Analysis method and system for analysis of and treatment plan for edema after hemorrhagic stroke

Abstract

Disclosed are an analysis method and a system for analysis of and a treatment plan for edema after a hemorrhagic stroke, comprising: collecting edema volume data of each patient at different time points and preprocessing the data; using a Gaussian fitting algorithm to perform fitting, constructing a curve of edema volume progression over time, and calculating a residue between the curve and a true value; dividing patients into multiple subgroups by means of a deep learning clustering algorithm, fitting the edema volume of each subgroup at different time points, and calculating a residue thereof with a true value; calculating a change in edema volume, determining a change trend, combining with a corresponding treatment plan, using an association rule algorithm to analyze a degree of correlation, and identifying a treatment plan having the highest degree of correlation with the edema progression trend. The present invention uses advanced algorithms to accurately analyze changes in edema volume after hemorrhagic stroke, improving the accuracy of predicting patient prognosis. By means of detailed analysis of different patient subgroups, it supports more personalized treatment plans and helps improve treatment effects.

Description

A method and system for analyzing edema after hemorrhagic stroke and analyzing treatment plans Technical Field

[0001] The present invention relates to the field of clinical intelligent diagnosis and treatment technology, and specifically to a method and system for analyzing edema after hemorrhagic stroke and analyzing treatment plans. Background Art

[0002] Hemorrhagic stroke, a severe cerebral hemorrhage caused by non-traumatic intraparenchymal vascular rupture, often leads to high mortality and severe neurological dysfunction due to its acute onset and rapid progression. This disease not only poses significant health risks to the patient but also places a heavy burden on society and the family economy. Particularly in the late stages of hemorrhagic stroke, hematoma expansion is considered a significant risk factor for poor prognosis, potentially increasing intracranial pressure, exacerbating neurological deterioration, and even endangering life.

[0003] Furthermore, perihematomal edema, a key marker of secondary injury after hemorrhagic stroke, has received increasing clinical attention in recent years. This edema may cause brain tissue compression, impair neuronal function, and lead to further brain tissue damage, worsening neurological dysfunction. Therefore, dynamic monitoring and early prediction of hematoma and perihematomal edema are crucial for improving patient survival and quality of life.

[0004] With the continuous advancement of medical imaging technology, noninvasive dynamic monitoring of brain tissue damage after hemorrhagic stroke has become possible. Simultaneously, the rapid development of artificial intelligence has revolutionized data processing and analysis in the medical field. By integrating multi-source data such as imaging features, clinical information, and treatment plans, it is possible to accurately predict the prognosis of patients after hemorrhagic stroke and evaluate personalized treatment efficacy.

[0005] The present invention proposes a comprehensive analysis method based on a Gaussian fitting algorithm, a deep learning clustering algorithm, and an association rule algorithm. This method aims to accurately predict and evaluate the changing trend of edema volume in patients after hemorrhagic stroke, as well as the correlation between these changes and different treatment plans. Through the Gaussian fitting algorithm, the present invention can construct a curve of edema volume changes over time and calculate the residual between the actual observation value and the fitting curve. Using the deep learning clustering algorithm, the present invention further distinguishes different patient subgroups and performs a special edema volume trend analysis on each subgroup. Finally, through the association rule algorithm, the present invention can identify the treatment plan with the highest correlation with the edema progression trend, providing important support for clinical treatment decision-making.

[0006] Summary of the Invention

[0007] In view of the above-mentioned problems, the present invention is proposed.

[0008] Therefore, the technical problem solved by the present invention is: how to effectively monitor and predict the development of hematoma expansion and surrounding edema after hemorrhagic stroke, and accurately evaluate the correlation between the treatment method and the patient's edema volume change trend, so as to formulate a personalized treatment plan.

