Medical corpus generation method, device and equipment and computer readable storage medium

Abstract

The application relates to the field of medical technology data processing, and discloses a medical corpus generation method, which comprises the following steps: extracting keywords of each target medical data; classifying the target medical data according to the keywords to obtain a classification label of each target medical data; filling the target medical data into a preset medical corpus template by using the classification label of each target medical data according to the preset medical corpus template to obtain a medical corpus. The application also discloses a medical corpus generation device, an electronic device and a computer readable storage medium. The application can improve the synthesis efficiency of the medical corpus.

Description

Technical Field

[0001] This invention relates to the field of medical technology data processing, and in particular to a method, apparatus, electronic device, and computer-readable storage medium for generating medical corpora. Background Technology

[0002] With the development of artificial intelligence technology, various types of medical models have been introduced into the medical field to predict the risk of disease occurrence, simulate disease progression, optimize clinical trial design, and assist doctors in diagnosis and treatment.

[0003] The design and training of medical models typically rely on a large amount of medical training data, such as patients' personal disease diagnosis and treatment information, genomes, physiological indicators, medical images, and symptom information. However, due to regional and platform limitations, as well as the need to protect personal data privacy, the available training data for medical models is insufficient to support the training of large-scale medical models, resulting in lower accuracy. Manually assembling medical data is often used, but this is slow and the objectivity and rationality of the data need improvement. Therefore, increasing the quantity of training data available for medical models has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method, apparatus, electronic device, and computer-readable storage medium for generating medical corpora, the main purpose of which is to improve the efficiency of medical corpus synthesis.

[0005] To achieve the above objectives, the present invention provides a method for generating medical corpus, comprising:

[0006] Retrieve target medical data that meets preset medical data characteristics from preset data sources;

[0007] Extract keywords from each of the target medical data;

[0008] The target medical data is classified according to the keywords to obtain a classification label for each target medical data.

[0009] Based on a preset medical corpus template, the target medical data is populated into the preset medical corpus template using the classification tags of each target medical data data, thus obtaining a medical corpus.

[0010] Optionally, obtaining target medical data that meets preset medical data characteristics from a preset data source includes:

[0011] Identify data units in the preset data source and obtain the term for each data unit;

[0012] Extract the text features of each term;

[0013] The similarity between the text features of each term and the preset medical data features is calculated sequentially.

[0014] Select data units corresponding to terms with similarity greater than or equal to a preset similarity threshold as the target medical data.

[0015] Optionally, the extraction of keywords for each of the target medical data includes:

[0016] Each of the target medical data is segmented into words to obtain a word segmentation set, and each word in the word segmentation set is converted into a word vector;

[0017] Using the word segmentation set and the word vector of each word segment, a text vector matrix of the target medical data is constructed;

[0018] Select one of the word segments in sequence as the target word segment, and calculate the key value of the target word segment based on the word vector of the target word segment and the text vector matrix;

[0019] A preset number of words are selected as keywords from the word segmentation set according to the order of the key values ​​from largest to smallest.

[0020] Optionally, the step of classifying the target medical data according to the keywords to obtain a classification label for each target medical data includes:

[0021] By utilizing the keywords of each target medical data point, textual features of the target medical data are generated;

[0022] Calculate the relative probability values ​​between the text features and the preset classification labels;

[0023] The score of each classification label is calculated based on the relative probability value, and the classification label with the highest score is taken as the classification label of the target medical data.

[0024] Optionally, the step of filling the target medical data into the preset medical corpus template using the classification tags of each target medical data point, based on a preset medical corpus template, to obtain medical corpus, includes:

[0025] The preset medical corpus template is decomposed into template information units;

[0026] Each template information unit is labeled using the classification tags;

[0027] Using permutation and combination, medical data corresponding to the category label are selected sequentially to fill the template information unit corresponding to the category label;

[0028] The medical corpus is obtained by collecting and filling the template information units.

[0029] Optionally, before extracting keywords from each of the target medical data, the method further includes:

[0030] The target medical data is anonymized according to a preset sensitive word dictionary;

[0031] The anonymized medical data is standardized according to a pre-defined standard medical terminology dictionary.

[0032] To address the above problems, the present invention also provides a medical corpus generation apparatus, the apparatus comprising:

[0033] The source data acquisition module is used to acquire target medical data that meets the preset medical data characteristics from the preset data source.

[0034] The source data classification module is used to extract keywords for each target medical data, classify the target medical data according to the keywords, and obtain a classification label for each target medical data.

[0035] The source data assembly module is used to fill the target medical data into the preset medical corpus template according to the preset medical corpus template and using the classification tags of each target medical data to obtain medical corpus.

[0036] Optionally, the source data acquisition module acquires target medical data that meets preset medical data characteristics from a preset data source using the following method:

[0037] Identify data units in the preset data source and obtain the term for each data unit;

[0038] Extract the text features of each term;

[0039] The similarity between the text features of each term and the preset medical data features is calculated sequentially.

[0040] Select data units corresponding to terms with similarity greater than or equal to a preset similarity threshold as the target medical data.

[0041] To address the above problems, the present invention also provides an electronic device, the electronic device comprising:

[0042] Memory, storing at least one computer program; and

[0043] The processor executes the program stored in the memory to implement the medical corpus generation method described above.

[0044] To address the aforementioned problems, the present invention also provides a computer-readable storage medium storing at least one computer program, which is executed by a processor in an electronic device to implement the medical corpus generation method described above.

[0045] This invention application obtains target medical data that meets preset medical data characteristics from a preset data source. By classifying each target medical data, classification management of medical data can be achieved. At the same time, based on the data organization form of a preset medical corpus template, target medical data under different classification tags are filled into the preset medical corpus template, completing the transformation from target medical data to medical corpus and improving the generation efficiency of medical corpus. Attached Figure Description

[0046] Figure 1 This is a flowchart illustrating a medical corpus generation method according to an embodiment of the present invention.

