Disease diagnosis result prediction method and system, electronic equipment and medium
By constructing a fine-tuning sample data set and model fine-tuning, the problem of inefficient acquisition of disease diagnosis results in the prior art is solved, and efficient and accurate prediction of disease diagnosis results is achieved.
Patent Information
- Application Number
- CN202411526821.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-29
- Publication Date
- 2025-02-14
AI Technical Summary
In the prior art, the efficiency of obtaining disease diagnosis results is low, and the target user is required to upload a variety of medical record image information by themselves.
Through the target object's historical disease diagnosis data and preset disease diagnosis classification standards, a fine-tuned sample data set is constructed, and the preset disease prediction model is fine-tuned to obtain a sub-predictive model specifically for the target object for prediction of disease diagnosis results.
Without relying on the medical record images uploaded by the target user, electronic receipt information is directly obtained through the hospital database, which improves the efficiency of obtaining disease diagnosis results and ensures prediction accuracy.
Smart Images
Figure CN119446491A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data processing technology, and in particular to a disease diagnosis result prediction method, system, electronic equipment and medium. Background Art
[0002] In the current insurance claims business scenario, it is necessary to determine the target user's specific insurance claims business based on the target user's disease diagnosis results and data related to the diagnosis results. In the current solution for obtaining the target user's disease diagnosis results, it is necessary to obtain the target user's disease diagnosis results by performing graphic and text analysis on the medical record images uploaded by the target user. This method requires the target user to upload a wide variety of medical record image information by himself, resulting in low efficiency in obtaining disease diagnosis results. Summary of the invention
[0003] Based on the above problems, in order to improve the efficiency of obtaining disease diagnosis results, the embodiments of the present application provide a disease diagnosis result prediction method, system, electronic device and medium.
[0004] The embodiments of the present application disclose the following technical solutions:
[0005] In a first aspect, an embodiment of the present application provides a method for predicting disease diagnosis results, comprising:
[0006] In response to a disease diagnosis result prediction instruction for a target object, historical disease diagnosis data of the target object is acquired; the historical disease diagnosis data includes: historical disease diagnosis results of the target object in historical medical consultation events, and first electronic receipt information corresponding to the historical medical consultation events alone;
[0007] Based on the historical disease diagnosis data and the preset disease diagnosis classification standard, construct a fine-tuning sample data set;
[0008] Fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model;
[0009] The second electronic receipt information of the target object in this medical consultation event is obtained, and the disease diagnosis result is predicted based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current medical consultation event.
[0010] In a possible implementation, constructing a fine-tuning sample data set based on the historical disease diagnosis data and a preset disease diagnosis classification standard includes:
[0011] Performing data standardization processing on the first electronic receipt information to obtain standardized first electronic receipt information;
[0012] Based on the preset disease diagnosis classification standard, the historical disease diagnosis results are classified to obtain historical disease classification results;
[0013] The fine-tuning sample data set is constructed according to the standardized first electronic receipt information and the historical disease classification results.
[0014] In a possible implementation, fine-tuning a preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model includes:
[0015] Write a model fine-tuning script file according to the preset fine-tuning framework;
[0016] Based on the model fine-tuning script file and the fine-tuning sample data set, the preset disease prediction model is fine-tuned to obtain the sub-prediction model.
[0017] In a possible implementation, the acquiring of the second electronic receipt information of the target object in the current medical consultation event, and predicting the disease diagnosis result according to the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current medical consultation event includes:
[0018] Performing a preliminary prediction on the second electronic receipt information by using the sub-prediction model to obtain a plurality of preliminary disease diagnosis prediction results;
[0019] Calculating each corresponding correlation score between each of the preliminary disease diagnosis prediction results and the second electronic receipt information;
[0020] With respect to the score ranking composed of each of the correlation scores, the preliminary disease diagnosis prediction result with the score ranking higher than a preset threshold is determined as the disease diagnosis prediction result of the target object in the current consultation event.