[0009] To solve the above technical problems, the present invention provides the following technical solutions: a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans, comprising: collecting edema volume data of each patient at different time points and performing preprocessing;

[0010] The preprocessed data were fitted using a Gaussian fitting algorithm to construct a curve showing the progression of edema volume over time for all patients, and the residual between the curve and the true value was calculated;

[0011] Patients were divided into multiple subgroups using a deep learning clustering algorithm. Gaussian functions were fitted to the edema volume of each subgroup at different time points, and the residuals between the fitted subgroup curves and the true values ​​of each subgroup were calculated.

[0012] Calculate the change in edema volume, determine the change trend, combine it with the corresponding treatment plan, use the association rule algorithm to analyze the correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

[0013] As a preferred embodiment of the method for analyzing edema and treating post-hemorrhagic stroke according to the present invention, the preprocessing includes cleaning the edema volume data of each patient at different time points, correcting errors and outliers, performing average value filling and linear interpolation on missing edema volume data, time-aligning the edema volume data according to the patient's examination time and onset time, and then standardizing the edema volume data.

[0014] As a preferred embodiment of the method for analyzing edema and treating post-hemorrhagic stroke according to the present invention, the method of fitting the preprocessed data set using a Gaussian fitting algorithm includes sorting the preprocessed edema volume data by patient and time point to form a structured data set, wherein each row represents the edema volume of a patient at a specific time point;

[0015] The Gaussian function formula is selected as:

[0016] Where v represents the time from onset to examination, a, b, and c are the parameters of the Gaussian function curve of the patient's edema with respect to time after fitting, a represents the peak value of the curve, b represents the mean value of the curve, and c represents the width of the curve;

[0017] For each patient's data, the least squares method was used to estimate the parameters a, b, and c of the Gaussian model, and the gradient descent method was used for iterative parameter optimization;

[0018] A Gaussian function was applied to the time series data of each patient to generate a fitting curve of edema volume over time.

[0019] As a preferred embodiment of the method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to the present invention, the calculation of the residual between the curve and the true value includes defining the residual as the difference between the predicted value of the Gaussian fitting curve and the actual observed value, and the calculation formula is:

[0020] Among them, R i represents the fitting residual at time point i, Y i represents the true value of the edema volume at time point i, represents the predicted value of the Gaussian fitting curve at the same time point;

[0021] For each patient's edema volume data at each time point, the residual between the data and the corresponding point of the Gaussian fitting curve was calculated.

[0022] As a preferred embodiment of the method for analyzing edema after hemorrhagic stroke and analyzing treatment plans described in the present invention, the method comprises: dividing patients into multiple subgroups by a deep learning clustering algorithm, extracting features from patient data using an autoencoder based on the personal information characteristics of different patients, and converting the original various feature data x into a low-dimensional latent variable h through the encoder during the encoding process. The calculation formula is h=σ(W1m+b1), wherein W1 and b1 are the weight and bias of the encoder, respectively, σ is the activation function, and m is the eigenvalue under the feature encoder;

[0023] During the decoding process, the latent variable h after the encoding process is decoded back to the original data dimension to obtain the reconstructed data The calculation formula is:

[0024] Where W2 represents the weight of the feature decoder, and b2 represents the bias of the feature decoder;

[0025] The parameters of the autoencoder are repeatedly optimized through the loss function, which is defined as minimizing the difference between the original data x and the reconstructed data The difference between them is: MinimizeLoss = dist(G,G R )

[0026] Among them, G represents the original feature information dataset of each patient; G RRepresents the original feature information dataset of each patient after being reconstructed by the encoder;

[0027] The low-dimensional representation h obtained by training the autoencoder is used for K-Means clustering. According to the pre-set number of clusters k, the algorithm will divide the patients into k clusters. Each cluster consists of patients with similar characteristics, represented by (C1, C2, ... C k ), then the minimized square error E is expressed as:

[0028] Among them, ||.|| represents the norm of the vector, l represents the low-dimensional representation of patient information, μ i Represents cluster C i The mean vector of .

