WO2021120688A1 - 医疗误诊检测方法、装置、电子设备及存储介质 - Google Patents
医疗误诊检测方法、装置、电子设备及存储介质 Download PDFInfo
- Publication number
- WO2021120688A1 WO2021120688A1 PCT/CN2020/112235 CN2020112235W WO2021120688A1 WO 2021120688 A1 WO2021120688 A1 WO 2021120688A1 CN 2020112235 W CN2020112235 W CN 2020112235W WO 2021120688 A1 WO2021120688 A1 WO 2021120688A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- disease
- medical
- entity
- vector
- diagnosed
- Prior art date
Links
Images
Classifications
-
- G—PHYSICS
- G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
- G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
- G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
- G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Definitions
- This application relates to the field of artificial intelligence technology, and in particular to a medical misdiagnosis detection method, device, electronic equipment, and storage medium.
- the first aspect of the present application provides a medical misdiagnosis detection method.
- the medical misdiagnosis detection method includes:
- the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis.
- a second aspect of the present application provides an electronic device including a processor and a memory, and the processor is configured to execute computer-readable instructions stored in the memory to implement the following steps:
- the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis.
- a third aspect of the present application provides a computer-readable storage medium having at least one computer-readable instruction stored thereon, and the at least one computer-readable instruction is executed by a processor to implement the following steps:
- the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis.
- a fourth aspect of the present application provides a medical misdiagnosis detection device, and the medical misdiagnosis detection device includes:
- the obtaining unit is configured to obtain the chief complaint data and the disease to be diagnosed from the medical misdiagnosis detection request when the medical misdiagnosis detection request is received;
- the determining unit is configured to determine the user to be diagnosed according to the medical misdiagnosis detection request, and obtain the current medical history of the user to be diagnosed;
- An extraction unit for extracting entities in the main complaint data and entities in the current disease history to obtain disease entities
- the obtaining unit is further configured to obtain a target entity associated with the disease entity from a pre-built graph neural network, and obtain the weight of the target entity;
- a conversion unit configured to convert the target entity into a medical knowledge feature vector based on the weight
- a processing unit configured to process the main complaint data by using a convolutional neural network to obtain a text feature vector
- a splicing unit for splicing the medical knowledge feature vector and the text feature vector to obtain a target vector
- the input unit is used to input the target vector into the discriminant model to obtain a list of diseases
- the detection unit is used to detect whether the disease to be diagnosed exists in the disease list
- the determining unit is further configured to determine the response result of the medical misdiagnosis detection request as a misdiagnosis when the disease to be diagnosed does not exist in the disease list.
- this application determines the user to be diagnosed according to the medical misdiagnosis detection request, can accurately determine the user to be diagnosed, extracts the entity in the main complaint data and the entity in the current medical history, and obtains the disease Entity, because the user’s medical history will affect the diagnosis result, the current medical history of the user to be diagnosed is considered when analyzing the disease entity, which can improve the accuracy of misdiagnosis detection.
- the target entity associated with the disease entity, and the weight of the target entity is obtained.
- the weight of the target entity can be accurately determined, and then The medical knowledge feature vector can be accurately generated, and then the target vector is input into the discriminant model to obtain a list of diseases. Since the target vector considers the main complaint data and the current medical history of the user to be diagnosed, the target vector can be accurately determined.
- the disease list when the disease to be diagnosed does not exist in the disease list, the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis. Since the disease list contains multiple predicted diseases, the misdiagnosis can be improved
- the detection rate in addition, since this application is tested before the diagnosis report is issued to the user to be diagnosed, the effect of real-time warning can be achieved. This application is used in smart cities and smart medical care to promote the construction of smart cities.
- Fig. 1 is a flowchart of a preferred embodiment of the medical misdiagnosis detection method of the present application.
- Fig. 2 is a functional block diagram of a preferred embodiment of the medical misdiagnosis detection device of the present application.
- FIG. 3 is a schematic structural diagram of an electronic device according to a preferred embodiment of the method for detecting medical misdiagnosis in this application.
- FIG. 1 it is a flowchart of a preferred embodiment of the medical misdiagnosis detection method of the present application. According to different needs, the order of the steps in the flowchart can be changed, and some steps can be omitted.
- the medical misdiagnosis detection method is applied to a smart city, thereby promoting the construction of a smart city.
- the medical misdiagnosis detection method is applied to one or more electronic devices.
- the electronic device is a device that can automatically perform numerical calculation and/or information processing according to pre-set or stored instructions. Its hardware includes but not Limited to microprocessors, application specific integrated circuits (ASICs), programmable gate arrays (Field-Programmable Gate Arrays, FPGAs), digital processors (Digital Signal Processors, DSPs), embedded devices, etc.
- the electronic device may be any electronic product that can interact with a user with a human machine, for example, a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive network television ( Internet Protocol Television, IPTV), smart wearable devices, etc.
- a personal computer for example, a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive network television ( Internet Protocol Television, IPTV), smart wearable devices, etc.
- PDA personal digital assistant
- IPTV Internet Protocol Television
- smart wearable devices etc.
- the electronic device may also include a network device and/or user equipment.
- the network device includes, but is not limited to, a single network server, a server group composed of multiple network servers, or a cloud composed of a large number of hosts or network servers based on cloud computing.
- the network where the electronic device is located includes, but is not limited to, the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (Virtual Private Network, VPN), etc.
- the medical misdiagnosis detection method is applied in the field of artificial intelligence.
- the medical misdiagnosis detection request may be triggered by a medical staff, or before it is detected that the medical staff has issued a diagnosis certificate to the patient, which is not limited in this application.
- the information carried in the medical misdiagnosis detection request includes, but is not limited to: an identification code, a preset label, the main complaint data, the disease to be diagnosed, and the like.
- the electronic device acquiring the chief complaint data and the disease to be diagnosed from the medical misdiagnosis detection request includes:
- a second identifier is acquired, and information corresponding to the second identifier is acquired from the message information as the disease to be diagnosed.
- the first identifier and the second identifier are predefined identifiers, for example, the first identifier may be intru.
- the analysis efficiency of the medical misdiagnosis detection request can be improved.
- the mapping relationship between the first identification and the main complaint data and the second identification The mapping relationship with the disease to be diagnosed can accurately determine the disease to be diagnosed in the main complaint data set.
- S11 Determine a user to be diagnosed according to the medical misdiagnosis detection request, and obtain the current medical history of the user to be diagnosed.
- the user to be diagnosed refers to any user currently treated by medical staff
- the current medical history refers to diseases that the user to be diagnosed has experienced in the past.
- the electronic device determining the user to be diagnosed according to the medical misdiagnosis detection request, and acquiring the current medical history of the user to be diagnosed includes:
- the identification code is used to determine the user to be diagnosed, and information corresponding to the user to be diagnosed is obtained from the archive database as the current medical history.
- the medical history of multiple users corresponding to the user is stored in the archive library.
- the identification code is unique, the user to be diagnosed can be accurately determined through the identification code.
- the medical misdiagnosis detection request can be analyzed by directly obtaining idle threads from the thread connection pool. The creation time of the thread, therefore, improves the parsing speed of the medical misdiagnosis detection request, thereby improving the efficiency of the misdiagnosis detection.
- S12 Extract the entity in the main complaint data and the entity in the current disease history to obtain the disease entity.
- the disease entity refers to a disease that may occur in the user to be diagnosed, for example, the disease entity may be a cough.
- the electronic device extracting the entity in the main complaint data and the entity in the current medical history to obtain the disease entity includes:
- the entity in the main complaint data and the entity in the history of current illness are merged to obtain the disease entity.
- S13 Obtain a target entity associated with the disease entity from a pre-built graph neural network, and obtain a weight of the target entity.
- the graph neural network includes multiple entities, the attributes of each entity, and the degree of association between the entities and the attributes.
- the medical misdiagnosis detection method before acquiring the target entity associated with the disease entity from a pre-built graph neural network, the medical misdiagnosis detection method further includes:
- the current disease, the symptom attribute, and the degree of association are entered into a graph template to obtain the graph neural network.
- the above embodiment since the graph template is constructed in advance, the above embodiment does not need to repeatedly create the graph template, and therefore, the determination efficiency of the graph neural network can be improved.
- the obtaining the weight of the target entity includes:
- the electronic device converting the target entity into a medical knowledge feature vector based on the weight includes:
- S15 Use a convolutional neural network to process the main complaint data to obtain a text feature vector.
- the electronic device using a convolutional neural network to process the main complaint data to obtain a text feature vector includes:
- each main complaint vector includes multiple dimensions
- the obtained vector values are spliced to obtain the text feature vector.
- the preset value can be self-defined configuration, for example, the preset value can be 2.
- the target vector is generated by concatenating the medical knowledge feature vector and the text feature vector.
- the list of diseases includes multiple predicted diseases with high probability.
- the method before the target vector is input into the discriminant model to obtain the disease list, the method further includes:
- each training data includes the symptom data of each training user, the medical history of each training user and the corresponding symptoms;
- the discriminant model is constructed according to the splicing vector and the disease vector.