[0021] In a possible implementation, after obtaining the second electronic receipt information of the target object in the current consultation event, and predicting the disease diagnosis result according to the second electronic receipt information and the sub-prediction model, and obtaining the disease diagnosis prediction result of the target object in the current consultation event, the method further includes:
[0022] When the disease diagnosis prediction result does not match the preset disease diagnosis classification standard, data marking is performed on the disease diagnosis prediction result.
[0023] In a second aspect, the present application embodiment provides a disease diagnosis result prediction system, comprising:
[0024] A historical data acquisition module, used to acquire historical disease diagnosis data of a target object; the historical diagnosis data includes: historical disease diagnosis results of the target object in historical medical consultation events, and first electronic receipt information corresponding solely to the historical disease diagnosis results;
[0025] A fine-tuning sample construction module, used to construct a fine-tuning sample data set based on the historical disease diagnosis data and the preset disease diagnosis classification standard;
[0026] A prediction model fine-tuning module, used to fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model;
[0027] The diagnosis result prediction module is used to obtain the second electronic receipt information of the target object in this consultation event, and predict the disease diagnosis result based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current consultation event.
[0028] In a possible implementation, the fine-tuning sample construction module is specifically used to:
[0029] Performing data standardization processing on the first electronic receipt information to obtain standardized first electronic receipt information;
[0030] Based on the preset disease diagnosis classification standard, the historical disease diagnosis results are classified to obtain historical disease classification results;
[0031] The fine-tuning sample data set is constructed according to the standardized first electronic receipt information and the historical disease classification results.
[0032] In a possible implementation, the prediction model fine-tuning module is specifically used to:
[0033] Write a model fine-tuning script file according to the preset fine-tuning framework;
[0034] Based on the model fine-tuning script file and the fine-tuning sample data set, the preset disease prediction model is fine-tuned to obtain the sub-prediction model.
[0035] In a third aspect, an embodiment of the present application provides an electronic device, the device comprising: a processor, a memory, and a system bus;
[0036] The processor and the memory are connected via the system bus;
[0037] The memory is used to store one or more programs, and the one or more programs include instructions, which, when executed by the processor, enable the processor to perform the steps of any possible disease diagnosis result prediction method in the first aspect.
[0038] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any possible disease diagnosis result prediction method in the first aspect.
[0039] Compared with the prior art, the present application has the following beneficial effects: the embodiment of the present application provides a disease diagnosis result prediction method, system, electronic device and medium, in which, when receiving a disease diagnosis result prediction instruction for a target object, a fine-tuning sample data set for a preset disease prediction model is constructed through the historical disease diagnosis data of the target object and the preset disease diagnosis classification standard, and the preset disease prediction model is fine-tuned to obtain a sub-prediction model specifically for predicting disease diagnosis results for the target object. Among them, the historical disease diagnosis data includes the disease diagnosis results in the historical consultation events, and the corresponding first electronic receipt information. Since the electronic receipt information of the target object can be directly obtained through the hospital's database, the prediction of the disease diagnosis result for the target object does not need to rely on the medical record images uploaded by the target object, thereby improving the efficiency of obtaining the disease diagnosis result. At the same time, through the fine-tuning sample data set specific to the target object, the model of the preset disease prediction model is fine-tuned, and the prediction accuracy of the disease diagnosis result can be simultaneously guaranteed. In this way, when predicting the disease diagnosis result of the target object, it is only necessary to directly obtain the second electronic receipt information of the target object in this consultation event, as well as its corresponding historical disease diagnosis data, to predict the corresponding disease diagnosis result. There is no need to rely on the target object to upload the image information of the medical record by himself, which effectively improves the efficiency of obtaining the disease diagnosis result. BRIEF DESCRIPTION OF THE DRAWINGS
[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.
[0041] Figure 1 A flowchart of a disease diagnosis result prediction method provided in an embodiment of the present application;
[0042] Figure 2 A schematic diagram of online prediction of disease diagnosis results provided in an embodiment of the present application;
[0043] Figure 3 A schematic diagram of a flow chart of another disease diagnosis result prediction method provided in an embodiment of the present application;
[0044] Figure 4 A schematic diagram of the structure of a disease diagnosis result prediction system provided in an embodiment of the present application;
[0045] Figure 5 A schematic diagram of the structure of an electronic device for predicting disease diagnosis results provided in an embodiment of the present application. DETAILED DESCRIPTION
[0046] In order to make the purpose, technical solutions and advantages of this application more clear, the following is a further detailed description of this application in combination with specific embodiments and with reference to the accompanying drawings. It should be noted that the embodiments described in the embodiments of this application are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.
