dimension is , which is the sum of the dimensions of the fusion features and the dimensions of the clinical status parameters.

[0120] The hidden layer contains multiple neurons, each of which receives weighted input from the input layer and is processed by a nonlinear activation function. hidden layer neurons The output is represented as:

[0121] ;

[0122] in, express activation function, Represents the connection input layer nodes and hidden layers The weight of the node, Represents the input layer The value of the element, Represents the hidden layer The bias term of each node.

[0123] The output layer uses the weighted sum of the outputs of the hidden layer and then maps the network output to the probability distribution of the three priority categories through the Softmax function, which is expressed as:

[0124] ;

[0125] in, Represents the connection of hidden layer The weights of the nodes and output layer, Represents the hidden layer The output of a neuron, represents the bias term of the output layer, represents the hidden layer dimension, express Function, expressed as:

[0126] ;

[0127] in, represents the weighted sum of each output neuron, Corresponding to low, medium and high priority, Represents the exponential sum of all categories, used to normalize the output.

[0128] Obtain the probability values ​​of the patient's three categories at time t: low priority, medium priority, and high priority, and select the category with the largest probability value as the patient's priority category.

[0129] The network is optimized using a multi-class cross entropy loss function. The deep learning model is trained by minimizing the loss. The objective function is expressed as:

[0130] ;

[0131] in, Indicates patient In time The true category label of the patient Belong to category ,but , otherwise 0. Indicates patient In time Corresponding category The predicted probability of . N represents the number of samples, Corresponding to low, medium, and high priority categories.

[0132] It should be noted that by minimizing the multi-class cross-entropy loss function, the model's classification performance can be optimized, minimizing the gap between the probability distribution of the model output and the actual label. Optimizing the loss function can improve the model's ability to identify different priority categories, ensuring that the model can accurately assign priorities to patients and optimize the allocation of medical resources.

[0133] S4: Based on the priority score, patients are divided into different emergency treatment levels and different treatment strategies are proposed.

[0134] For patients with the highest probability, those in the high-priority category must be treated immediately by the hospital. All relevant medical staff must attend to the patient as soon as possible. The patient will be assigned to a dedicated intensive care unit, and all relevant medical resources will be immediately activated and provided with priority.

[0135] Patients will be monitored in real time 24 hours a day during the emergency process, and medical staff will check key physiological parameters every hour to ensure a rapid response to any changes in their condition.

[0136] If necessary, emergency treatment and surgery will be carried out immediately to ensure that the patient's condition is stabilized as soon as possible.

[0137] For patients whose probability values ​​belong to the medium priority category and whose conditions need to be treated in the short term, the hospital will arrange diagnosis and treatment within 48 hours and ensure that the patients are admitted to the ward as soon as possible.

[0138] Assess the patient's condition every 24 hours to check for worsening conditions and other emergencies. Transfer patients whose conditions suddenly worsen to high-priority treatment procedures.

[0139] Based on the results of clinical evaluation, patients are provided with necessary drug treatment to ensure that their condition remains stable.

[0140] For patients whose probability values ​​belong to the low priority category, the hospital will arrange for them to be diagnosed and treated through regular outpatient clinics.

[0141] For low-priority patients, the hospital will prioritize the allocation of routine resources, including general wards and routine examination equipment.

[0142] Patients will be scheduled for follow-up visits within an appropriate timeframe to ensure their condition has not deteriorated further. If any changes occur, their priority will be reassessed based on the latest assessment of their condition. We prioritize the patient's quality of life and provide symptom relief and rehabilitation based on their condition.

[0143] The above embodiments also include a medical big data intelligent analysis system based on deep learning, specifically:

[0144] The preprocessing module obtains patient imaging data, case text data and wearable device monitoring data, performs preprocessing, and standardizes the encoding of each modality data.

[0145] The calculation module calculates the dynamic weights of each modality data and synthesizes multi-level fusion weights. Combining the fusion weights with the patient's clinical status parameters, a dual-factor resource decision model based on modality value and clinical urgency is constructed to obtain a priority score.

[0146] The classification module divides patients into different emergency treatment levels according to priority scores and proposes different treatment strategies.