[0029] As a preferred embodiment of the method for analyzing edema after hemorrhagic stroke and analyzing treatment plans of the present invention, wherein: the residuals of the calculated fitting subclass curves and the true values ​​of the subclasses include:

[0030] Gaussian fitting was performed on the patient data in each subgroup divided by the deep learning clustering algorithm, and the residual between the Gaussian fitting curve of each subgroup and the actual edema volume data of each patient in the subgroup was calculated.

[0031] As a preferred embodiment of the method for analyzing edema after hemorrhagic stroke and analyzing treatment plans of the present invention, the method of analyzing the correlation between treatment methods and edema progression trends using an association rule algorithm includes extracting edema volume changes and corresponding treatment plan information from a preprocessed data set;

[0032] Define item sets X and Y, where X represents a specific treatment plan and Y represents a specific edema volume change trend;

[0033] To calculate support, the support of data item set X, support(X), is the ratio of the number of transactions containing X in D to the total number of transactions in D. The formula is:

[0034] Where D represents the development of edema, x represents the specific treatment plan selected, count(x) represents the number of specific treatment plans selected, and count(D) represents the number of cases where the edema development trend is inhibited;

[0035] The support of the association rule X=>Y is equal to the support of the item set X∪Y, which can be expressed as follows:

[0036] Where X represents a specific treatment option selected, Y represents another specific treatment option selected, and count(X∪Y) represents the number of treatment options selected;

[0037] To calculate the confidence level of X=>Y, use the formula:

[0038] A rule that evaluates the likelihood of another treatment being present given a particular treatment; where Support(X∩Y) represents the support for both treatments being present, and Support(X) represents the support for the presence of the first treatment.

[0039] The Apriori algorithm is used to iteratively generate frequent item sets and corresponding association rules, and the rules with the highest support and confidence are identified from the generated association rules.

[0040] A system for analyzing edema after hemorrhagic stroke and analyzing treatment plans, characterized by: including:

[0041] Data preprocessing module: preprocess the edema data of each patient at different time points;

[0042] Gaussian fitting module: uses Gaussian fitting algorithm to fit the above data set, constructs a curve of edema volume progression over time for all patients, and calculates the residual between the curve and the true value;

[0043] Patient subgroup classification module: Patients are divided into multiple subgroups using a deep learning clustering algorithm. Gaussian functions are fitted to the edema volume of patients in each subgroup at different time points, and the residuals between the fitted subgroup curves and the true values ​​of each subgroup are calculated.

[0044] Treatment plan association analysis module: Calculate the change in edema volume, determine the change trend, combine it with the corresponding treatment plan, use the association rule algorithm to analyze the degree of correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

[0045] 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.

[0046] A computer-readable storage medium stores a computer program, which implements the steps of the method described above when executed by a processor.

[0047] The beneficial effects of this invention include: Utilizing advanced algorithms to accurately analyze changes in edema volume after hemorrhagic stroke, the accuracy of patient prognosis prediction is improved. Through detailed analysis of different patient subgroups, this invention supports more personalized treatment plans, helping to improve treatment outcomes. Early identification and accurate prediction of hematoma and edema progression facilitates timely and effective measures, improving patient prognosis and quality of life. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] 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:

[0049] FIG1 is an overall flow chart of a method for analyzing edema after hemorrhagic stroke and analyzing a treatment plan provided by a first embodiment of the present invention;

[0050] FIG2 is a flowchart of an association rule Apriori algorithm for a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by the first embodiment of the present invention;

[0051] FIG3 is a fitting diagram of edema volume vs. time for all patients in a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by a second embodiment of the present invention;

[0052] FIG4 is a graph showing changes in edema volume over time for subclass 0 patients in a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by a second embodiment of the present invention;

[0053] FIG5 is a graph showing changes in edema volume over time for subcategory 1 patients in a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by a second embodiment of the present invention;

[0054] FIG6 is a graph showing changes in edema volume over time for subcategory 2 patients in a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by a second embodiment of the present invention;

[0055] FIG7 is a graph showing changes in edema volume over time for subcategory 3 patients in a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans provided by a second embodiment of the present invention. DETAILED DESCRIPTION

[0056] 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.