[0047] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "include" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right", etc. are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.
[0048] As described above, in the current insurance claims business scenario, it is necessary to determine the target user's specific insurance claims business based on the target user's disease diagnosis results and data related to the diagnosis results. In the current solution for obtaining the target user's disease diagnosis results, it is necessary to obtain the target user's disease diagnosis results by performing graphic and text analysis on the medical record images uploaded by the target user. This method requires the target user to upload a wide variety of medical record image information by himself, resulting in low efficiency in obtaining disease diagnosis results.
[0049] In order to solve the above problems, the embodiments of the present application provide a disease diagnosis result prediction method, system, electronic device and medium. In the method, when a disease diagnosis result prediction instruction for a target object is received, a fine-tuning sample data set for a preset disease prediction model is constructed through the historical disease diagnosis data of the target object and the preset disease diagnosis classification standard, and the preset disease prediction model is fine-tuned to obtain a sub-prediction model specifically for predicting disease diagnosis results for the target object. Among them, the historical disease diagnosis data includes the disease diagnosis results in the historical consultation events, and the corresponding first electronic receipt information. Since the electronic receipt information of the target object can be directly obtained through the hospital's database, the prediction of the disease diagnosis result for the target object does not need to rely on the medical record images uploaded by the target object, thereby improving the efficiency of obtaining the disease diagnosis result. At the same time, the preset disease prediction model is fine-tuned through the fine-tuning sample data set specific to the target object, and the prediction accuracy of the disease diagnosis result can be simultaneously guaranteed. In this way, when predicting the disease diagnosis result of the target object, it is only necessary to directly obtain the second electronic receipt information of the target object in this consultation event, as well as its corresponding historical disease diagnosis data, to predict the corresponding disease diagnosis result. There is no need to rely on the target object to upload the image information of the medical record by himself, which effectively improves the efficiency of obtaining the disease diagnosis result.
[0050] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0051] It should be noted that in the disease diagnosis result prediction method provided in the embodiment of the present application, the target object may be a target patient for whom disease diagnosis result prediction is required, and the disease diagnosis result prediction method provided in the embodiment of the present application may be applied to the disease diagnosis result prediction for target patients in the hospital. In order to facilitate the understanding of the prediction method provided in the embodiment of the present application, the target patient in the hospital is used as an example of the target object to introduce the above-mentioned disease diagnosis result prediction method.
[0052] See also Figure 1 , which is a flow chart of a disease diagnosis result prediction method provided in an embodiment of the present application, and specifically includes the following steps:
[0053] S101: In response to a disease diagnosis result prediction instruction for a target object, historical disease diagnosis data of the target object is obtained; the historical disease diagnosis data includes: historical disease diagnosis results of the target object in historical medical consultation events, and first electronic receipt information corresponding solely to the historical disease diagnosis results.
[0054] The disease diagnosis result prediction method provided in the embodiment of the present application is intended to quickly complete the prediction of the disease diagnosis result through the electronic receipt information of the target patient and the preset disease prediction model. However, in order to ensure the prediction accuracy of the disease diagnosis result as much as possible, when responding to the disease diagnosis result prediction instruction of the target patient, it is necessary to obtain the target patient's consultation data in the past period of time, and use this to fine-tune the prediction model, so that the output result of the prediction model can be closer to the actual situation of the target patient, thereby improving the accuracy of the disease diagnosis result prediction.
[0055] First, respond to the disease diagnosis result prediction instruction for the target patient to obtain the historical disease diagnosis data of the target patient, wherein the historical disease diagnosis data of the target patient includes the historical disease diagnosis results of the target patient in the past historical consultation events, and the first electronic receipt information generated in each historical consultation event.
