Method and system for processing hereditary angioedema symptom data
By extracting the characteristic vectors of hereditary angioedema symptoms data and inputting the trained probability prediction model of disease, the probability of suffering from hereditary angioedema is predicted, which solves the problem that diagnosis relies on expert experience in the prior art, improves diagnostic accuracy and reduces misdiagnosis and mistreatment.
Patent Information
- Application Number
- CN202411995886.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-31
- Publication Date
- 2025-05-13
AI Technical Summary
In the prior art, the diagnosis of hereditary angioedema depends on expert experience, is low in popularity, and is often misdiagnosed and mistreated, and lacks effective auxiliary diagnosis methods.
By obtaining the pending symptom data, extracting the symptom feature vector, and inputting it into the trained probability prediction model of disease, predicting the probability of suffering from hereditary angioedema. The model is trained based on sample symptom data and disease probability truth value labels, and adopts a gradient-enhanced decision tree model and a binary logic loss function.
This method can assist physicians in diagnosis, improve the diagnostic accuracy of hereditary angioedema, reduce misdiagnosis and mistreatment, and provide effective auxiliary diagnostic support in areas with low popularity.
Smart Images
Figure CN119993478A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of medical data processing, and in particular to a method and system for processing hereditary angioedema symptom data. Background Art
[0002] The incidence of hereditary angioedema (HAE) is about 1 / 50,000 to 1 / 10,000. It is a rare disease, and the awareness among doctors and patients is still very low, resulting in frequent misdiagnosis and mistreatment. At present, hereditary angioedema mainly relies on manual diagnosis by experts in this rare disease. The diagnostic logic and methods are highly dependent on expert experience and the popularity is low. Only top tertiary hospitals that specialize in this rare disease have the ability to diagnose it.
[0003] In recent years, with the development of data processing technology, the technology of using artificial intelligence for image and text recognition and image and text analysis has developed rapidly, which has made it possible to realize the concept of artificial intelligence analysis of symptoms of hereditary angioedema to assist in diagnosis. Therefore, how to realize a method that can analyze the symptom data of hereditary angioedema in combination with artificial intelligence technology to assist doctors in diagnosing hereditary angioedema is a technical problem that needs to be solved urgently. Summary of the invention
[0004] The present invention provides a method and system for processing hereditary angioedema symptom data, which are used to solve the above-mentioned technical problems existing in the prior art.
[0005] The present invention provides a method for processing hereditary angioedema symptom data, comprising the following steps.
[0006] Get pending symptom data.
[0007] Extracting the symptom feature vector of the symptom data to be processed.
[0008] The symptom feature vector is input into a disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0009] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0010] According to a method for processing hereditary angioedema symptom data provided by the present invention, a symptom feature vector of the symptom data to be processed is extracted, including: for any symptom feature in the symptom data to be processed, if the any symptom feature belongs to hereditary angioedema symptoms, the any symptom feature is set to 1, otherwise it is set to 0, so as to obtain the symptom feature vector of the symptom data to be processed.
[0011] According to a method for processing hereditary angioedema symptom data provided by the present invention, the training process of the disease probability prediction model is as follows.
[0012] Obtain sample symptom data and a true value label of the disease probability corresponding to the sample symptom data.
[0013] A sample symptom feature vector of the sample symptom data is extracted.
[0014] The sample symptom feature vector and the true value label of the disease probability are substituted into the disease probability prediction model, and the loss function of the disease probability prediction model is solved. When the loss function converges, the model training is completed, otherwise, the parameters of the disease probability prediction model are adjusted until the loss function converges.
[0015] A method for processing hereditary angioedema symptom data according to the present invention further includes the following steps before extracting the sample symptom feature vector of the sample symptom data.
[0016] The sample symptom data is cleaned to delete the sample symptom data with missing data items and the repeated sample symptom data.
[0017] According to a method for processing hereditary angioedema symptom data provided by the present invention, the disease probability prediction model is a gradient boosting decision tree model, wherein the depth of each decision tree is determined based on the number of sample symptom features in the sample symptom feature vector.
[0018] According to a method for processing hereditary angioedema symptom data provided by the present invention, the loss function is a binary logistic loss function.
[0019] The present invention also provides a hereditary angioedema symptom data processing system, comprising the following modules.
[0020] The data acquisition module is used to acquire the symptom data to be processed.
[0021] The feature extraction module is used to extract the symptom feature vector of the symptom data to be processed.
[0022] The model execution module is used to input the symptom feature vector into the disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0023] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0024] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for processing hereditary angioedema symptom data as described in any one of the above is implemented.
[0025] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the hereditary angioedema symptom data processing method as described in any one of the above.