[0057] Example 1

[0058] 1 and 2 , an embodiment of the present invention provides a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans, including:

[0059] S1: Preprocess the edema volume data of each patient at different time points.

[0060] The edema volume data of each patient at different time points were cleaned, errors and outliers were corrected, and the missing edema volume data were filled with mean values ​​and linear interpolated. The edema volume data were time-aligned according to the patient's examination time and onset time, and then the edema volume data were standardized.

[0061] S2: Use the Gaussian fitting algorithm to fit the preprocessed data, construct a curve showing the progression of edema volume over time for all patients, and calculate the residual between the curve and the true value.

[0062] The pre-processed edema volume data were sorted by patients and time points to form a structured dataset, where each row represented the edema volume of a patient at a specific time point.

[0063] The Gaussian function formula is selected as:

[0064] Where x represents the time point from onset to examination, a, b, and c are the parameters of the Gaussian function, representing the peak value, mean value, and width of the curve, respectively.

[0065] For each patient's data, the least squares method was used to estimate the parameters a, b, and c of the Gaussian model;

[0066] A Gaussian function was applied to the time series data of each patient to generate a fitting curve of edema volume over time.

[0067] The residual is defined as the difference between the predicted value of the Gaussian fitting curve and the actual observed value, and is calculated as:

[0068] Among them, R i represents the residual of the i-th data point, Y i represents the actual observed value, represents the predicted value of the Gaussian fitting curve at the same time point;

[0069] For each patient's edema volume data at each time point, the residual between the data and the corresponding point of the Gaussian fitting curve was calculated.

[0070] S3: Patients are divided into multiple subgroups through deep learning clustering algorithm. Gaussian function is fitted to the edema amount of patients in each subgroup at different time points, and the residuals of the fitted subclass curves and the true values ​​of each subclass are calculated.

[0071] The autoencoder is used to extract features from patient data. During the encoding process, the original data x is converted into a low-dimensional latent variable h through the encoder. The calculation formula is: Where W1 and b1 are the weight and bias of the encoder respectively, and σ is the activation function;

[0072] During the decoding process, the latent variable h is decoded back to the original data dimension to obtain the reconstructed data The calculation formula is Where W2 and b2 are the weights and biases of the decoder; the parameters of the autoencoder are optimized to minimize the difference between the original data x and the reconstructed data The difference between them is: MinimizeLoss = dist(X,X R )

[0073] The low-dimensional representation h obtained by training the autoencoder is used for K-Means clustering. According to the pre-set number of clusters k, the algorithm will divide the patients into k clusters. Each cluster consists of patients with similar characteristics, represented by (C1, C2, ... C k ), then the minimized square error E is expressed as:

[0074] Among them, μ i is cluster c i The mean vector of is expressed as:

[0075] Gaussian fitting was performed on the patient data in each subgroup divided by the deep learning clustering algorithm, and the residual between the Gaussian fitting curve of each subgroup and the actual edema volume data of each patient in the subgroup was calculated.

[0076] It should be noted that the deep learning clustering algorithm (Autoencoder K-Means Clustering) combines the feature learning of autoencoders with the clustering capabilities of K-means. It first uses the autoencoder training data to obtain a low-dimensional encoded representation H. It then uses the K-Means clustering algorithm to cluster H into K clusters, improving the performance of K-means clustering through the feature learning of the autoencoder.