- the inputting the target vector into a discriminant model to obtain a disease list includes:
- the top N predicted diseases are selected from the queue, and the selected predicted diseases are merged to obtain the list of diseases, where N is a positive integer.
- the electronic device traverses the disease list, and when the disease to be diagnosed is the same as the predicted disease traversed in the disease list, it is determined that the disease to be diagnosed exists in all the diseases. In the disease list, or when the disease to be diagnosed is different from all predicted diseases traversed in the disease list, it is determined that the disease to be diagnosed does not exist in the disease list.
- the response result of the medical misdiagnosis detection request is determined to be a correct diagnosis, reward information is generated, and the user who issues the disease to be diagnosed is determined, And send the reward information to the terminal device of the issuing user.
- the reward information can be sent to the issuing user, thereby improving the experience of the issuing user.
- the above response result can also be stored in a node of a blockchain.
- the medical misdiagnosis detection method further includes:
- the ciphertext can be generated and sent in time after the misdiagnosis is determined, so as to achieve the effect of timely warning.
- it can also send out an alarm in time without receiving feedback within the preset time, which can achieve real-time warning.
- the effect is to ensure that the ciphertext can be received in time, thereby avoiding damage.
- this application determines the user to be diagnosed according to the medical misdiagnosis detection request, can accurately determine the user to be diagnosed, extracts the entity in the main complaint data and the entity in the current medical history, and obtains the disease Entity, since the user’s medical history will affect the diagnosis result, the current medical history of the user to be diagnosed is considered when analyzing the disease entity, which can improve the accuracy of misdiagnosis detection.
- the target entity associated with the disease entity, and the weight of the target entity is obtained.
- the weight of the target entity can be accurately determined, and then The medical knowledge feature vector can be accurately generated, and then the target vector is input into the discriminant model to obtain a list of diseases. Because the target vector takes into account the main complaint data and the current medical history of the user to be diagnosed, the target vector can be accurately determined.
- Disease list when the disease to be diagnosed does not exist in the disease list, the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis. Since the disease list contains multiple predicted diseases, the misdiagnosis can be improved The detection rate, in addition, since this application is tested before the diagnosis report is issued to the user to be diagnosed, the effect of real-time warning can be achieved.
- the medical misdiagnosis detection device 11 includes an acquisition unit 110, a determination unit 111, an extraction unit 112, a conversion unit 113, a processing unit 114, a splicing unit 115, an input unit 116, a detection unit 117, a calculation unit 118, an entry unit 119, and a construction unit 120.
- the module/unit referred to in this application refers to a series of computer program segments that can be executed by the processor 13 and can complete fixed functions, and are stored in the memory 12. In this embodiment, the functions of each module/unit will be described in detail in subsequent embodiments.
- the acquiring unit 110 When receiving the medical misdiagnosis detection request, acquires the main complaint data and the disease to be diagnosed from the medical misdiagnosis detection request.
- the medical misdiagnosis detection request may be triggered by a medical staff, or before it is detected that the medical staff has issued a diagnosis certificate to the patient, which is not limited in this application.
- the information carried in the medical misdiagnosis detection request includes, but is not limited to: an identification code, a preset label, the main complaint data, the disease to be diagnosed, and the like.
- the acquiring unit 110 acquiring the chief complaint data and the disease to be diagnosed from the medical misdiagnosis detection request includes:
- a second identifier is acquired, and information corresponding to the second identifier is acquired from the message information as the disease to be diagnosed.
- the first identifier and the second identifier are predefined identifiers, for example, the first identifier may be intru.
- the parsing efficiency of the medical misdiagnosis detection request can be improved.
- the mapping relationship between the first identification and the main complaint data, and the second identification The mapping relationship with the disease to be diagnosed can accurately determine the disease to be diagnosed in the main complaint data set.
- the determining unit 111 determines the user to be diagnosed according to the medical misdiagnosis detection request, and obtains the current medical history of the user to be diagnosed.
- the user to be diagnosed refers to any user currently treated by medical staff
- the current medical history refers to diseases that the user to be diagnosed has experienced in the past.
- the determining unit 111 determining the user to be diagnosed according to the medical misdiagnosis detection request, and obtaining the current medical history of the user to be diagnosed includes:
- the identification code is used to determine the user to be diagnosed, and information corresponding to the user to be diagnosed is obtained from the archive database as the current medical history.
- the medical history of multiple users corresponding to the user is stored in the archive library.
- the identification code is unique, the user to be diagnosed can be accurately determined through the identification code.
- the medical misdiagnosis detection request can be analyzed by directly obtaining idle threads from the thread connection pool. The creation time of the thread, therefore, improves the parsing speed of the medical misdiagnosis detection request, thereby improving the efficiency of the misdiagnosis detection.
- the extraction unit 112 extracts entities in the main complaint data and entities in the current medical history to obtain disease entities.
- the disease entity refers to a disease that may occur in the user to be diagnosed, for example, the disease entity may be a cough.
- the extraction unit 112 extracts entities in the main complaint data and entities in the current medical history, and obtaining disease entities includes:
- the entity in the main complaint data and the entity in the history of current illness are merged to obtain the disease entity.
- the obtaining unit 110 obtains the target entity associated with the disease entity from a pre-built graph neural network, and obtains the weight of the target entity.
- the graph neural network includes multiple entities, the attributes of each entity, and the degree of association between the entities and the attributes.
- the acquiring unit 110 before acquiring the target entity associated with the disease entity from a pre-built graph neural network, acquires the current disease, and acquires the target entity associated with the current disease Symptom attributes;
- the conversion unit 113 converts the current disease into a disease vector, and converts the symptom attribute into a symptom vector;
- the calculation unit 118 uses the attention mechanism to calculate the degree of association between the symptom vector and the disease vector;
- the entry unit 119 enters the current disease, the symptom attribute, and the degree of association into the graph template to obtain the graph neural network.
- the above embodiment since the graph template is constructed in advance, the above embodiment does not need to repeatedly create the graph template, and therefore, the determination efficiency of the graph neural network can be improved.
- the obtaining unit 110 obtaining the weight of the target entity includes:
- the conversion unit 113 converts the target entity into a medical knowledge feature vector based on the weight.
- the conversion unit 113 converting the target entity into a medical knowledge feature vector based on the weight includes:
- the processing unit 114 uses a convolutional neural network to process the main complaint data to obtain a text feature vector.
- the processing unit 114 using a convolutional neural network to process the main complaint data to obtain a text feature vector includes:
- each main complaint vector includes multiple dimensions
- the obtained vector values are spliced to obtain the text feature vector.
- the preset value can be self-defined configuration, for example, the preset value can be 2.
- the splicing unit 115 splices the medical knowledge feature vector and the text feature vector to obtain a target vector.
- the target vector is generated by concatenating the medical knowledge feature vector and the text feature vector.
- the input unit 116 inputs the target vector into the discriminant model to obtain a list of diseases.
- the list of diseases includes multiple predicted diseases with high probability.
- the obtaining unit 110 before inputting the target vector into the discriminant model to obtain the disease list, obtains a plurality of training data, and each training data includes symptom data of each training user , Medical history and corresponding symptoms of each training user;
- the conversion unit 113 converts the symptom data into a symptom vector, converts the medical history into a medical history vector, and converts the corresponding symptom into a symptom vector;
- the splicing unit 115 splices the symptom vector and the medical history vector to obtain a spliced vector
- the construction unit 120 constructs the discriminant model according to the splicing vector and the disease vector.
- the input unit 116 inputting the target vector into a discriminant model to obtain a disease list includes:
- the top N predicted diseases are selected from the queue, and the selected predicted diseases are merged to obtain the list of diseases, where N is a positive integer.
- the detection unit 117 detects whether the disease to be diagnosed exists in the disease list.
- the detection unit 117 traverses the disease list, and when the disease to be diagnosed is the same as the predicted disease traversed in the disease list, it is determined that the disease to be diagnosed exists in In the disease list, or when the disease to be diagnosed is different from all predicted diseases traversed in the disease list, it is determined that the disease to be diagnosed does not exist in the disease list.
- the response result of the medical misdiagnosis detection request is determined to be a correct diagnosis, reward information is generated, and the user who issues the disease to be diagnosed is determined, And send the reward information to the terminal device of the issuing user.
- the reward information can be sent to the issuing user, thereby improving the experience of the issuing user.
- the determining unit 111 determines the response result of the medical misdiagnosis detection request as a misdiagnosis.
- the above response result can also be stored in a node of a blockchain.
- the generating unit 121 after the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis, the generating unit 121 generates a misdiagnosis report according to the user to be diagnosed, the response result, and the disease list;
- the encryption unit 122 uses a symmetric encryption algorithm to encrypt the diagnosis report to obtain a ciphertext
- the determining unit 111 determines the terminal that issued the medical misdiagnosis detection request and the issuance time of the medical misdiagnosis detection request;
- the obtaining unit 110 obtains the log table of the terminal, and obtains the login account corresponding to the issuing time from the log table;
- the sending unit 123 sends the ciphertext to the login account
- the sending unit 123 issues an alarm reminder.
- the ciphertext can be generated and sent in time after the misdiagnosis is determined, so as to achieve the effect of timely warning.