[0056] In one possible implementation, the historical consultation events of the target patient can be obtained based on a specific time screening length. For example, the consultation events of the target patient in the hospital within the past six months are screened as historical consultation events. In one possible implementation, in order to ensure the accuracy of subsequent model fine-tuning, on the basis of historical consultation events based on a specific time screening length, the historical consultation events can also be secondary screened according to the specified disease diagnosis results. For example, if the specified disease diagnosis result is set to respiratory tract infection, then the consultation events within six months with the disease diagnosis result of respiratory tract infection can be determined as historical consultation events.
[0057] The historical disease diagnosis data includes, in addition to the historical disease diagnosis results corresponding to each historical medical consultation event of the target patient, the first electronic receipt information corresponding to each historical medical consultation event of the target patient.
[0058] The first electronic receipt information is used to represent the target patient's personal information, as well as all relevant medical consultation matters and prescription drugs involved in the corresponding consultation event. For the target patient's electronic receipt information, please refer to the following example:
[0059] "Target patient age: 5 years old, gender: female, type of visit: outpatient, department: internal medicine, examination items: digital imaging, digital photography (DR), examination site: chest, test items: high-sensitivity C-reactive protein assay, blood routine (five categories), drug name: desloratadine dry suspension, (Etanjing) ambrobutrol oral solution, (Nesona) mometasone furoate nasal spray, medical device items: none, surgical items: none."
[0060] It can be seen that the target patient's first electronic receipt information corresponding to each historical consultation event can be used to determine the consultation and treatment items performed by the target patient in the historical consultation event, as well as the prescribed drugs assigned. The specific first electronic receipt information and the actual historical disease diagnosis results can be used as historical disease diagnosis data to effectively reflect the target patient's disease diagnosis in the past period of time, and it can be used as a basis for fine-tuning the preset disease prediction model, which can effectively improve the prediction accuracy of the disease diagnosis results.
[0061] S102: Constructing a fine-tuning sample data set based on the historical disease diagnosis data and the preset disease diagnosis classification standard.
[0062] Accordingly, after obtaining the historical disease diagnosis data of the target patient in the past period of time, in order to ensure that the model fine-tuning of the preset disease prediction model can be effectively completed, it is necessary to construct a fine-tuning sample data set with a standard data structure through the historical disease diagnosis data and the preset disease diagnosis classification standards, and use this as the data basis for model fine-tuning.
[0063] Specifically, it is necessary to first perform data standardization processing on the first electronic receipt information of the target patient to convert the first electronic receipt information into receipt information with a standard data structure. Taking the above example of the electronic receipt information of the target patient as an example, after performing data standardization processing on the first electronic receipt information, the first electronic receipt information obtained after standardization processing is shown in the following table:
[0064]
[0065] After completing the data standardization processing for the first electronic receipt information, the historical disease diagnosis results of the target patient are further classified by pre-set disease diagnosis classification standards to increase the standardization and accuracy of the data.
[0066] In a possible implementation, the preset disease diagnosis classification standard may be the ICD10 standard. The ICD10 standard is a globally unified medical classification system developed and published by the World Health Organization (WHO). It is used to encode diseases, symptoms, signs, abnormal findings, external causes, and the social environment of the disease to facilitate the standardization, statistics, and analysis of medical data.
[0067] Specifically, the preset disease diagnosis classification standards determined by ICD10 can be seen in the following table:
[0068]
[0069]
[0070] It can be seen that by classifying the historical disease diagnosis results of the target patients according to the ICD10 standard and using the classified disease diagnosis names as the mapping standard, the historical disease classification results can be obtained.
[0071] In this way, after obtaining the standardized first electronic invoice information and the historical disease classification results, the standardized first electronic invoice information is mapped with its corresponding historical disease classification results, and assembled according to the case dimension to construct a fine-tuning sample data set. An example of the fine-tuning sample data set is as follows:
[0072] "instruction":"Infer the target patient's disease diagnosis based on the following information"
[0073] "input":"Target patient age: 5 years old, gender: female, type of visit: outpatient, department: internal medicine, examination items: digital imaging, digital photography (DR), examination site: chest, test items: high-sensitivity C-reactive protein assay, blood routine (five categories), drug name: desloratadine dry suspension, (Etanjing) ambrobutrol oral solution, (Nasonex) mometasone furoate nasal spray, medical device project: none, surgical project: none"
[0074] "output":"Disease diagnosis: upper respiratory tract infection, allergic rhinitis"
[0075] S103: Fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model.