[0026] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the method for processing hereditary angioedema symptom data as described in any one of the above is implemented.
[0027] The hereditary angioedema symptom data processing method and system provided by the present invention obtains the symptom data to be processed; extracts the symptom feature vector of the symptom data to be processed; inputs the symptom feature vector into the disease probability prediction model, and obtains the probability of suffering from hereditary angioedema output by the disease probability prediction model; wherein the disease probability prediction model is trained based on the sample symptom data and the disease probability true value label corresponding to the sample symptom data. In the present invention, the disease probability prediction model is trained by the sample symptom data and its corresponding label, and then the disease probability prediction model is used to predict the symptom data to be processed of the patient to be diagnosed, so as to obtain the probability of suffering from hereditary angioedema, which is used as a reference for the physician's diagnosis and is helpful for the final diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS
[0028] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0029] Figure 1 This is one of the flow charts of the method for processing hereditary angioedema symptom data provided by the present invention.
[0030] Figure 2 It is a curve diagram of the evaluation index AUC (Area Under Curve) when evaluating the model in the hereditary angioedema symptom data processing method provided by the present invention.
[0031] Figure 3It is a curve diagram of the evaluation index precision (Precision)-recall rate (Recall) when evaluating the model in the hereditary angioedema symptom data processing method provided by the present invention.
[0032] Figure 4 It is a schematic diagram of the structure of the hereditary angioedema symptom data processing system provided by the present invention.
[0033] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION
[0034] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.
[0035] The hereditary angioedema symptom data processing method of the embodiment of the present invention is as follows: Figure 1 As shown, the following steps S110 to S130 are included.
[0036] Step S110: Acquire the symptom data to be processed. In this embodiment, the symptom data to be processed is extracted from the medical records of the patient to be diagnosed, which are usually in a txt format. Specifically, the symptom data to be processed is extracted by identifying keywords related to the symptoms of hereditary angioedema in the medical records.
[0037] It should be noted that the symptoms of hereditary angioedema include: 'edema', 'swelling', 'edema', 'abdominal pain', 'abdominal distension', 'ascites', 'laryngeal edema', 'dyspnea', 'laryngeal edema', 'shortness of breath', 'shortness of breath', 'stomach pain', 'vomiting', 'intermittent', 'attack', 'rash', 'itching' and 'urticaria'. If the symptoms of the patient to be diagnosed include some or most of the above symptoms, it means that the patient may suffer from hereditary angioedema. Among them, 'intermittent' and 'attack' indicate the onset of symptoms. 'intermittent' means continuous and non-sudden onset, and 'attack' means sudden onset. If the symptoms occur continuously and non-suddenly, it means that the patient may not have this disease.
[0038] Step S120: extract the symptom feature vector of the symptom data to be processed, mainly to perform numerical mapping on the non-numerical symptom features in the symptom data to be processed. Specifically, for any symptom feature in the symptom data to be processed, if any symptom feature belongs to the symptom of hereditary angioedema, then any symptom feature is set to 1, otherwise it is set to 0, so as to obtain the symptom feature vector of the symptom data to be processed. For example: the current symptom data of a patient to be diagnosed is ['edema', 'swelling', 'edema', 'abdominal pain', 'abdominal distension', 'ascites', 'laryngeal edema', 'dyspnea', 'laryngeal edema', 'holding breath', 'short breathing', 'stomach pain', 'vomiting', 'attack'], and there are no symptoms of 'rash', 'itching' and 'urticaria', then the symptom feature vector of the patient to be diagnosed is [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 0].
[0039] Step S130: Input the symptom feature vector into the disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model. The disease probability prediction model is trained based on sample symptom data and the disease probability true value label corresponding to the sample symptom data. The probability of suffering from hereditary angioedema output by the disease probability prediction model can provide a reference for doctors. For example, if the disease probability is greater than 60%, the doctor can confirm whether the patient suffers from hereditary angioedema based on other clinical manifestations of the patient.
[0040] Specifically, the sample symptom data includes positive samples and negative samples. The positive samples are: symptom data of patients who are determined to have hereditary angioedema, and the corresponding disease probability true value label is 1. The negative samples are: symptom data of patients who are determined not to have hereditary angioedema, that is, symptom data of suspected patients but ultimately determined not to be ill, and the corresponding disease probability true value label is 0. In this embodiment, all sample symptom data are divided into training set, validation set and test set. The training set is used to train the disease probability prediction model, the validation set is used to evaluate the performance and parameter tuning of the disease probability prediction model, and the test set is used to evaluate the final effect of the disease probability prediction model. Reasonable division can accurately evaluate the generalization ability of the model and avoid overfitting or underfitting problems. For example, in this embodiment, the ratio of the number of sample data in the training set, validation set and test set is: 975:420:373, and the proportion of positive samples in each set is: 0.55171, 0.237557 and 0.210973 respectively. Figure 2 and Figure 3As shown in the figure, based on the test set, the trained disease probability prediction model is evaluated using evaluation indicators such as AUC and precision-recall. AUC is defined as the area under the ROC curve and the coordinate axis. The AUC value is 0.96, and the disease probability prediction model is in line with expectations. It can also be seen from the precision-recall curve that the disease probability prediction model is in line with expectations.