[0147] The computer device may be a server. The computer device includes a processor, a memory, an input / output (I / O) interface, and a communication interface. The processor, memory, and I / O interface are connected via a system bus, and the communication interface is connected to the system bus via the I / O interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The database of the computer device is used to store data cluster data of the power monitoring system. The I / O interface of the computer device is used to exchange information between the processor and external devices. The communication interface of the computer device is used to communicate with external terminals via a network connection. When the computer program is executed by the processor, it implements a medical big data intelligent analysis method based on deep learning.

[0148] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the above-mentioned embodiments. In particular, any reference to memory, database, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may be, but are not limited to, general-purpose processors, central processing units (CPUs), graphics processors (GPUs), digital signal processors (DSPs), programmable logic devices (PLDs), data processing logic devices based on quantum computing, and the like.

[0149] Example 2 is an embodiment of the present invention, which provides a medical big data intelligent analysis method and system based on deep learning. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through simulation experiments.

[0150] The goal of this example is to verify the effectiveness of a deep learning-based intelligent analysis method and system for medical big data, particularly its application in multimodal data fusion, clinical priority assessment, and resource allocation. The primary subject of the experiment is 100 patient data from the emergency department of a large hospital, covering imaging data, case text data, and wearable device monitoring data. The specific implementation process is as follows:

[0151] Imaging Data: CT imaging data from 100 patients, ranging in condition from mild respiratory illness to acute myocardial infarction, were collected through the hospital's imaging department. The imaging data were stored in DICOM format and uploaded to the hospital's data management system.

[0152] Case text data: Patient medical records, including medical history, physical signs, and diagnosis information, are collected through the hospital's electronic medical record system (EMR). The medical record text is processed by a natural language processing (NLP) model to extract key information such as symptoms and past medical history.

[0153] Wearable device monitoring data: Smartwatches, smart heart rate monitors, and other wearable devices are used to collect patients' real-time vital signs, including heart rate, blood oxygen saturation, and body temperature. All wearable device data is uploaded to the cloud platform through the hospital's smart terminals and standardized.

[0154] Data preprocessing and feature extraction:

[0155] Image data preprocessing: Use the U-Net model to denoise the image data and normalize it to the [0, 1] range to facilitate input into the subsequent deep learning model. Use the ResNet50 model to extract high-level feature vectors from the image data.

[0156] Case text data processing: Case text data is converted into vector representations through word segmentation and word embedding models (BERT). The BERT model can capture the contextual semantic relationships in the text, thereby generating accurate text vector representations.

[0157] Wearable device data processing: Use the LSTM model to perform time series modeling on wearable device monitoring data and extract time series feature vectors. Data standardization ensures that data from different devices can be analyzed at the same scale.

[0158] Training and optimization of deep learning models:

[0159] A Bayesian optimization algorithm was used to optimize model hyperparameters, while weighted least squares was also used to adjust the model's weighting of each modality's data. Preprocessed image, text, and wearable device data were fused and input into a deep learning model. This model, based on a multi-layer perceptron (MLP) network structure, outputs a priority score for each patient using a softmax function.

[0160] The model is trained using a multi-class cross entropy loss function to ensure that the final model achieves optimal performance on different priority classifications.

[0161] During the training process, the model parameters are optimized by comparing the model prediction results with the priority labels assessed by experts, thereby improving the accuracy and stability of the model in clinical applications.

[0162] Evaluation criteria:

[0163] Accuracy: Use the test set to evaluate the accuracy of the model in classifying patient priorities.

[0164] Resource allocation efficiency: Patients are assigned to different treatment priorities based on the priority scores predicted by the model to evaluate the effectiveness of the model in actual clinical resource allocation.

[0165] Patient survival rate and treatment time: The effectiveness of this deep learning model was verified by comparing the survival rate and treatment time of emergency patients after priority was assigned using this model.

[0166] The experimental results are shown in Table 1.

[0167] Table 1 Experimental data of the model of the present invention in the priority assessment of emergency patients

[0168]

[0169] As can be seen from the experimental data table, the deep learning-based intelligent analysis method for medical big data has significant advantages in assessing the priority of emergency patients. In the high-priority patient prediction probability column, the prediction probability processed by the deep learning model is significantly higher than that of traditional manual evaluation methods and is highly consistent with the priority labels annotated by clinical experts. For example, the high priority prediction probabilities for patients 1, 3, and 5 are 75.2%, 80.1%, and 85.5%, respectively. These data show that the model can maintain a high level of accuracy in predicting high-priority patients. Moreover, the priority scores predicted by this method can effectively guide the allocation of hospital resources, ensuring that high-priority patients in the emergency room receive timely treatment.