[0076] After the fine-tuning sample data set is constructed, the preset disease prediction model is fine-tuned through the fine-tuning sample data set, so as to optimize the prediction model by combining the disease diagnosis results of the target patient in the historical consultation events and the corresponding electronic receipt information, and obtain a sub-prediction model specifically for disease diagnosis structure prediction for the target patient.
[0077] In a possible implementation, the preset disease prediction model can adopt an open source medical health base model (such as LLama3, Bian Que, etc.). A wealth of medical knowledge is integrated in the open source medical health base model, such as drug instructions, medical encyclopedia knowledge, etc. Through the rich medical knowledge in the open source medical health base model, the disease diagnosis results of the target patient can be effectively predicted based on the surgical items, drug types, and other information indicated in the bill information, thereby improving the efficiency of obtaining disease diagnosis results.
[0078] Specifically, in the process of fine-tuning the preset disease prediction model through the fine-tuning sample set, it is necessary to pre-select a specific preset fine-tuning framework to write a model fine-tuning script file, and based on the model fine-tuning script file and the fine-tuning sample data set, fine-tune the preset disease prediction model to obtain a sub-prediction model.
[0079] Next, taking the preset disease prediction model as LLama3 and the preset fine-tuning framework as LLaMA Factory as an example, the fine-tuning process of the preset disease prediction model is explained.
[0080] After LLaMA Factory and LLama3 have been downloaded, create a configuration file related to fine-tuning training, and then write a model fine-tuning script file based on the configuration file in its framework. The fine-tuning script file includes the Lora fine-tuning method.
[0081] When the fine-tuning script file has been written, import the previously generated fine-tuning sample data set, set the corresponding fine-tuning parameters batch and learning_rate, and determine the save path of the sub-prediction model. Finally, in the LLaMAFactory directory, run the corresponding model fine-tuning script to fine-tune the model LLama3 according to the fine-tuning sample data set specific to the target patient, and obtain the corresponding sub-prediction model Bill_Diagnose_LLM file. Finally, perform an effect comparison test on the generated sub-prediction model to determine the model that is used to predict the disease diagnosis results of the target patient.
[0082] S104: Obtain the second electronic receipt information of the target object in this medical consultation event, and predict the disease diagnosis result based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current medical consultation event.
[0083] Finally, the second electronic receipt information of the target patient in this consultation event is obtained from the hospital's management database through direct acquisition, and the second electronic receipt information is used as the input data of the sub-prediction model to predict the disease diagnosis result, so as to obtain the disease diagnosis result of the target patient in this consultation event.
[0084] Specifically, the process of predicting the disease diagnosis result by using the second electronic receipt information of the target patient and the sub-prediction model can be found in Figure 2 The schematic diagram of online prediction of disease diagnosis results shown is combined for understanding.
[0085] It can be seen that, unlike the diagnosis result prediction method that requires the target patient to upload medical record image information by himself, the target patient's electronic receipt information can be directly obtained from the hospital's management database for electronic receipt information through a direct connection. Since the electronic receipt information can be directly obtained through a data direct connection, the method of obtaining the electronic receipt information is more convenient and rapid than the medical record image information uploaded by the user in the past. At the same time, because the sub-prediction model used to predict the target patient's disease diagnosis result is obtained by fine-tuning the target patient's own historical disease diagnosis data, the prediction of the target patient's disease diagnosis result is also guaranteed. In this way, the embodiment of the present application achieves the improvement of the prediction efficiency of the disease diagnosis result while ensuring the accuracy of the disease prediction.
[0086] Next, the process of predicting the disease diagnosis result through the second electronic receipt information of the target object in the above step S104 will be introduced in conjunction with the drawings of the specific embodiments.