[0041] The hereditary angioedema symptom data processing method of the present embodiment trains a disease probability prediction model through sample symptom data and its corresponding labels, and then uses the disease probability prediction model to predict the symptom data to be processed of the patient to be diagnosed, thereby obtaining the probability of suffering from hereditary angioedema for reference by doctors and facilitating the final diagnosis.
[0042] In some embodiments, the training process of the disease probability prediction model is as follows.
[0043] Obtain sample symptom data and a true value label of the probability of illness corresponding to the sample symptom data. In this embodiment, the sample symptom data is the symptom data of confirmed patients and suspected patients (finally confirmed as not having the disease) in the historical medical treatment data, and the true value label of the probability of illness corresponding to the sample symptom data is the true value of the final diagnosis and the undiagnosed. If the disease is confirmed, the label corresponding to the sample symptom data is 1, and if the disease is confirmed as not having the disease, the label corresponding to the sample symptom data is 0.
[0044] Extract the sample symptom feature vector of the sample symptom data. Specifically, extract the sample symptom feature vector of the sample symptom data. The method of extracting the sample symptom feature vector is similar to the above step S120 and will not be repeated here.
[0045] The sample symptom feature vector and the true value label of the disease probability are substituted into the disease probability prediction model to solve the loss function of the disease probability prediction model. When the loss function converges, the model training is completed. Otherwise, the parameters of the disease probability prediction model are adjusted until the loss function converges. Different models have different loss functions.
[0046] In this embodiment, an existing mature classification model can be used as a basic model, and the basic model can be trained using the above-mentioned sample symptom data and its corresponding disease probability true value label to obtain a disease probability prediction model.
[0047] In some embodiments, before extracting the sample symptom feature vector of the sample symptom data, it also includes: performing data cleaning on the sample symptom data to delete sample symptom data with missing data items and repeated sample symptom data.
[0048] Specifically, in the sample symptom data, if some symptom features in the symptom feature vector are not determined to be 0 or 1, and the value corresponding to the symptom feature is empty, the sample symptom data is deleted. If there are multiple repeated sample symptom data, only one piece of data is retained.
[0049] In this embodiment, data cleaning of sample symptom data is helpful for model training and improves the accuracy and robustness of the model.
[0050] In some embodiments, the disease probability prediction model may be a gradient boosting decision tree model (GradientBoosting Decision Tree, GBDT), and the depth of each decision tree may be determined according to the number of sample symptom features in the sample symptom feature vector, for example: the depth of each decision tree is 4-8. The total number of decision trees may also be determined according to the number of sample symptom features, and each decision tree is classified according to different symptom features.
[0051] In some embodiments, the loss function is a binary logistic loss function (binary:logistic). Specifically, the loss function of each decision tree is a binary logistic loss function. When the loss function converges, the current decision tree training is completed, that is, the construction is completed. After all decision trees are constructed, the entire disease probability prediction model training is completed.
[0052] The hereditary angioedema symptom data processing system provided by the present invention is described below. The hereditary angioedema symptom data processing system described below and the hereditary angioedema symptom data processing method described above can be referenced to each other.
[0053] The hereditary angioedema symptom data processing system of the embodiment of the present invention is as follows: Figure 4 As shown, it includes the following modules.
[0054] The data acquisition module 410 is used to acquire the symptom data to be processed.
[0055] The feature extraction module 420 is used to extract the symptom feature vector of the symptom data to be processed.
[0056] The model execution module 430 is used to input the symptom feature vector into the disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0057] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0058] The hereditary angioedema symptom data processing system of the present embodiment trains a disease probability prediction model through sample symptom data and its corresponding labels, and then uses the disease probability prediction model to predict the symptom data to be processed of the patient to be diagnosed, thereby obtaining the probability of suffering from hereditary angioedema for reference by doctors and facilitating the final diagnosis.
[0059] In some embodiments, the feature extraction module 420 is specifically used to set any symptom feature in the symptom data to be processed to 1 if it belongs to hereditary angioedema symptoms, otherwise it is set to 0, so as to obtain the symptom feature vector of the symptom data to be processed.
[0060] In some embodiments, the training process of the disease probability prediction model is as follows.