- it can also send out an alarm in time without receiving feedback within the preset time, which can achieve real-time warning.
- the effect is to ensure that the ciphertext can be received in time, thereby avoiding damage.
- this application determines the user to be diagnosed according to the medical misdiagnosis detection request, can accurately determine the user to be diagnosed, extracts the entity in the main complaint data and the entity in the current medical history, and obtains the disease Entity, because the user’s medical history will affect the diagnosis result, the current medical history of the user to be diagnosed is considered when analyzing the disease entity, which can improve the accuracy of misdiagnosis detection.
- the target entity associated with the disease entity, and the weight of the target entity is obtained.
- the weight of the target entity can be accurately determined, and then The medical knowledge feature vector can be accurately generated, and then the target vector is input into the discriminant model to obtain a list of diseases. Because the target vector takes into account the main complaint data and the current medical history of the user to be diagnosed, the target vector can be accurately determined.
- the disease list when the disease to be diagnosed does not exist in the disease list, the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis. Since the disease list contains multiple predicted diseases, the misdiagnosis can be improved The detection rate, in addition, since this application is tested before the diagnosis report is issued to the user to be diagnosed, the effect of real-time warning can be achieved.
- FIG. 3 it is a schematic structural diagram of an electronic device in a preferred embodiment of the method for detecting medical misdiagnosis in this application.
- the electronic device 1 includes, but is not limited to, a memory 12, a processor 13, and a computer program stored in the memory 12 and running on the processor 13, such as Medical misdiagnosis detection procedures.
- the schematic diagram is only an example of the electronic device 1 and does not constitute a limitation on the electronic device 1. It may include more or less components than those shown in the figure, or a combination of certain components, or different components. Components, for example, the electronic device 1 may also include an input/output device, a network access device, a bus, and the like.
- the processor 13 may be a central processing unit (Central Processing Unit, CPU), other general-purpose processors, digital signal processors (Digital Signal Processor, DSP), application specific integrated circuits (ASIC), Field-Programmable Gate Array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc.
- the general-purpose processor can be a microprocessor or the processor can also be any conventional processor, etc.
- the processor 13 is the computing core and control center of the electronic device 1 and connects the entire electronic device with various interfaces and lines. Each part of 1, and executes the operating system of the electronic device 1, and various installed applications, program codes, etc.
- the processor 13 executes the operating system of the electronic device 1 and various installed applications.
- the processor 13 executes the application program to implement the steps in the foregoing embodiments of the medical misdiagnosis detection method, for example, the steps shown in FIG. 1.
- the computer program may be divided into one or more modules/units, and the one or more modules/units are stored in the memory 12 and executed by the processor 13 to complete this Application.
- the one or more modules/units may be a series of computer-readable instruction segments capable of completing specific functions, and the instruction segments are used to describe the execution process of the computer program in the electronic device 1.
- the computer program can be divided into an acquisition unit 110, a determination unit 111, an extraction unit 112, a conversion unit 113, a processing unit 114, a splicing unit 115, an input unit 116, a detection unit 117, a calculation unit 118, an entry unit 119,
- the construction unit 120, the generation unit 121, the encryption unit 122, and the transmission unit 123 are examples of the computer program.
- the memory 12 may be used to store the computer program and/or module.
- the processor 13 runs or executes the computer program and/or module stored in the memory 12 and calls data stored in the memory 12, The various functions of the electronic device 1 are realized.
- the memory 12 may mainly include a storage program area and a storage data area.
- the storage program area may store an operating system, an application program required by at least one function (such as a sound playback function, an image playback function, etc.), etc.; the storage data area may Stores data, etc. created based on the use of electronic devices.
- the memory 12 may include non-volatile and volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a Secure Digital (SD) card, and a flash memory card ( Flash Card), at least one magnetic disk storage device, flash memory device, or other storage device.
- non-volatile and volatile memory such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a Secure Digital (SD) card, and a flash memory card ( Flash Card), at least one magnetic disk storage device, flash memory device, or other storage device.
- the memory 12 may be an external memory and/or an internal memory of the electronic device 1. Further, the memory 12 may be a memory in a physical form, such as a memory stick, a TF card (Trans-flash Card), and so on.