[0087] See also Figure 3 , which is a flow chart of another disease diagnosis result prediction method provided in an embodiment of the present application, which specifically includes the following steps:
[0088] S201: Perform a preliminary prediction on the second electronic invoice information using the sub-prediction model to obtain a plurality of preliminary disease diagnosis prediction results.
[0089] First, the second electronic receipt information of the target patient in this consultation event is used as the input data of the sub-prediction model to perform a preliminary prediction of the disease diagnosis result, and multiple preliminary disease diagnosis prediction results can be obtained.
[0090] S202: Calculate each corresponding correlation score between each of the preliminary disease diagnosis prediction results and the second electronic receipt information.
[0091] After obtaining multiple preliminary disease diagnosis prediction results, in order to select the prediction result with the highest accuracy from the multiple preliminary disease diagnosis prediction results, it is necessary to calculate the correlation scores corresponding to each preliminary disease prediction result and the second electronic receipt information. The higher the correlation score, the greater the correlation between the preliminary disease diagnosis prediction result and the second electronic receipt information, and the more accurate the corresponding prediction result.
[0092] The correlation score between the disease diagnosis prediction result and the second electronic receipt information can be calculated by a similarity measurement method. For example, similarity calculation based on ICD coding or text similarity based on specific medical service items (such as using TF-IDF, cosine similarity, etc.). If the medical service item information in the second electronic receipt is highly correlated with the preliminary disease diagnosis prediction result, its correlation score will be higher.
[0093] S203: For the score ranking composed of the respective correlation scores, the preliminary disease diagnosis prediction result having the score ranking higher than a preset threshold is determined as the disease diagnosis prediction result of the target subject in the current consultation event.
[0094] Finally, the correlation scores between each preliminary disease diagnosis prediction result and the second electronic receipt information are ranked by score, and the prediction result with a score ranking higher than the preset threshold is determined as the disease diagnosis prediction result of the target patient in this consultation event, thereby improving the prediction accuracy of the disease diagnosis result.
[0095] In addition, in a possible implementation, in order to reasonably avoid some unreasonable outputs, after determining the disease diagnosis prediction results of the target patient in this consultation event, the obtained disease diagnosis prediction results can be double-checked in combination with the aforementioned preset disease diagnosis classification standards. When the disease diagnosis prediction results do not match the preset disease diagnosis classification standards, the disease diagnosis prediction results are data-marked to improve the accuracy of the effective output and the quality of the output data.
[0096] For example, when the diagnostic prediction results output by the sub-prediction model do not match the preset disease diagnosis classification standards, the disease diagnosis results input to the model are reassigned to "other" to ensure the accuracy of the model input data.
[0097] An embodiment of the present application provides a method for predicting disease diagnosis results, in which, upon receiving a disease diagnosis result prediction instruction for a target object, a fine-tuning sample data set for a preset disease prediction model is constructed through the historical disease diagnosis data of the target object and the preset disease diagnosis classification standard, and the preset disease prediction model is fine-tuned to obtain a sub-prediction model specifically for predicting disease diagnosis results for the target object. Among them, the historical disease diagnosis data includes the disease diagnosis results in historical consultation events, and the corresponding first electronic receipt information. Since the electronic receipt information of the target object can be directly obtained through the hospital's database, the prediction of the disease diagnosis result for the target object does not need to rely on the medical record images uploaded by the target object, thereby improving the efficiency of obtaining the disease diagnosis results. At the same time, by fine-tuning the preset disease prediction model through the fine-tuning sample data set specific to the target object, the prediction accuracy of the disease diagnosis result can be simultaneously guaranteed. In this way, when predicting the disease diagnosis result of the target object, it is only necessary to directly obtain the second electronic receipt information of the target object in this consultation event, as well as its corresponding historical disease diagnosis data, to predict the corresponding disease diagnosis result. There is no need to rely on the target object to upload the image information of the medical record by himself, which effectively improves the efficiency of obtaining the disease diagnosis result.