[0170] Furthermore, a comparison of the resource allocation efficiency (treatment time, minutes) data reveals that this method significantly improves resource allocation efficiency. For example, patient 1's treatment took 30 minutes, while traditional methods typically require much longer to make a similar diagnosis and allocation. This demonstrates that applying deep learning models to emergency patient priority assessment can significantly reduce patient wait times and improve the efficiency of hospital emergency resources.

[0171] As can be seen from the processing times for both case text data and wearable device data, the entire data processing process was completed within a reasonable timeframe, ensuring real-time data availability and the model's rapid responsiveness. In particular, the use of deep neural networks for time series modeling in the processing of wearable device data effectively captured the patient's dynamic changes, further improving prediction accuracy.

[0172] The above data analysis demonstrates that the deep learning model presented in this paper can accurately and rapidly assess the priority of emergency patients, optimize resource allocation processes, and improve the efficiency and effectiveness of hospital emergency treatment. Compared to existing technologies, this paper demonstrates significant innovation and novelty in multimodal data fusion and priority assessment, enabling more intelligent and accurate decision-making support in emergency medical settings, with significant clinical application prospects.

[0173] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A medical big data intelligent analysis method based on deep learning, characterized in that: include: Obtain patient imaging data, case text data, and wearable device monitoring data, perform preprocessing, and standardize the coding of each modality data; Calculate the dynamic weight of each modality data and synthesize multi-level fusion weights; Combining the fusion weights and the patient's clinical status parameters, a dual-factor resource decision-making model of modality value and clinical urgency was constructed to obtain a priority score. According to the priority score, patients are divided into different emergency treatment levels and different treatment strategies are proposed; The calculation of the dynamic weight of each modal data includes initializing the weight of each modal , and iteratively update it using the Bayesian optimization algorithm, which is expressed as: in, represents the dynamic weight of the i-th mode at time t, represents the final multimodal feature vector, represents the initial parameters of the i-th mode, Indicates the current time; In order to further improve the accuracy of the dynamic weight, the weighted least squares method is introduced to fine-tune the dynamic weight according to the quality of each modal data. The optimized weight is expressed as: Where n represents the number of modes, represents the standard deviation of the i-th mode, indicating the relative contribution of this mode to the final decision; represents the multimodal feature vector of the jth patient at time t, Represents the estimated feature vector of the i-th mode at time t based on the prediction model; The synthesis of multi-level fusion weights includes normalizing the dynamic weights of each modality and fusing the features of different modalities using a weighted average method, which is expressed as: in, represents the normalized weight of the i-th mode; Based on the normalized weights, the features of each modality are weighted and fused to obtain the fused multimodal feature vector. The process of multimodal fusion is expressed as: in, represents the fused multimodal feature vector, Represents the original vector of the i-th mode at time t.

2. The deep learning-based intelligent analysis method for medical big data according to claim 1, characterized in that: The preprocessing includes collecting multimodal data of patients from hospital imaging systems, electronic medical records, and wearable devices, including imaging data, case text data, and wearable device data; The imaging data is denoised and normalized; the case text data is segmented and vectorized, and the word embedding model is used to convert the text data into a vector representation; the wearable device data is detected and corrected for outliers and standardized.

3. The deep learning-based intelligent analysis method for medical big data according to claim 2, characterized in that: The standardized encoding of each modality data includes extracting features from the image data using a deep convolutional network and obtaining a high-level feature vector of the image through a pre-trained model; Encode text data through a natural language processing model to obtain a vector representation of the text; The monitoring data of the wearable device is modeled into a time series using a deep neural network to obtain the time series feature vector of the device monitoring data; The obtained vector representations of each modal data are integrated to form a unified multimodal feature vector, which is expressed as: in, represents the final multimodal feature vector, Represents the encoding characteristics of image data, Represents the encoding characteristics of text data, Indicates the data encoding characteristics of wearable devices.

4. The deep learning-based intelligent analysis method for medical big data according to claim 3, characterized in that: The patient's clinical status parameters are screened from the collected patient imaging data, case text data, and wearable device monitoring data, including vital signs data, disease score, and clinical observation data; Vital sign data include heart rate, respiratory rate, blood oxygen saturation and blood pressure, the condition score is the acute condition score, and clinical observation data include body temperature, weight and clinical symptom description.