- TF card Trans-flash Card
- the integrated module/unit of the electronic device 1 may be stored in a computer-readable storage medium, which may be non-easy.
- a volatile storage medium can also be a volatile storage medium.
- the computer program includes computer-readable instruction code
- the computer-readable instruction code may be in the form of source code, object code, executable file, or some intermediate form.
- the computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory) , Random access memory.
- the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
- Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
- the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.
- the memory 12 in the electronic device 1 stores multiple instructions to implement a medical misdiagnosis detection method, and the processor 13 can execute the multiple instructions to achieve:
- the response result of the medical misdiagnosis detection request is determined to be a misdiagnosis.
- modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the objectives of the solutions of the embodiments.
- the functional modules in the various embodiments of the present application may be integrated into one processing unit, or each unit may exist alone physically, or two or more units may be integrated into one unit.
- the above-mentioned integrated unit may be implemented in the form of hardware, or may be implemented in the form of hardware plus software functional modules.
Landscapes
- Engineering & Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Biomedical Technology (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Computing Systems (AREA)
- Public Health (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Molecular Biology (AREA)
- Medical Informatics (AREA)
- Pathology (AREA)
- Databases & Information Systems (AREA)
- Epidemiology (AREA)
- Primary Health Care (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Measuring And Recording Apparatus For Diagnosis (AREA)
- Medical Treatment And Welfare Office Work (AREA)
Abstract
一种医疗误诊检测方法、装置、电子设备及存储介质,该方法从医疗误诊检测请求中获取主诉数据及待诊断疾病(S10);确定待诊断用户,获取待诊断用户的现病史(S11);抽取主诉数据及现病史中的实体,得到疾病实体(S12);获取与疾病实体关联的目标实体,获取目标实体的权重(S13);基于权重将目标实体转换为医学知识特征向量(S14);处理主诉数据,得到文本特征向量(S15);拼接医学知识特征向量及文本特征向量,得到目标向量(S16);将目标向量输入至判别模型中,得到疾病列表(S17);当待诊断疾病不存在于疾病列表中时,将医疗误诊检测请求的响应结果确定为误诊(S19)。该方法能够提高误诊检测率及实时预警。
Description
本申请要求于2020年07月28日提交中国专利局,申请号为202010739726.2,发明名称为“医疗误诊检测方法、装置、电子设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
本申请涉及人工智能技术领域,尤其涉及一种医疗误诊检测方法、装置、电子设备及存储介质。
由于我国人口基数巨大,导致医院的就诊量也很大,这就要求医生在极短时间内根据患者的主诉数据及相关经验做出诊断,由于全身诊断自身的难度及时间限制,导致医生的误诊率较高。为了降低误诊率,现有的误诊检测系统也应运而生。然而,发明人意识到现有的误诊检测系统往往是事后检测,达不到实时预警的效果。
发明内容
鉴于以上内容,有必要提出一种医疗误诊检测方法、装置、电子设备及存储介质,不仅能够提高误诊检测率,还能够达到实时预警的效果。
本申请的第一方面提供一种医疗误诊检测方法,所述医疗误诊检测方法包括:
当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;
根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;
抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;
从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;
基于所述权重将所述目标实体转换为医学知识特征向量;
利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;
拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;
将所述目标向量输入至判别模型中,得到疾病列表;
检测所述待诊断疾病是否存在于所述疾病列表中;
当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
本申请的第二方面提供一种电子设备,所述电子设备包括处理器和存储器,所述处理器用于执行所述存储器中存储的计算机可读指令以实现以下步骤:
当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;
根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;
抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;
从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;
基于所述权重将所述目标实体转换为医学知识特征向量;
利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;
拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;
将所述目标向量输入至判别模型中,得到疾病列表;
检测所述待诊断疾病是否存在于所述疾病列表中;
当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
本申请的第三方面提供一种计算机可读存储介质,所述计算机可读存储介质上存储有至少一个计算机可读指令,所述至少一个计算机可读指令被处理器执行以实现以下步骤:
当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;
根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;
抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;
从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;
基于所述权重将所述目标实体转换为医学知识特征向量;
利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;
拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;
将所述目标向量输入至判别模型中,得到疾病列表;
检测所述待诊断疾病是否存在于所述疾病列表中;
当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
本申请的第四方面提供一种医疗误诊检测装置,所述医疗误诊检测装置包括:
获取单元,用于当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;
确定单元,用于根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;
抽取单元,用于抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;
所述获取单元,还用于从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;
转换单元,用于基于所述权重将所述目标实体转换为医学知识特征向量;
处理单元,用于利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;
拼接单元,用于拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;
输入单元,用于将所述目标向量输入至判别模型中,得到疾病列表;
检测单元,用于检测所述待诊断疾病是否存在于所述疾病列表中;
所述确定单元,还用于当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
由以上技术方案可以看出,本申请根据所述医疗误诊检测请求确定待诊断用户,能够准确确定所述待诊断用户,抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体,由于用户的病史会影响诊断结果,因此,在分析疾病实体时考虑了所述待诊断用户的现病史,能够提高误诊检测的准确率,通过从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重,由于权重并不是任意设定的,而是从预先构建的图神经网络中获取的,因此能够准确确定所述目标实体的权重,进而能够准确生成医学知识特征向量,进而将所述目标向量输入至判别模型中,得到疾病列表,由于所述目标向量中考虑了主诉数据、待诊断用户的现病史,因此,能够准确确定出所述疾病列表,当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检 测请求的响应结果确定为误诊,由于所述疾病列表中包含了多个预测疾病,因此能够提高误诊检测率,此外,由于本申请是在给待诊断用户下达诊断报告前检测的,因此能够达到实时预警的效果。本申请应用于智慧城市和智慧医疗中,从而推动智慧城市的建设。
图1是本申请医疗误诊检测方法的较佳实施例的流程图。
图2是本申请医疗误诊检测装置的较佳实施例的功能模块图。
图3是本申请实现医疗误诊检测方法的较佳实施例的电子设备的结构示意图。
为了使本申请的目的、技术方案和优点更加清楚,下面结合附图和具体实施例对本申请进行详细描述。
如图1所示,是本申请医疗误诊检测方法的较佳实施例的流程图。根据不同的需求,该流程图中步骤的顺序可以改变,某些步骤可以省略。