[0098] A disease diagnosis result prediction system provided in an embodiment of the present application is introduced below. The disease diagnosis result prediction system described below and the disease diagnosis result prediction method described above can be referenced to each other.
[0099] See also Figure 4 , which is a schematic diagram of the structure of a disease diagnosis result prediction system provided in an embodiment of the present application, and specifically includes the following modules:
[0100] The historical data acquisition module 100 is used to acquire the historical disease diagnosis data of the target object; the historical diagnosis data includes: the historical disease diagnosis results of the target object in the historical medical consultation events, and the first electronic receipt information corresponding to the historical disease diagnosis results;
[0101] A fine-tuning sample construction module 200 is used to construct a fine-tuning sample data set based on the historical disease diagnosis data and the preset disease diagnosis classification standard;
[0102] A prediction model fine-tuning module 300 is used to fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model;
[0103] The diagnosis result prediction module 400 is used to obtain the second electronic receipt information of the target object in this consultation event, and predict the disease diagnosis result based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current consultation event.
[0104] In a possible implementation, the fine-tuning sample construction module 200 is specifically used to:
[0105] Performing data standardization processing on the first electronic receipt information to obtain standardized first electronic receipt information;
[0106] Based on the preset disease diagnosis classification standard, the historical disease diagnosis results are classified to obtain historical disease classification results;
[0107] The fine-tuning sample data set is constructed according to the standardized first electronic receipt information and the historical disease classification results.
[0108] In a possible implementation, the prediction model fine-tuning module 300 is specifically used to:
[0109] Write a model fine-tuning script file according to the preset fine-tuning framework;
[0110] Based on the model fine-tuning script file and the fine-tuning sample data set, the preset disease prediction model is fine-tuned to obtain the sub-prediction model.
[0111] See also Figure 5 , which is a schematic diagram of the structure of an electronic device for predicting disease diagnosis results provided by an embodiment of the present application, including:
[0112] A memory 11, used for storing computer programs;
[0113] The processor 12 is used to implement the steps of a disease diagnosis result prediction method described in any of the above method embodiments when executing the computer program.
[0114] In this embodiment, the device may be a vehicle-mounted computer, a PC (Personal Computer), or a terminal device such as a smart phone, a tablet computer, a PDA, or a portable computer.
[0115] The device may include a memory 11 , a processor 12 , and a bus 13 .
[0116] Among them, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the device, such as a hard disk of the device. In other embodiments, the memory 11 can also be an external storage device of the device, such as a plug-in hard disk equipped on the device, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Further, the memory 11 can also include both an internal storage unit of the device and an external storage device. The memory 11 can not only be used to store application software and various types of data installed in the device, such as program codes for executing fault prediction methods, etc., but can also be used to temporarily store data that has been output or is to be output. The processor 12 can be a central processing unit (CPU) in some embodiments.
[0117] In some embodiments, the processor 12 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor or other data processing chip, used to run program codes or process data stored in the memory 11, such as program codes for executing corresponding methods.
[0118] The bus 13 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or only one type of bus.
[0119] Furthermore, the device may also include a network interface 14, which may optionally include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is generally used to establish a communication connection between the device and other electronic devices.
[0120] Optionally, the device may further include a user interface 15, which may include a display (Display), an input unit such as a keyboard (Keyboard), and the optional user interface 15 may also include a standard wired interface and a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, and an OLED (Organic Light-Emitting Diode) touch device, etc. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the device and to display a visual user interface.
[0121] Figure 5 Only the device with components 11-15 is shown, and it can be understood by those skilled in the art that Figure 5 The structure shown does not constitute a limitation of the device, and may include fewer or more components than shown, or combine certain components, or arrange the components differently.
[0122] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the embodiment of the present application also provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the disease diagnosis result prediction method described in any of the above embodiments.
[0123] The computer-readable media of the embodiments of the present application include permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.
[0124] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the disease diagnosis result prediction method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0125] It should be noted that each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for methods, systems, electronic devices and media, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments. The methods, systems, electronic devices and media described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components indicated 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 scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.