[0061] Obtain sample symptom data and a true value label of the disease probability corresponding to the sample symptom data.
[0062] A sample symptom feature vector of the sample symptom data is extracted.
[0063] The sample symptom feature vector and the true value label of the disease probability are substituted into the disease probability prediction model, and the loss function of the disease probability prediction model is solved. When the loss function converges, the model training is completed, otherwise, the parameters of the disease probability prediction model are adjusted until the loss function converges.
[0064] In some embodiments, the hereditary angioedema symptom data processing system further includes: a data preprocessing module for performing data cleaning on the sample symptom data before extracting the sample symptom feature vector of the sample symptom data to delete sample symptom data with missing data items and repeated sample symptom data.
[0065] In some embodiments, the disease probability prediction model is a gradient boosting decision tree model, wherein the depth of each decision tree is determined based on the number of sample symptom features in the sample symptom feature vector.
[0066] In some embodiments, the loss function is a binary logistic loss function.
[0067] Figure 5 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5As shown, the electronic device may include: a processor 510, a communications interface 520, a memory 530, and a communication bus 540, wherein the processor 510, the communications interface 520, and the memory 530 communicate with each other via the communication bus 540. The processor 510 may call the logic instructions in the memory 530 to execute the hereditary angioedema symptom data processing method, which includes the following steps.
[0068] Get pending symptom data.
[0069] Extracting the symptom feature vector of the symptom data to be processed.
[0070] The symptom feature vector is input into a disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0071] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0072] In addition, the logic instructions in the above-mentioned memory 530 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.
[0073] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the hereditary angioedema symptom data processing method provided by the above-mentioned methods, which includes the following steps.
[0074] Get pending symptom data.
[0075] Extracting the symptom feature vector of the symptom data to be processed.
[0076] The symptom feature vector is input into a disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0077] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0078] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for processing hereditary angioedema symptom data provided by the above methods is implemented, and the method includes the following steps.
[0079] Get pending symptom data.
[0080] Extracting the symptom feature vector of the symptom data to be processed.
[0081] The symptom feature vector is input into a disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model.
[0082] The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
[0083] The system embodiment described above is merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.
[0084] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.
[0085] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for processing hereditary angioedema symptom data, characterized in that: include: Obtaining symptom data to be processed; Extracting a symptom feature vector of the symptom data to be processed; Inputting the symptom feature vector into a disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model; The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
2. The method for processing hereditary angioedema symptom data according to claim 1, characterized in that: Extracting the symptom feature vector of the symptom data to be processed includes: For any symptom feature in the symptom data to be processed, if the symptom feature belongs to hereditary angioedema symptoms, then the symptom feature is set to 1, otherwise it is set to 0, so as to obtain the symptom feature vector of the symptom data to be processed.
3. The method for processing hereditary angioedema symptom data according to claim 1, characterized in that: The training process of the disease probability prediction model is as follows: Obtain sample symptom data and a true value label of the disease probability corresponding to the sample symptom data; Extracting a sample symptom feature vector of the sample symptom data; The sample symptom feature vector and the true value label of the disease probability are substituted into the disease probability prediction model, and the loss function of the disease probability prediction model is solved. When the loss function converges, the model training is completed, otherwise, the parameters of the disease probability prediction model are adjusted until the loss function converges.
4. The method for processing hereditary angioedema symptom data according to claim 3, characterized in that: Before extracting the sample symptom feature vector of the sample symptom data, the method further includes: The sample symptom data is cleaned to delete the sample symptom data with missing data items and the repeated sample symptom data.
5. The method for processing hereditary angioedema symptom data according to claim 3, characterized in that: The disease probability prediction model is a gradient boosting decision tree model, wherein the depth of each decision tree is determined based on the number of sample symptom features in the sample symptom feature vector.
6. The method for processing hereditary angioedema symptom data according to claim 3, characterized in that: The loss function is a binary logistic loss function.
7. A hereditary angioedema symptom data processing system, characterized in that: include: A data acquisition module, used for acquiring symptom data to be processed; A feature extraction module, used for extracting a symptom feature vector of the symptom data to be processed; A model execution module, used for inputting the symptom feature vector into the disease probability prediction model to obtain the probability of suffering from hereditary angioedema output by the disease probability prediction model; The disease probability prediction model is obtained by training based on sample symptom data and the disease probability true value labels corresponding to the sample symptom data.
8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the hereditary angioedema symptom data processing method according to any one of claims 1 to 6 is implemented.
9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the hereditary angioedema symptom data processing method according to any one of claims 1 to 6 is implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the hereditary angioedema symptom data processing method according to any one of claims 1 to 6 is implemented.