所述医疗误诊检测方法应用于智慧城市,从而推动智慧城市的建设。所述医疗误诊检测方法应用于一个或者多个电子设备中,所述电子设备是一种能够按照事先设定或存储的指令,自动进行数值计算和/或信息处理的设备,其硬件包括但不限于微处理器、专用集成电路(Application Specific Integrated Circuit,ASIC)、可编程门阵列(Field-Programmable Gate Array,FPGA)、数字处理器(Digital Signal Processor,DSP)、嵌入式设备等。
所述电子设备可以是任何一种可与用户进行人机交互的电子产品,例如,个人计算机、平板电脑、智能手机、个人数字助理(Personal Digital Assistant,PDA)、游戏机、交互式网络电视(Internet Protocol Television,IPTV)、智能式穿戴式设备等。
所述电子设备还可以包括网络设备和/或用户设备。其中,所述网络设备包括,但不限于单个网络服务器、多个网络服务器组成的服务器组或基于云计算(Cloud Computing)的由大量主机或网络服务器构成的云。
所述电子设备所处的网络包括但不限于互联网、广域网、城域网、局域网、虚拟专用网络(Virtual Private Network,VPN)等。
在本申请的至少一个实施例中,所述医疗误诊检测方法应用于人工智能领域中。
S10,当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病。
在本申请的至少一个实施例中,所述医疗误诊检测请求可以由医护人员触发,也可以在检测到所述医护人员给病人下达诊断书之前触发,本申请对此不作限制。
进一步地,所述医疗误诊检测请求携带的信息包括,但不限于:身份识别码、预设标签、所述主诉数据、所述待诊断疾病等。
在本申请的至少一个实施例中,所述电子设备从所述医疗误诊检测请求中获取主诉数据及待诊断疾病包括:
解析所述医疗误诊检测请求的报文,得到所述医疗误诊检测请求携带的报文信息;
获取第一标识,并从所述报文信息中获取与所述第一标识对应的信息,作为所述主诉数据;
获取第二标识,并从所述报文信息中获取与所述第二标识对应的信息,作为所述待诊断疾病。
其中,所述第一标识及所述第二标识为预先定义好的标识,例如,所述第一标识可以是intru。
通过上述实施方式,由于无需解析所述医疗误诊检测请求的报文头,因此,能够提高所述医疗误诊检测请求的解析效率,此外,通过第一标识与主诉数据的映射关系,及第二标识 与待诊断疾病的映射关系,能够准确确定所述主诉数据集所述待诊断疾病。
S11,根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史。
在本申请的至少一个实施例中,所述待诊断用户是指医护人员当前医治的任意用户,所述现病史是指所述待诊断用户在过去发生过的疾病。
在本申请的至少一个实施例中,所述电子设备根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史包括:
从线程连接池中获取任意闲置线程;
利用所述任意闲置线程解析所述医疗误诊检测请求的方法体,得到所述医疗误诊检测请求携带的报文信息;
获取预设标签,并从所述报文信息中获取与所述预设标签对应的信息,作为身份识别码;
利用所述身份识别码确定所述待诊断用户,并从建档库中获取与所述待诊断用户对应的信息,作为所述现病史。
其中,所述建档库中存储多个用户与所述用户对应的病史。
由于所述身份识别码具有唯一性,因此,通过所述身份识别码能够准确确定所述待诊断用户,另外,通过从线程连接池中直接获取闲置线程解析所述医疗误诊检测请求,由于节省了线程的创建时间,因此,提高了所述医疗误诊检测请求的解析速度,进而提高误诊检测的效率。
S12,抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体。
在本申请的至少一个实施例中,所述疾病实体是指所述待诊断用户可能发生的疾病,例如,所述疾病实体可以是咳嗽。
在本申请的至少一个实施例中,所述电子设备抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体包括:
对所述主诉数据进行分词处理,得到第一分词,并对所述现病史进行分词处理,得到第二分词;
遍历预先构建的词典,并将遍历到的第一分词确定为所述主诉数据中的实体,将遍历到的第二分词确定为所述现病史中的实体;
融合所述主诉数据中的实体及所述现病史中的实体,得到所述疾病实体。
其中,所述词典中存储多个病原实体。
S13,从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重。
在本申请的至少一个实施例中,所述图神经网络中包括多个实体、每个实体的属性、实体与属性的关联度。
在本申请的至少一个实施例中,在从预先构建的图神经网络中获取与所述疾病实体关联的目标实体之前,所述医疗误诊检测方法还包括:
获取当前疾病,并获取与所述当前疾病相关联的症状属性;
将所述当前疾病转换为疾病向量,并将所述症状属性转换为症状向量;
利用注意力机制计算所述症状向量与所述疾病向量的关联度;
将所述当前疾病、所述症状属性、所述关联度录入至图模板中,得到所述图神经网络。
通过上述实施方式,由于图模板是预先构建的,因此,上述实施方式无需重复创建图模板,因此,能够提高所述图神经网络的确定效率。
在本申请的至少一个实施例中,所述获取所述目标实体的权重包括:
从所述图神经网络中获取所述疾病实体与所述目标实体的关联度,作为目标关联度;
对所述目标关联度进行归一化处理,得到所述权重。
S14,基于所述权重将所述目标实体转换为医学知识特征向量。
在本申请的至少一个实施例中,所述电子设备基于所述权重将所述目标实体转换为医学知识特征向量包括:
获取所述目标实体的向量值,得到实体向量;
基于所述权重对所述实体向量进行加权和运算,得到所述医学知识特征向量。
S15,利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量。
在本申请的至少一个实施例中,所述电子设备利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量包括:
根据预设数值对所述主诉数据进行分词,得到多个主诉分词;
对所述多个主诉分词进行向量化处理,得到多个主诉向量,每个主诉向量包括多个维度;
确定每个主诉向量中向量值最大的维度,得到目标维度,并获取所述目标维度对应的向量值;
拼接获取的向量值,得到所述文本特征向量。
其中,所述预设数值可以自定义配置,例如,所述预设数值可以是2。
通过上述实施方式,能够减少所述主诉数据的丢失。
S16,拼接所述医学知识特征向量及所述文本特征向量,得到目标向量。
在本申请的至少一个实施例中,所述目标向量是由所述医学知识特征向量及所述文本特征向量拼接而生成的。
S17,将所述目标向量输入至判别模型中,得到疾病列表。
在本申请的至少一个实施例中,所述疾病列表中包括多种概率较大的预测疾病。
在本申请的至少一个实施例中,在将所述目标向量输入至判别模型中,得到疾病列表之前,所述方法还包括:
获取多个训练数据,每个训练数据包括每个训练用户的症状数据、每个训练用户的病史及对应病症;
将所述症状数据转换为症状向量,将所述病史转换为病史向量,并将所述对应病症转换为病症向量;
拼接所述症状向量与所述病史向量,得到拼接向量;
依据所述拼接向量及所述病症向量构建所述判别模型。
通过上述实施方式,由于训练数据源自于真实数据,因此,能够生成准确的判别模型,有利于后续疾病列表的确定。
在本申请的至少一个实施例中,所述将所述目标向量输入至判别模型中,得到疾病列表包括:
将所述目标向量输入至所述判别模型,得到多种预测疾病及每种预测疾病的概率;
依照所述概率从大至小的顺序对所述多种预测疾病进行排序,得到队列;
从所述队列中选取前N个预测疾病,并融合选取到的预测疾病,得到所述疾病列表,其中,N为正整数。
通过融合概率较大的预测疾病,能够得到准确的疾病列表。
S18,检测所述待诊断疾病是否存在于所述疾病列表中。
在本申请的至少一个实施例中,所述电子设备遍历所述疾病列表,当所述待诊断疾病与在所述疾病列表中遍历到的预测疾病相同时,确定所述待诊断疾病存在于所述疾病列表中,或者,当所述待诊断疾病与在所述疾病列表中遍历到的所有预测疾病都不相同时,确定所述待诊断疾病不存在于所述疾病列表中。
在其他实施例中,当所述待诊断疾病存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为正确诊断,生成奖励信息,确定所述待诊断疾病的发出用户,并向所述发出用户的终端设备发送所述奖励信息。
通过上述实施方式,能够在所述发出用户做出准确的诊断时,向所述发出用户发送所述奖励信息,从而提高所述发出用户的体验。
S19,当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
需要强调的是,为进一步保证上述响应结果的私密和安全性,上述响应结果还可以存储于一区块链的节点中。
在本申请的至少一个实施例中,在将所述医疗误诊检测请求的响应结果确定为误诊之后,所述医疗误诊检测方法还包括:
根据所述待诊断用户、所述响应结果及所述疾病列表生成误诊报告;
采用对称加密算法对所述诊断报告进行加密处理,得到密文;
确定发出所述医疗误诊检测请求的终端及所述医疗误诊检测请求的发出时间;
获取所述终端的日志表,并从所述日志表中获取与所述发出时间对应的登录账户;
将所述密文发送至所述登录账户上;
当在预设时间内未接收到所述登录账户的反馈信息时,发出警报提醒。
通过上述实施方式,能够在确定误诊后,及时生成并发送所述密文,以达到及时预警的效果,此外,还能在预设时间内未接收到反馈,及时发出警报,能够达到实时预警的效果,确保所述密文能够被及时接收,从而避免造成损害。
由以上技术方案可以看出,本申请根据所述医疗误诊检测请求确定待诊断用户,能够准确确定所述待诊断用户,抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体,由于用户的病史会影响诊断结果,因此,在分析疾病实体时考虑了所述待诊断用户的现病史,能够提高误诊检测的准确率,通过从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重,由于权重并不是任意设定的,而是从预先构建的图神经网络中获取的,因此能够准确确定所述目标实体的权重,进而能够准确生成医学知识特征向量,进而将所述目标向量输入至判别模型中,得到疾病列表,由于所述目标向量中考虑了主诉数据、待诊断用户的现病史,因此,能够准确确定出所述疾病列表,当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊,由于所述疾病列表中包含了多个预测疾病,因此能够提高误诊检测率,此外,由于本申请是在给待诊断用户下达诊断报告前检测的,因此能够达到实时预警的效果。
如图2所示,是本申请医疗误诊检测装置的较佳实施例的功能模块图。所述医疗误诊检测装置11包括获取单元110、确定单元111、抽取单元112、转换单元113、处理单元114、拼接单元115、输入单元116、检测单元117、计算单元118、录入单元119、构建单元120、生成单元121、加密单元122及发送单元123。本申请所称的模块/单元是指一种能够被处理器13所执行,并且能够完成固定功能的一系列计算机程序段,其存储在存储器12中。在本实施例中,关于各模块/单元的功能将在后续的实施例中详述。
当接收到医疗误诊检测请求时,获取单元110从所述医疗误诊检测请求中获取主诉数据及待诊断疾病。
在本申请的至少一个实施例中,所述医疗误诊检测请求可以由医护人员触发,也可以在检测到所述医护人员给病人下达诊断书之前触发,本申请对此不作限制。
进一步地,所述医疗误诊检测请求携带的信息包括,但不限于:身份识别码、预设标签、所述主诉数据、所述待诊断疾病等。
在本申请的至少一个实施例中,所述获取单元110从所述医疗误诊检测请求中获取主诉数据及待诊断疾病包括:
解析所述医疗误诊检测请求的报文,得到所述医疗误诊检测请求携带的报文信息;
获取第一标识,并从所述报文信息中获取与所述第一标识对应的信息,作为所述主诉数 据;
获取第二标识,并从所述报文信息中获取与所述第二标识对应的信息,作为所述待诊断疾病。
其中,所述第一标识及所述第二标识为预先定义好的标识,例如,所述第一标识可以是intru。
通过上述实施方式,由于无需解析所述医疗误诊检测请求的报文头,因此,能够提高所述医疗误诊检测请求的解析效率,此外,通过第一标识与主诉数据的映射关系,及第二标识与待诊断疾病的映射关系,能够准确确定所述主诉数据集所述待诊断疾病。
确定单元111根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史。
在本申请的至少一个实施例中,所述待诊断用户是指医护人员当前医治的任意用户,所述现病史是指所述待诊断用户在过去发生过的疾病。
在本申请的至少一个实施例中,所述确定单元111根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史包括:
从线程连接池中获取任意闲置线程;
利用所述任意闲置线程解析所述医疗误诊检测请求的方法体,得到所述医疗误诊检测请求携带的报文信息;
获取预设标签,并从所述报文信息中获取与所述预设标签对应的信息,作为身份识别码;
利用所述身份识别码确定所述待诊断用户,并从建档库中获取与所述待诊断用户对应的信息,作为所述现病史。
其中,所述建档库中存储多个用户与所述用户对应的病史。
由于所述身份识别码具有唯一性,因此,通过所述身份识别码能够准确确定所述待诊断用户,另外,通过从线程连接池中直接获取闲置线程解析所述医疗误诊检测请求,由于节省了线程的创建时间,因此,提高了所述医疗误诊检测请求的解析速度,进而提高误诊检测的效率。
抽取单元112抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体。
在本申请的至少一个实施例中,所述疾病实体是指所述待诊断用户可能发生的疾病,例如,所述疾病实体可以是咳嗽。
在本申请的至少一个实施例中,所述抽取单元112抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体包括:
对所述主诉数据进行分词处理,得到第一分词,并对所述现病史进行分词处理,得到第二分词;
遍历预先构建的词典,并将遍历到的第一分词确定为所述主诉数据中的实体,将遍历到的第二分词确定为所述现病史中的实体;
融合所述主诉数据中的实体及所述现病史中的实体,得到所述疾病实体。
其中,所述词典中存储多个病原实体。
所述获取单元110从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重。
在本申请的至少一个实施例中,所述图神经网络中包括多个实体、每个实体的属性、实体与属性的关联度。
在本申请的至少一个实施例中,在从预先构建的图神经网络中获取与所述疾病实体关联的目标实体之前,所述获取单元110获取当前疾病,并获取与所述当前疾病相关联的症状属性;
转换单元113将所述当前疾病转换为疾病向量,并将所述症状属性转换为症状向量;
计算单元118利用注意力机制计算所述症状向量与所述疾病向量的关联度;
录入单元119将所述当前疾病、所述症状属性、所述关联度录入至图模板中,得到所述图神经网络。
通过上述实施方式,由于图模板是预先构建的,因此,上述实施方式无需重复创建图模板,因此,能够提高所述图神经网络的确定效率。
在本申请的至少一个实施例中,所述获取单元110获取所述目标实体的权重包括:
从所述图神经网络中获取所述疾病实体与所述目标实体的关联度,作为目标关联度;
对所述目标关联度进行归一化处理,得到所述权重。
所述转换单元113基于所述权重将所述目标实体转换为医学知识特征向量。
在本申请的至少一个实施例中,所述转换单元113基于所述权重将所述目标实体转换为医学知识特征向量包括:
获取所述目标实体的向量值,得到实体向量;
基于所述权重对所述实体向量进行加权和运算,得到所述医学知识特征向量。
处理单元114利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量。
在本申请的至少一个实施例中,所述处理单元114利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量包括:
根据预设数值对所述主诉数据进行分词,得到多个主诉分词;
对所述多个主诉分词进行向量化处理,得到多个主诉向量,每个主诉向量包括多个维度;
确定每个主诉向量中向量值最大的维度,得到目标维度,并获取所述目标维度对应的向量值;
拼接获取的向量值,得到所述文本特征向量。
其中,所述预设数值可以自定义配置,例如,所述预设数值可以是2。
通过上述实施方式,能够减少所述主诉数据的丢失。
拼接单元115拼接所述医学知识特征向量及所述文本特征向量,得到目标向量。
在本申请的至少一个实施例中,所述目标向量是由所述医学知识特征向量及所述文本特征向量拼接而生成的。
输入单元116将所述目标向量输入至判别模型中,得到疾病列表。
在本申请的至少一个实施例中,所述疾病列表中包括多种概率较大的预测疾病。
在本申请的至少一个实施例中,在将所述目标向量输入至判别模型中,得到疾病列表之前,所述获取单元110获取多个训练数据,每个训练数据包括每个训练用户的症状数据、每个训练用户的病史及对应病症;
所述转换单元113将所述症状数据转换为症状向量,将所述病史转换为病史向量,并将所述对应病症转换为病症向量;
所述拼接单元115拼接所述症状向量与所述病史向量,得到拼接向量;
构建单元120依据所述拼接向量及所述病症向量构建所述判别模型。
通过上述实施方式,由于训练数据源自于真实数据,因此,能够生成准确的判别模型,有利于后续疾病列表的确定。
在本申请的至少一个实施例中,所述输入单元116将所述目标向量输入至判别模型中,得到疾病列表包括:
将所述目标向量输入至所述判别模型,得到多种预测疾病及每种预测疾病的概率;
依照所述概率从大至小的顺序对所述多种预测疾病进行排序,得到队列;
从所述队列中选取前N个预测疾病,并融合选取到的预测疾病,得到所述疾病列表,其中,N为正整数。
通过融合概率较大的预测疾病,能够得到准确的疾病列表。
检测单元117检测所述待诊断疾病是否存在于所述疾病列表中。
在本申请的至少一个实施例中,所述检测单元117遍历所述疾病列表,当所述待诊断疾病与在所述疾病列表中遍历到的预测疾病相同时,确定所述待诊断疾病存在于所述疾病列表中,或者,当所述待诊断疾病与在所述疾病列表中遍历到的所有预测疾病都不相同时,确定所述待诊断疾病不存在于所述疾病列表中。
在其他实施例中,当所述待诊断疾病存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为正确诊断,生成奖励信息,确定所述待诊断疾病的发出用户,并向所述发出用户的终端设备发送所述奖励信息。
通过上述实施方式,能够在所述发出用户做出准确的诊断时,向所述发出用户发送所述奖励信息,从而提高所述发出用户的体验。
当所述待诊断疾病不存在于所述疾病列表中时,所述确定单元111将所述医疗误诊检测请求的响应结果确定为误诊。
需要强调的是,为进一步保证上述响应结果的私密和安全性,上述响应结果还可以存储于一区块链的节点中。
在本申请的至少一个实施例中,在将所述医疗误诊检测请求的响应结果确定为误诊之后,生成单元121根据所述待诊断用户、所述响应结果及所述疾病列表生成误诊报告;
加密单元122采用对称加密算法对所述诊断报告进行加密处理,得到密文;
所述确定单元111确定发出所述医疗误诊检测请求的终端及所述医疗误诊检测请求的发出时间;
所述获取单元110获取所述终端的日志表,并从所述日志表中获取与所述发出时间对应的登录账户;
发送单元123将所述密文发送至所述登录账户上;
当在预设时间内未接收到所述登录账户的反馈信息时,所述发送单元123发出警报提醒。
通过上述实施方式,能够在确定误诊后,及时生成并发送所述密文,以达到及时预警的效果,此外,还能在预设时间内未接收到反馈,及时发出警报,能够达到实时预警的效果,确保所述密文能够被及时接收,从而避免造成损害。
由以上技术方案可以看出,本申请根据所述医疗误诊检测请求确定待诊断用户,能够准确确定所述待诊断用户,抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体,由于用户的病史会影响诊断结果,因此,在分析疾病实体时考虑了所述待诊断用户的现病史,能够提高误诊检测的准确率,通过从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重,由于权重并不是任意设定的,而是从预先构建的图神经网络中获取的,因此能够准确确定所述目标实体的权重,进而能够准确生成医学知识特征向量,进而将所述目标向量输入至判别模型中,得到疾病列表,由于所述目标向量中考虑了主诉数据、待诊断用户的现病史,因此,能够准确确定出所述疾病列表,当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊,由于所述疾病列表中包含了多个预测疾病,因此能够提高误诊检测率,此外,由于本申请是在给待诊断用户下达诊断报告前检测的,因此能够达到实时预警的效果。
如图3所示,是本申请实现医疗误诊检测方法的较佳实施例的电子设备的结构示意图。
在本申请的一个实施例中,所述电子设备1包括,但不限于,存储器12、处理器13,以及存储在所述存储器12中并可在所述处理器13上运行的计算机程序,例如医疗误诊检测程序。
本领域技术人员可以理解,所述示意图仅仅是电子设备1的示例,并不构成对电子设备1的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件,例如所述电子设备1还可以包括输入输出设备、网络接入设备、总线等。
所述处理器13可以是中央处理单元(Central Processing Unit,CPU),还可以是其他通用处理器、数字信号处理器(Digital Signal Processor,DSP)、专用集成电路(Application Specific Integrated Circuit,ASIC)、现场可编程门阵列(Field-Programmable Gate Array,FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等,所述处理器13是所述电子设备1的运算核心和控制中心,利用各种接口和线路连接整个电子设备1的各个部分,及执行所述电子设备1的操作系统以及安装的各类应用程序、程序代码等。
所述处理器13执行所述电子设备1的操作系统以及安装的各类应用程序。所述处理器13执行所述应用程序以实现上述各个医疗误诊检测方法实施例中的步骤,例如图1所示的步骤。
示例性的,所述计算机程序可以被分割成一个或多个模块/单元,所述一个或者多个模块/单元被存储在所述存储器12中,并由所述处理器13执行,以完成本申请。所述一个或多个模块/单元可以是能够完成特定功能的一系列计算机可读指令段,该指令段用于描述所述计算机程序在所述电子设备1中的执行过程。例如,所述计算机程序可以被分割成获取单元110、确定单元111、抽取单元112、转换单元113、处理单元114、拼接单元115、输入单元116、检测单元117、计算单元118、录入单元119、构建单元120、生成单元121、加密单元122及发送单元123。
所述存储器12可用于存储所述计算机程序和/或模块,所述处理器13通过运行或执行存储在所述存储器12内的计算机程序和/或模块,以及调用存储在存储器12内的数据,实现所述电子设备1的各种功能。所述存储器12可主要包括存储程序区和存储数据区,其中,存储程序区可存储操作系统、至少一个功能所需的应用程序(比如声音播放功能、图像播放功能等)等;存储数据区可存储根据电子设备的使用所创建的数据等。此外,存储器12可以包括非易失性和易失性存储器,例如硬盘、内存、插接式硬盘,智能存储卡(Smart Media Card,SMC),安全数字(Secure Digital,SD)卡,闪存卡(Flash Card)、至少一个磁盘存储器件、闪存器件、或其他存储器件。
所述存储器12可以是电子设备1的外部存储器和/或内部存储器。进一步地,所述存储器12可以是具有实物形式的存储器,如内存条、TF卡(Trans-flash Card)等等。
所述电子设备1集成的模块/单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中,所述计算机可读存储介质可以是非易失性的存储介质,也可以是易失的存储介质。基于这样的理解,本申请实现上述实施例方法中的全部或部分流程,也可以通过计算机程序来指令相关的硬件来完成,所述的计算机程序可存储于一计算机可读存储介质中,该计算机程序在被处理器执行时,可实现上述各个方法实施例的步骤。
其中,所述计算机程序包括计算机可读指令代码,所述计算机可读指令代码可以为源代码形式、对象代码形式、可执行文件或某些中间形式等。所述计算机可读介质可以包括:能够携带所述计算机程序代码的任何实体或装置、记录介质、U盘、移动硬盘、磁碟、光盘、计算机存储器、只读存储器(ROM,Read-Only Memory)、随机存取存储器。
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
结合图1,所述电子设备1中的所述存储器12存储多个指令以实现一种医疗误诊检测方法,所述处理器13可执行所述多个指令从而实现:
当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;
根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;
抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;
从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;
基于所述权重将所述目标实体转换为医学知识特征向量;
利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;
拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;
将所述目标向量输入至判别模型中,得到疾病列表;
检测所述待诊断疾病是否存在于所述疾病列表中;
当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
具体地,所述处理器13对上述指令的具体实现方法可参考图1对应实施例中相关步骤的描述,在此不赘述。
在本申请所提供的几个实施例中,应该理解到,所揭露的系统,装置和方法,可以通过其它的方式实现。例如,以上所描述的装置实施例仅仅是示意性的,例如,所述模块的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式。
所述作为分离部件说明的模块可以是或者也可以不是物理上分开的,作为模块显示的部件可以是或者也可以不是物理单元,既可以位于一个地方,或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。
另外,在本申请各个实施例中的各功能模块可以集成在一个处理单元中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现,也可以采用硬件加软件功能模块的形式实现。
因此,无论从哪一点来看,均应将实施例看作是示范性的,而且是非限制性的,本申请的范围由所附权利要求而不是上述说明限定,因此旨在将落在权利要求的等同要件的含义和范围内的所有变化涵括在本申请内。不应将权利要求中的任何附关联图标记视为限制所涉及的权利要求。
此外,显然“包括”一词不排除其他单元或步骤,单数不排除复数。系统权利要求中陈述的多个单元或装置也可以由一个单元或装置通过软件或者硬件来实现。第二等词语用来表示名称,而并不表示任何特定的顺序。
最后应说明的是,以上实施例仅用以说明本申请的技术方案而非限制,尽管参照较佳实施例对本申请进行了详细说明,本领域的普通技术人员应当理解,可以对本申请的技术方案进行修改或等同替换,而不脱离本申请技术方案的精神和范围。
Claims (20)
- 一种医疗误诊检测方法,其中,所述医疗误诊检测方法包括:当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;基于所述权重将所述目标实体转换为医学知识特征向量;利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;将所述目标向量输入至判别模型中,得到疾病列表;检测所述待诊断疾病是否存在于所述疾病列表中;当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
- 根据权利要求1所述的医疗误诊检测方法,其中,所述根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史包括:从线程连接池中获取任意闲置线程;利用所述任意闲置线程解析所述医疗误诊检测请求的方法体,得到所述医疗误诊检测请求携带的报文信息;获取预设标签,并从所述报文信息中获取与所述预设标签对应的信息,作为身份识别码;利用所述身份识别码确定所述待诊断用户,并从建档库中获取与所述待诊断用户对应的信息,作为所述现病史。
- 根据权利要求1所述的医疗误诊检测方法,其中,所述抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体包括:对所述主诉数据进行分词处理,得到第一分词,并对所述现病史进行分词处理,得到第二分词;遍历预先构建的词典,并将遍历到的第一分词确定为所述主诉数据中的实体,将遍历到的第二分词确定为所述现病史中的实体;融合所述主诉数据中的实体及所述现病史中的实体,得到所述疾病实体。
- 根据权利要求1所述的医疗误诊检测方法,其中,在从预先构建的图神经网络中获取与所述疾病实体关联的目标实体之前,所述医疗误诊检测方法还包括:获取当前疾病,并获取与所述当前疾病相关联的症状属性;将所述当前疾病转换为疾病向量,并将所述症状属性转换为症状向量;利用注意力机制计算所述症状向量与所述疾病向量的关联度;将所述当前疾病、所述症状属性、所述关联度录入至图模板中,得到所述图神经网络。
- 根据权利要求1所述的医疗误诊检测方法,其中,所述基于所述权重将所述目标实体转换为医学知识特征向量包括:获取所述目标实体的向量值,得到实体向量;基于所述权重对所述实体向量进行加权和运算,得到所述医学知识特征向量。
- 根据权利要求1所述的医疗误诊检测方法,其中,所述利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量包括:根据预设数值对所述主诉数据进行分词,得到多个主诉分词;对所述多个主诉分词进行向量化处理,得到多个主诉向量,每个主诉向量包括多个维度;确定每个主诉向量中向量值最大的维度,得到目标维度,并获取所述目标维度对应的向量值;拼接获取的向量值,得到所述文本特征向量。
- 根据权利要求1所述的医疗误诊检测方法,其中,在将所述医疗误诊检测请求的响应结果确定为误诊之后,所述医疗误诊检测方法还包括:根据所述待诊断用户、所述响应结果及所述疾病列表生成误诊报告;采用对称加密算法对所述诊断报告进行加密处理,得到密文;确定发出所述医疗误诊检测请求的终端及所述医疗误诊检测请求的发出时间;获取所述终端的日志表,并从所述日志表中获取与所述发出时间对应的登录账户;将所述密文发送至所述登录账户上;当在预设时间内未接收到所述登录账户的反馈信息时,发出警报提醒。
- 一种电子设备,其中,所述电子设备包括处理器和存储器,所述处理器用于执行存储器中存储的至少一个计算机可读指令以实现以下步骤:当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;基于所述权重将所述目标实体转换为医学知识特征向量;利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;将所述目标向量输入至判别模型中,得到疾病列表;检测所述待诊断疾病是否存在于所述疾病列表中;当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
- 根据权利要求8所述的电子设备,其中,在所述根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史时,所述处理器执行所述至少一个计算机可读指令以实现以下步骤:从线程连接池中获取任意闲置线程;利用所述任意闲置线程解析所述医疗误诊检测请求的方法体,得到所述医疗误诊检测请求携带的报文信息;获取预设标签,并从所述报文信息中获取与所述预设标签对应的信息,作为身份识别码;利用所述身份识别码确定所述待诊断用户,并从建档库中获取与所述待诊断用户对应的信息,作为所述现病史。
- 根据权利要求8所述的电子设备,其中,在所述抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体时,所述处理器执行所述至少一个计算机可读指令以实现以下步骤:对所述主诉数据进行分词处理,得到第一分词,并对所述现病史进行分词处理,得到第二分词;遍历预先构建的词典,并将遍历到的第一分词确定为所述主诉数据中的实体,将遍历到的第二分词确定为所述现病史中的实体;融合所述主诉数据中的实体及所述现病史中的实体,得到所述疾病实体。
- 根据权利要求8所述的电子设备,其中,在从预先构建的图神经网络中获取与所述疾病实体关联的目标实体之前,所述处理器执行所述至少一个计算机可读指令还用以实现以 下步骤:获取当前疾病,并获取与所述当前疾病相关联的症状属性;将所述当前疾病转换为疾病向量,并将所述症状属性转换为症状向量;利用注意力机制计算所述症状向量与所述疾病向量的关联度;将所述当前疾病、所述症状属性、所述关联度录入至图模板中,得到所述图神经网络。
- 根据权利要求8所述的电子设备,其中,在所述基于所述权重将所述目标实体转换为医学知识特征向量时,所述处理器执行所述至少一个计算机可读指令以实现以下步骤:获取所述目标实体的向量值,得到实体向量;基于所述权重对所述实体向量进行加权和运算,得到所述医学知识特征向量。
- 根据权利要求8所述的电子设备,其中,在所述利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量时,所述处理器执行所述至少一个计算机可读指令以实现以下步骤:根据预设数值对所述主诉数据进行分词,得到多个主诉分词;对所述多个主诉分词进行向量化处理,得到多个主诉向量,每个主诉向量包括多个维度;确定每个主诉向量中向量值最大的维度,得到目标维度,并获取所述目标维度对应的向量值;拼接获取的向量值,得到所述文本特征向量。
- 一种计算机可读存储介质,其中,所述计算机可读存储介质存储有至少一个计算机可读指令,所述至少一个计算机可读指令被处理器执行时实现以下步骤:当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;基于所述权重将所述目标实体转换为医学知识特征向量;利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;将所述目标向量输入至判别模型中,得到疾病列表;检测所述待诊断疾病是否存在于所述疾病列表中;当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
- 根据权利要求14所述的存储介质,其中,在所述根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史时,所述至少一个计算机可读指令被处理器执行以实现以下步骤:从线程连接池中获取任意闲置线程;利用所述任意闲置线程解析所述医疗误诊检测请求的方法体,得到所述医疗误诊检测请求携带的报文信息;获取预设标签,并从所述报文信息中获取与所述预设标签对应的信息,作为身份识别码;利用所述身份识别码确定所述待诊断用户,并从建档库中获取与所述待诊断用户对应的信息,作为所述现病史。
- 根据权利要求14所述的存储介质,其中,在所述抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体时,所述至少一个计算机可读指令被处理器执行以实现以下步骤:对所述主诉数据进行分词处理,得到第一分词,并对所述现病史进行分词处理,得到第 二分词;遍历预先构建的词典,并将遍历到的第一分词确定为所述主诉数据中的实体,将遍历到的第二分词确定为所述现病史中的实体;融合所述主诉数据中的实体及所述现病史中的实体,得到所述疾病实体。
- 根据权利要求14所述的存储介质,其中,在从预先构建的图神经网络中获取与所述疾病实体关联的目标实体之前,所述至少一个计算机可读指令被处理器执行还用以实现以下步骤:获取当前疾病,并获取与所述当前疾病相关联的症状属性;将所述当前疾病转换为疾病向量,并将所述症状属性转换为症状向量;利用注意力机制计算所述症状向量与所述疾病向量的关联度;将所述当前疾病、所述症状属性、所述关联度录入至图模板中,得到所述图神经网络。
- 根据权利要求14所述的存储介质,其中,在所述基于所述权重将所述目标实体转换为医学知识特征向量时,所述至少一个计算机可读指令被处理器执行时还用以实现以下步骤:获取所述目标实体的向量值,得到实体向量;基于所述权重对所述实体向量进行加权和运算,得到所述医学知识特征向量。
- 根据权利要求14所述的存储介质,其中,在所述利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量时,所述至少一个计算机可读指令被处理器执行以实现以下步骤:根据预设数值对所述主诉数据进行分词,得到多个主诉分词;对所述多个主诉分词进行向量化处理,得到多个主诉向量,每个主诉向量包括多个维度;确定每个主诉向量中向量值最大的维度,得到目标维度,并获取所述目标维度对应的向量值;拼接获取的向量值,得到所述文本特征向量。
- 一种医疗误诊检测装置,其中,所述医疗误诊检测装置包括:获取单元,用于当接收到医疗误诊检测请求时,从所述医疗误诊检测请求中获取主诉数据及待诊断疾病;确定单元,用于根据所述医疗误诊检测请求确定待诊断用户,并获取所述待诊断用户的现病史;抽取单元,用于抽取所述主诉数据中的实体及所述现病史中的实体,得到疾病实体;所述获取单元,还用于从预先构建的图神经网络中获取与所述疾病实体关联的目标实体,并获取所述目标实体的权重;转换单元,用于基于所述权重将所述目标实体转换为医学知识特征向量;处理单元,用于利用卷积神经网络对所述主诉数据进行处理,得到文本特征向量;拼接单元,用于拼接所述医学知识特征向量及所述文本特征向量,得到目标向量;输入单元,用于将所述目标向量输入至判别模型中,得到疾病列表;检测单元,用于检测所述待诊断疾病是否存在于所述疾病列表中;所述确定单元,还用于当所述待诊断疾病不存在于所述疾病列表中时,将所述医疗误诊检测请求的响应结果确定为误诊。
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010739726.2A CN111883251A (zh) | 2020-07-28 | 2020-07-28 | 医疗误诊检测方法、装置、电子设备及存储介质 |