[0077] S4: Calculate the change in edema volume, determine the change trend, combine it with the corresponding treatment plan, use the association rule algorithm to analyze the correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

[0078] Extract the edema volume change and corresponding treatment plan information from the preprocessed data set; define item sets X and Y, where X represents a specific treatment plan and Y represents a specific edema volume change trend;

[0079] To calculate support, the support of data item set X, support(X), is the ratio of the number of transactions containing X in D to the total number of transactions in D, as shown in the following formula:

[0080] Where D represents the development of edema, x represents the specific treatment plan selected, count(x) represents the number of specific treatment plans selected, and count(D) is the number of edema development trends suppressed. The support of the association rule X=>Y is equal to the support of the item set X∪Y, as shown in the following formula:

[0081] Where X represents a specific treatment option that was selected, Y represents another specific treatment option that was selected, and count(X∪Y) represents the number of treatment options that were selected. To calculate the confidence level that X=>Y, use the formula:

[0082] A rule that estimates the likelihood of a treatment being present given a given treatment; where Support(X∩Y) represents the support for both treatments being present, and Support(X) represents the support for the preceding treatment being present.

[0083] The Apriori algorithm is used to iteratively generate frequent item sets and corresponding association rules, and the rules with the highest support and confidence are identified from the generated association rules.

[0084] It should be noted that the Apriori algorithm is an iterative method, as shown in Figure 2. Using the k-order frequent item set L k Connect to generate (k+1) order candidate set C k+1 , and then perform pruning to obtain the (k+1)-order frequent item set L k+1 , scan the database multiple times until no frequent itemsets are found. After the frequent itemsets are mined, for a k-order frequent itemset L k , 2(2k-1) meaningful rules can be generated (the preceding and following items are not empty), but not all of them meet the conditions. k Only when the confidence of the non-empty subset S is greater than the minimum confidence, an association rule "s→(1-s)" that meets the requirements is generated.

[0085] The above embodiments also include a system for analyzing edema after hemorrhagic stroke and analyzing treatment plans, specifically:

[0086] Data preprocessing module: preprocess the edema data of each patient at different time points.

[0087] Gaussian fitting module: Use the Gaussian fitting algorithm to fit the above data set, construct a curve showing the progression of edema volume over time for all patients, and calculate the residual between the curve and the true value.

[0088] Patient subgroup classification module: Patients are divided into multiple subgroups through deep learning clustering algorithm, Gaussian function is fitted to the edema amount of patients in each subgroup at different time points, and the residuals of the fitted subclass curves and the true values ​​of each subclass are calculated.

[0089] Treatment plan association analysis module: Calculate the change in edema volume, determine the change trend, combine it with the corresponding treatment plan, use the association rule algorithm to analyze the degree of correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

[0090] The computer device may be a server. The computer device includes a processor, a memory, an input / output interface (I / O) and a communication interface. The processor, the memory and the input / output interface are connected via a system bus, and the communication interface is connected to the system bus via the input / output 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 the 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 input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans is implemented.

[0091] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may 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 may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided in this application 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 processor involved in the various embodiments provided in this application may be a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, etc., but are not limited to these.

[0092] Example 2

[0093] Figures 3 to 7 are an embodiment of the present invention, which provides a method for analyzing edema after hemorrhagic stroke and analyzing treatment plans. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and simulation / comparative experiments.

[0094] Figure 3 is a Gaussian function fitting curve of edema change over time for all patients and a Gaussian function fitting curve for each of the 100 patients divided into four subclasses using a deep learning clustering algorithm according to the present invention.

[0095] In real-world settings, the time from onset to the first hospitalization and examination is random, independent, and countable. The number and timing of subsequent examinations also exhibit strong statistical independence and randomness. Therefore, for a large amount of patient data, we removed data with zero subsequent examinations. The results of the Gaussian Fitting Algorithm curve fitting for a total of 450 patients are shown in Figure 3 below.