[0126] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A method for predicting disease diagnosis results, characterized in that: include: In response to a disease diagnosis result prediction instruction for a target object, acquiring historical disease diagnosis data of the target object; The historical disease diagnosis data includes: the historical disease diagnosis results of the target object in the historical medical consultation events, and the first electronic receipt information corresponding to the historical medical consultation events; Based on the historical disease diagnosis data and the preset disease diagnosis classification standard, construct a fine-tuning sample data set; Fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model; The second electronic receipt information of the target object in this medical consultation event is obtained, and the disease diagnosis result is predicted based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current medical consultation event.
2. The method according to claim 1, characterized in that The constructing of a fine-tuning sample data set based on the historical disease diagnosis data and the preset disease diagnosis classification standard includes: Performing data standardization processing on the first electronic receipt information to obtain standardized first electronic receipt information; Based on the preset disease diagnosis classification standard, the historical disease diagnosis results are classified to obtain historical disease classification results; The fine-tuning sample data set is constructed according to the standardized first electronic receipt information and the historical disease classification results.
3. The method according to claim 1, characterized in that The fine-tuning of the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model includes: Write a model fine-tuning script file according to the preset fine-tuning framework; Based on the model fine-tuning script file and the fine-tuning sample data set, the preset disease prediction model is fine-tuned to obtain the sub-prediction model.
4. The method according to claim 1, characterized in that The acquiring the second electronic receipt information of the target object in the current consultation event, and predicting the disease diagnosis result according to the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current consultation event, includes: Performing a preliminary prediction on the second electronic receipt information by using the sub-prediction model to obtain a plurality of preliminary disease diagnosis prediction results; Calculating each corresponding correlation score between each of the preliminary disease diagnosis prediction results and the second electronic receipt information; With respect to the score ranking composed of each of the correlation scores, the preliminary disease diagnosis prediction result with the score ranking higher than a preset threshold is determined as the disease diagnosis prediction result of the target object in the current consultation event.
5. The method according to claim 1, characterized in that After obtaining the second electronic receipt information of the target object in the current consultation event, and predicting the disease diagnosis result according to the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current consultation event, the method further includes: When the disease diagnosis prediction result does not match the preset disease diagnosis classification standard, data marking is performed on the disease diagnosis prediction result.
6. A disease diagnosis result prediction system, characterized in that: include: A historical data acquisition module is used to acquire historical disease diagnosis data of the target object; The historical diagnosis data includes: the historical disease diagnosis results of the target object in the historical medical consultation events, and the first electronic receipt information corresponding to the historical disease diagnosis results; A fine-tuning sample construction module, used to construct a fine-tuning sample data set based on the historical disease diagnosis data and the preset disease diagnosis classification standard; A prediction model fine-tuning module, used to fine-tune the preset disease prediction model according to the fine-tuning sample data set to obtain a sub-prediction model of the preset disease prediction model; The diagnosis result prediction module is used to obtain the second electronic receipt information of the target object in this consultation event, and predict the disease diagnosis result based on the second electronic receipt information and the sub-prediction model to obtain the disease diagnosis prediction result of the target object in the current consultation event.
7. The system according to claim 6, characterized in that The fine-tuning sample construction module is specifically used for: Performing data standardization processing on the first electronic receipt information to obtain standardized first electronic receipt information; Based on the preset disease diagnosis classification standard, the historical disease diagnosis results are classified to obtain historical disease classification results; The fine-tuning sample data set is constructed according to the standardized first electronic receipt information and the historical disease classification results.
8. The system according to claim 6, characterized in that The prediction model fine-tuning module is specifically used for: Write a model fine-tuning script file according to the preset fine-tuning framework; Based on the model fine-tuning script file and the fine-tuning sample data set, the preset disease prediction model is fine-tuned to obtain the sub-prediction model.
9. An electronic device, characterized in that: The device comprises: a processor, a memory and a system bus; The processor and the memory are connected via the system bus; The memory is used to store one or more programs, and the one or more programs include instructions, which, when executed by the processor, enable the processor to perform the steps of the disease diagnosis result prediction method described in any one of claims 1-5.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the disease diagnosis result prediction method described in any one of claims 1 to 5 is implemented.