CN202010739726.2 | 2020-07-28 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2021120688A1 true WO2021120688A1 (zh) | 2021-06-24 |
Family
ID=73201859
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2020/112235 WO2021120688A1 (zh) | 2020-07-28 | 2020-08-28 | 医疗误诊检测方法、装置、电子设备及存储介质 |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN111883251A (zh) |
WO (1) | WO2021120688A1 (zh) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113657114A (zh) * | 2021-08-31 | 2021-11-16 | 平安医疗健康管理股份有限公司 | 疾病名称对码列表的生成方法、装置、设备及存储介质 |
CN113921144A (zh) * | 2021-09-23 | 2022-01-11 | 清华大学 | 疾病预测集处理方法、装置、电子设备及存储介质 |
CN115830017A (zh) * | 2023-02-09 | 2023-03-21 | 智慧眼科技股份有限公司 | 基于图文多模态融合的肿瘤检测系统、方法、设备及介质 |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112541066B (zh) * | 2020-12-11 | 2022-10-25 | 清华大学 | 基于文本结构化的医技报告检测方法及相关设备 |
CN112435745B (zh) * | 2020-12-18 | 2024-04-05 | 深圳赛安特技术服务有限公司 | 就诊策略推荐方法、装置、电子设备及存储介质 |
CN113327691B (zh) * | 2021-06-01 | 2022-08-12 | 平安科技(深圳)有限公司 | 基于语言模型的问询方法、装置、计算机设备及存储介质 |
CN116682551B (zh) * | 2023-07-27 | 2023-12-22 | 腾讯科技(深圳)有限公司 | 疾病预测方法、疾病预测模型训练方法和装置 |
CN118116584A (zh) * | 2024-04-23 | 2024-05-31 | 鼎泰(南京)临床医学研究有限公司 | 一种基于大数据的可调整医疗辅助诊断系统及方法 |
Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106951719A (zh) * | 2017-04-10 | 2017-07-14 | 荣科科技股份有限公司 | 临床诊断模型的构建方法及构建系统、临床诊断系统 |
CN107833629A (zh) * | 2017-10-25 | 2018-03-23 | 厦门大学 | 基于深度学习的辅助诊断方法及系统 |
CN109545384A (zh) * | 2018-11-21 | 2019-03-29 | 上海依智医疗技术有限公司 | 一种医学诊断方法及装置 |
CN110277165A (zh) * | 2019-06-27 | 2019-09-24 | 清华大学 | 基于图神经网络的辅助诊断方法、装置、设备及存储介质 |
CN111063429A (zh) * | 2019-10-25 | 2020-04-24 | 中国科学院自动化研究所 | 一种医疗咨询方法、装置、设备和计算机可读存储介质 |
CN111161819A (zh) * | 2019-12-31 | 2020-05-15 | 重庆亚德科技股份有限公司 | 一种中医病历数据处理系统及方法 |
CN111225612A (zh) * | 2017-10-17 | 2020-06-02 | 萨蒂什·拉奥 | 基于机器学习的神经障碍识别和监测系统 |
WO2020123723A1 (en) * | 2018-12-11 | 2020-06-18 | K Health Inc. | System and method for providing health information |
CN111402979A (zh) * | 2020-03-24 | 2020-07-10 | 清华大学 | 病情描述与诊断一致性检测方法及装置 |
Family Cites Families (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
GB2561156A (en) * | 2017-03-24 | 2018-10-10 | Clinova Ltd | Apparatus, method and computer program |
CN110162779B (zh) * | 2019-04-04 | 2023-08-04 | 北京百度网讯科技有限公司 | 病历质量的评估方法、装置及设备 |
CN110534206A (zh) * | 2019-08-26 | 2019-12-03 | 北京好医生云医院管理技术有限公司 | 一种医疗诊断辅助系统的工作方法 |
CN110675951A (zh) * | 2019-08-26 | 2020-01-10 | 北京百度网讯科技有限公司 | 智能化的疾病诊断方法及装置、计算机设备与可读介质 |
CN111091906B (zh) * | 2019-10-31 | 2023-06-20 | 中电药明数据科技(成都)有限公司 | 一种基于真实世界数据的辅助医疗诊断方法及系统 |
CN111383769B (zh) * | 2020-01-08 | 2024-04-12 | 科大讯飞股份有限公司 | 主诉与诊断一致性检测方法、装置、设备及存储介质 |
-
2020
- 2020-07-28 CN CN202010739726.2A patent/CN111883251A/zh active Pending
- 2020-08-28 WO PCT/CN2020/112235 patent/WO2021120688A1/zh active Application Filing
Patent Citations (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106951719A (zh) * | 2017-04-10 | 2017-07-14 | 荣科科技股份有限公司 | 临床诊断模型的构建方法及构建系统、临床诊断系统 |
CN111225612A (zh) * | 2017-10-17 | 2020-06-02 | 萨蒂什·拉奥 | 基于机器学习的神经障碍识别和监测系统 |
CN107833629A (zh) * | 2017-10-25 | 2018-03-23 | 厦门大学 | 基于深度学习的辅助诊断方法及系统 |
CN109545384A (zh) * | 2018-11-21 | 2019-03-29 | 上海依智医疗技术有限公司 | 一种医学诊断方法及装置 |
WO2020123723A1 (en) * | 2018-12-11 | 2020-06-18 | K Health Inc. | System and method for providing health information |
CN110277165A (zh) * | 2019-06-27 | 2019-09-24 | 清华大学 | 基于图神经网络的辅助诊断方法、装置、设备及存储介质 |
CN111063429A (zh) * | 2019-10-25 | 2020-04-24 | 中国科学院自动化研究所 | 一种医疗咨询方法、装置、设备和计算机可读存储介质 |
CN111161819A (zh) * | 2019-12-31 | 2020-05-15 | 重庆亚德科技股份有限公司 | 一种中医病历数据处理系统及方法 |
CN111402979A (zh) * | 2020-03-24 | 2020-07-10 | 清华大学 | 病情描述与诊断一致性检测方法及装置 |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113657114A (zh) * | 2021-08-31 | 2021-11-16 | 平安医疗健康管理股份有限公司 | 疾病名称对码列表的生成方法、装置、设备及存储介质 |
CN113921144A (zh) * | 2021-09-23 | 2022-01-11 | 清华大学 | 疾病预测集处理方法、装置、电子设备及存储介质 |
CN115830017A (zh) * | 2023-02-09 | 2023-03-21 | 智慧眼科技股份有限公司 | 基于图文多模态融合的肿瘤检测系统、方法、设备及介质 |
Also Published As
Publication number | Publication date |
---|---|
CN111883251A (zh) | 2020-11-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2021120688A1 (zh) | 医疗误诊检测方法、装置、电子设备及存储介质 | |
WO2021114736A1 (zh) | 医疗问诊辅助方法、装置、电子设备及介质 | |
WO2021151299A1 (zh) | 基于人工智能的数据增强方法、装置、电子设备及介质 | |
CN111694840B (zh) | 数据同步方法、装置、服务器及存储介质 | |
WO2022105115A1 (zh) | 问答对匹配方法、装置、电子设备及存储介质 | |
WO2021217867A1 (zh) | 基于XGBoost的数据分类方法、装置、计算机设备及存储介质 | |
WO2021159738A1 (zh) | 基于医学领域的数据推荐方法、设备、服务器及存储介质 | |
CN113032528B (zh) | 案件分析方法、装置、设备及存储介质 | |
CN111985241B (zh) | 医学信息查询方法、装置、电子设备及介质 | |
WO2021196825A1 (zh) | 摘要生成方法、装置、电子设备及介质 | |
CN111459967A (zh) | 结构化查询语句生成方法、装置、电子设备及介质 | |
CN111651363B (zh) | 测试数据获取方法、装置、电子设备及介质 | |
CN113724830B (zh) | 基于人工智能的用药风险检测方法及相关设备 | |
WO2021151302A1 (zh) | 基于机器学习的药品质控分析方法、装置、设备及介质 | |
CN111986763B (zh) | 疾病数据分析方法、装置、电子设备及存储介质 | |
WO2021174923A1 (zh) | 概念词序列生成方法、装置、计算机设备及存储介质 | |
CN113470775B (zh) | 信息采集方法、装置、设备及存储介质 | |
CN111986793A (zh) | 基于人工智能的导诊处理方法、装置、计算机设备及介质 | |
CN111933241B (zh) | 医疗数据解析方法、装置、电子设备及存储介质 | |
CN112435745B (zh) | 就诊策略推荐方法、装置、电子设备及存储介质 | |
CN112199494B (zh) | 医疗信息搜索方法、装置、电子设备及存储介质 | |
CN114360732B (zh) | 医疗数据分析方法、装置、电子设备及存储介质 | |
US20220027612A1 (en) | Detecting and processing sections spanning processed document partitions | |
CN113380414B (zh) | 基于大数据的数据采集方法及系统 | |
KR20240134901A (ko) | 이미지 분석 및 통찰력 생성 |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 20902449 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
32PN | Ep: public notification in the ep bulletin as address of the adressee cannot be established |
Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 22/03/2023) |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 20902449 Country of ref document: EP Kind code of ref document: A1 |