5. The deep learning-based intelligent analysis method for medical big data according to claim 4, characterized in that: The construction of the modality value-clinical urgency dual-factor resource decision model includes combining the fused multimodal feature vector and the patient's clinical status parameters into a unified input vector. Assuming there are m clinical status parameters, then Will contain fusion weights and patient clinical status parameters The splicing result is expressed as: A multilayer perceptron network model is used to learn the relationship between the fused features and clinical status parameters, including input layer, hidden layer and output layer; The input layer receives input consisting of multimodal features and clinical status , the input vector The dimension is , is the sum of the dimensions of the fusion features and the dimensions of the clinical status parameters; The hidden layer contains multiple neurons. Each neuron receives the weighted sum input from the input layer and processes it through a nonlinear activation function. The jth hidden layer neuron The output is represented as: in, represents the ReLU activation function, represents the weight connecting the i-th node in the input layer and the j-th node in the hidden layer, Represents the value of the i-th element in the input layer, represents the bias term of the jth node in the hidden layer; The output layer uses the weighted sum of the outputs of the hidden layer and then maps the network output to the probability distribution of the three priority categories through the Softmax function, which is expressed as: in, represents the weight connecting the jth node of the hidden layer and the output layer, represents the output of the jth neuron in the hidden layer, represents the bias term of the output layer, represents the hidden layer dimension, Represents the softmax function, expressed as: in, Represents the weighted sum of each output neuron, i=1,2,3 corresponds to low, medium, and high priority, Represents the sum of the exponentials of all categories, used to normalize the output; Obtain the probability values ​​of the patient's three categories at time t: low priority, medium priority, and high priority, and select the category with the largest probability value as the patient's priority category.

6. The deep learning-based intelligent analysis method for medical big data according to claim 5, characterized in that: The constructing of the modality value-clinical urgency dual-factor resource decision model further includes optimizing the network using a multi-class cross entropy loss function and training the deep learning model by minimizing the loss. The objective function is expressed as: in, represents the true category label of patient i at time t. If patient i belongs to category c, then =1, otherwise 0; represents the predicted probability of patient i corresponding to category c at time t; N represents the number of samples, Corresponding to low, medium, and high priority categories.

7. The deep learning-based intelligent analysis method for medical big data according to claim 6, characterized in that: According to the priority score, patients are divided into different emergency treatment levels and different treatment strategies are proposed, including: for patients with the highest probability value belonging to the high priority category, the hospital must immediately provide emergency treatment, and all relevant medical staff must receive the patient as soon as possible; the patient is assigned to a dedicated intensive care unit, and all relevant medical resources are immediately activated and provided with priority; Patients will be monitored 24 hours a day in real time during the emergency room, with medical staff checking key physiological parameters every hour to ensure a rapid response to any changes in their condition; If necessary, emergency treatment and surgery will be carried out immediately to ensure that the patient's condition is stabilized as soon as possible; For patients with medium-priority probabilities whose conditions require short-term treatment, the hospital will arrange diagnosis and treatment within 48 hours and ensure that the patient is admitted to the ward as soon as possible. Assess the patient's condition every 24 hours to check for worsening conditions or other emergencies. If the patient's condition suddenly becomes serious, transfer them to a high-priority treatment process. Provide patients with necessary medication based on clinical assessment results to ensure their condition remains stable; For patients whose probability values ​​belong to the low priority category, the hospital will arrange for them to be diagnosed and treated through regular outpatient clinics; For low-priority patients, the hospital will prioritize the allocation of routine resources, including general wards and routine examination equipment; The patient should arrange a follow-up appointment within an appropriate time after the consultation to ensure that the condition has not deteriorated further. If there is a change, the patient's priority should be reconsidered based on the latest assessment of the condition; Focus on the patient's quality of life and provide symptom relief and rehabilitation treatment according to the patient's condition.

8. A deep learning-based medical big data intelligent analysis system using the method according to any one of claims 1 to 7, characterized in that: The preprocessing module obtains patient imaging data, case text data, and wearable device monitoring data, performs preprocessing, and standardizes the encoding of each modality data; The calculation module calculates the dynamic weight of each modal data and synthesizes the multi-level fusion weight; Combining the fusion weights and the patient's clinical status parameters, a dual-factor resource decision-making model of modality value and clinical urgency was constructed to obtain a priority score. The classification module divides patients into different emergency treatment levels according to priority scores and proposes different treatment strategies.

Images