[0096] By observing Figure 3, it can be seen that the time points with larger edema volumes are mostly concentrated within xx hours after the onset of complications. The data show a trend of increasing and decreasing over time, with obvious peaks. It can be seen from the fitted data (curve) that it has relatively outstanding accuracy. It is obvious that the Gaussian Fitting Algorithm function fitting curve can well retain the distribution characteristics of the data and the superiority of the smoothness of the fitting curve. By calculating the residuals of the true value and the fitting curve corresponding to the time point, the serial number of the residual corresponding to the time point is found, and the residual values ​​corresponding to all the serial numbers of each patient are matched. Then, the mean square error is calculated for the residual values ​​of all the serial numbers corresponding to the patient, and the residual value that can finally reflect the true value of the patient and the fitting curve of the edema of all patients over time is obtained.

[0097] Similarly, to further explore individual differences in the temporal progression of edema volume, we incorporated patient history, medical history, and onset-related information. Using a deep learning algorithm, we clustered the 100 patients into four subclusters (Subcluster 0, Subcluster 1, Subcluster 2, and Subcluster 3). The curves were fitted using a Gaussian fitting algorithm, as shown in Figure 4.

[0098] As can be seen from Figures 4 to 7, the edema volume of patients in different subcategories is consistent with the following: within 1000 hours after onset, the edema volume changes more significantly, and after 1000 hours, the edema volume changes tend to be stable. By calculating the residuals between the true values ​​of each subcategory and its fitted curve, it can be concluded through data analysis that the residuals between the true data of each patient in each subcategory obtained based on the deep learning clustering algorithm and the curve fitted by the Gaussian (Guess) function are significantly lower than the residuals of the fitted curve of the edema volume change over time for all patients. It can be concluded that the curves of the subgroups obtained by the deep learning clustering algorithm and then fitted by the Gaussian (Guess) function are more consistent with the change pattern of the real data and closer to the actual situation, which is of great significance for analyzing the changes in patient conditions in the future medical field.

[0099] 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 method for analyzing edema after hemorrhagic stroke and analyzing treatment plans, characterized in that: include: The edema volume data of each patient at different time points were collected and preprocessed; The preprocessed data are fitted using a Gaussian fitting algorithm to construct a curve of the edema volume of all patients over time, and the residual between the curve and the true value is calculated; The patients were divided into multiple subgroups through deep learning clustering algorithm, and the edema volume of each subgroup was fitted with Gaussian function at different time points, and the residuals of the fitted subclass curves and the true values ​​of each subclass were calculated; Calculate the change in edema volume, determine the change trend, combine it with the corresponding treatment plan, use the association rule algorithm to analyze the correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

2. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 1, characterized in that: The preprocessing includes cleaning the edema volume data of each patient at different time points, correcting errors and outliers, performing mean value filling and linear interpolation on the missing edema volume data, time-aligning the edema volume data according to the patient's examination time and onset time, and then standardizing the edema volume data.

3. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 2, characterized in that: The fitting of the preprocessed data set using the Gaussian fitting algorithm includes sorting the preprocessed edema volume data by patient and time point to form a structured data set, where each row represents the edema volume of a patient at a specific time point; The Gaussian function formula is selected as: Wherein, v represents the time point from onset to examination, a, b and c are the parameters of the Gaussian function curve of the patient's edema with respect to time after fitting, a represents the peak value of the curve, b represents the mean value of the curve and c represents the width of the curve; For each patient’s data, the least squares method was used to estimate the parameters a, b, and c of the Gaussian model, and the gradient descent method was used for iterative parameter optimization; A Gaussian function was applied to the time series data of each patient to generate a fitting curve of edema volume over time.

4. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 3, characterized in that: The calculation of the residual between the curve and the true value includes defining the residual as the difference between the predicted value of the Gaussian fitting curve and the actual observed value, and the calculation formula is: Among them, R i represents the fitting residual at time point i, Y i represents the true value of the edema volume at time point i, represents the predicted value of the Gaussian fitting curve at the same time point; For each patient's edema volume data at each time point, the residual between the data and the corresponding point of the Gaussian fitting curve was calculated.

5. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 4, characterized in that: The method of dividing patients into multiple subgroups by deep learning clustering algorithm includes extracting features of patient data using autoencoders according to the personal information features of different patients. During the encoding process, the original various feature data x are converted into low-dimensional latent variables h by the encoder. The calculation formula is h=σ(W1m+b1), where W1 and b1 are the weight and bias of the encoder respectively, σ is the activation function, and m is the feature value under the feature encoder; In the decoding process, the latent variable h after the encoding process is decoded back to the original data dimension to obtain the reconstructed data The calculation formula is: Where W2 represents the weight of the feature decoder, and b2 represents the bias of the feature decoder; The parameters of the autoencoder are repeatedly optimized through the loss function, which is defined as minimizing the difference between the original data x and the reconstructed data The difference between the two is: MinimizeLoss=dist(G,G R ) Where G represents the original feature information dataset of each patient; G R Represents the original feature information dataset of each patient after being reconstructed by the encoder; The low-dimensional representation h obtained by training the autoencoder is used for K-Means clustering. According to the pre-set number of clusters k, the algorithm will divide the patients into k clusters, each of which consists of patients with similar characteristics, represented by (C1, C2, … C k ), then the minimized square error E is expressed as: Among them, ||.|| represents the norm of the vector, l represents the low-dimensional representation of patient information, μ i Represents cluster C i The mean vector of .

6. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 5, characterized in that: The residuals of the calculated fitting curves of each subclass and the true values ​​of each subclass include: Gaussian fitting was performed on the patient data in each subgroup divided by the deep learning clustering algorithm, and the residual between the Gaussian fitting curve of each subgroup and the actual edema volume data of each patient in the subgroup was calculated.

7. The method for analyzing edema after hemorrhagic stroke and analyzing treatment plans according to claim 6, characterized in that: The use of an association rule algorithm to analyze the degree of association between the treatment method and the edema progression trend includes extracting edema volume change and corresponding treatment plan information from a preprocessed data set; Define item sets X and Y, where X represents a specific treatment plan and Y represents a specific edema volume change trend; To calculate the support, the support of the data item set X, support(X), is the ratio of the number of transactions containing X in D to the total number of transactions in D. The formula is: Where D represents the development of edema, x represents the specific treatment plan selected, count(x) represents the number of specific treatment plans selected, and count(D) represents the number of edema development trends that have been inhibited; The support of the association rule X=>Y is equal to the support of the item set X∪Y, which can be expressed as: Where X represents a specific treatment option selected, Y represents another specific treatment option selected, and count(X∪Y) represents the number of treatment options selected; To calculate the confidence level of X=>Y, use the formula: Rules that assess the likelihood of another treatment option being present given a particular treatment option; where Support(X∩Y) represents the support for both treatment options, and Support(X) represents the support for the former treatment option being present; The Apriori algorithm is used to iteratively generate frequent item sets and corresponding association rules, and from the generated association rules, the rules with the highest support and confidence are identified.

8. A system for analyzing edema after hemorrhagic stroke and analyzing treatment plans using the method according to any one of claims 1 to 7, characterized in that: Data preprocessing module: preprocess the edema volume data of each patient at different time points; Gaussian fitting module: Use Gaussian fitting algorithm to fit the above data set, construct a curve of edema volume progression over time for all patients, and calculate the residual between the curve and the true value; Patient subgroup classification module: The deep learning clustering algorithm is used to divide patients into multiple subgroups. The edema volume of each subgroup is fitted with a Gaussian function at different time points, and the residuals of the fitted subgroup curves and the true values ​​of each subgroup are calculated. Treatment plan association analysis module: calculate the change in edema volume, determine the change trend, combine the corresponding treatment plan, use the association rule algorithm to analyze the degree of correlation between the treatment method and the edema progression trend, and identify the treatment plan with the highest correlation with the edema progression trend.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images