[0047] Figure 2 A detailed implementation flowchart of one step of the medical corpus generation method provided in an embodiment of the present invention is shown below;

[0048] Figure 3 A detailed implementation flowchart of one step of the medical corpus generation method provided in an embodiment of the present invention is shown below;

[0049] Figure 4 This is a functional block diagram of a medical corpus generation device provided in an embodiment of the present invention;

[0050] Figure 5 This is a schematic diagram of the structure of an electronic device for implementing the medical corpus generation method according to an embodiment of the present invention.

[0051] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0052] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0053] This application provides a method for generating medical corpus. The execution entity of the medical corpus generation method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the medical corpus generation method can be executed by software or hardware installed on a terminal device or a server device, and the software can be a blockchain platform. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0054] Reference Figure 1 The diagram shown is a flowchart illustrating a medical corpus generation method according to an embodiment of the present invention. In this embodiment, the medical corpus generation method includes:

[0055] S1. Obtain target medical data that meets the preset medical data characteristics from the preset data source;

[0056] In this embodiment of the invention, the preset data source refers to an authorized data area for storing medical-related data, such as a hospital's electronic pathology system, medical paper websites, and medical databases. It should be noted that the preset data source includes, but is not limited to, databases, blockchains, and web caches.

[0057] In this embodiment of the invention, the preset medical data features can be set according to the needs of actual medical model training. For example, if the medical model is used to simulate the evolution of a specific disease, the corresponding medical data features can be the features of that specific disease, including but not limited to the medical terminology features, symptom features, gene features, indicator features, etc. of that specific disease.

[0058] In this embodiment of the invention, the target medical data includes, but is not limited to, patient medical records, medical papers, etc.

[0059] In this embodiment of the invention, a crawler strategy can be designed based on the preset medical data characteristics, and a crawler program can be written to obtain medical data that meets the preset medical data characteristics from the preset data source. Alternatively, API interfaces and data mining techniques can be used to obtain medical data that meets the preset medical data characteristics.

[0060] For details, please refer to Figure 2 As shown, S1 includes:

[0061] S11. Identify the data units in the preset data source and obtain the terms for each data unit;

[0062] S12. Extract the text features of each term;

[0063] S13. Calculate the similarity between the text features of each term and the preset medical data features in sequence;

[0064] S14. Select the data unit corresponding to the term with a similarity greater than or equal to the preset similarity threshold as the target medical data.

[0065] In this embodiment of the invention, the data unit may be a medical paper, a disease diagnosis form, a laboratory test report, etc.

[0066] In this embodiment of the invention, the term for each data unit refers to the text that summarizes the data unit, such as the title of a medical paper or the name of a disease diagnosis report.

[0067] In this embodiment of the invention, the terms are text content expressed in natural language. Directly analyzing each term would consume significant computing resources, resulting in low analysis efficiency. Therefore, NLP (Natural Language Processing) technology can be used to convert the terms into text features. These text features include, but are not limited to, medical terminology and medical treatment features corresponding to each term.

[0068] In this embodiment of the invention, the text features of each term and the preset medical data features can be vectorized, and then the distance between the vectorized text features and the vectorized data features can be calculated using the Euclidean distance formula or the Mahalanobis distance formula. The calculated distance value is used to characterize the similarity between the two.

[0069] In this embodiment of the invention, the preset similarity threshold can be set according to actual business needs.

[0070] In this embodiment of the invention, the target medical data meets preset medical data characteristics and can be used as the data basis for synthesizing medical corpora, thus ensuring the accuracy of the source data for synthesizing medical corpora.

[0071] S2. Extract keywords for each target medical data, classify the target medical data according to the keywords, and obtain classification labels for each target medical data;

[0072] In this embodiment of the invention, the keywords of the target medical data refer to the core key information contained in the target medical data, such as treatment method information, disease description information, description information of disease types such as chronic and acute, and description information of complications.

[0073] It should be noted that in this embodiment of the invention, the keywords of each target medical data are different from the terms of each data unit in the preset data source. The scope represented by the terms of each data unit is broader than that of the keywords of each target medical data. For example, if the term corresponding to a certain medical data is acute gastroenteritis, the keywords corresponding to the medical data can be descriptive information, diagnostic information, or descriptive information of related complications of acute gastroenteritis.

[0074] It is understandable that target medical data obtained from different preset data sources may have inconsistent terminology or symbolism, and may contain some patient privacy information. Therefore, before extracting keywords from each target medical data, the following preprocessing can be performed on each target medical data:

[0075] The target medical data is anonymized according to a preset sensitive word dictionary;

[0076] The anonymized medical data is standardized according to a pre-defined standard medical terminology dictionary.

[0077] In this embodiment of the invention, the preset sensitive word dictionary refers to entries that are considered personal privacy according to relevant business regulations, such as patients' names, genders, contact information, home addresses, social security information, etc. The preset standard medical terminology dictionary refers to a collection of medical terms that conform to international standards.

[0078] Specifically, the extraction of keywords for each of the target medical data includes:

[0079] Each of the target medical data is segmented into words to obtain a word segmentation set, and each word in the word segmentation set is converted into a word vector;

[0080] Using the word segmentation set and the word vector of each word segment, a text vector matrix of the target medical data is constructed;

[0081] Select one of the word segments in sequence as the target word segment, and calculate the key value of the target word segment based on the word vector of the target word segment and the text vector matrix;

[0082] A preset number of words are selected as keywords from the word segmentation set according to the order of the key values ​​from largest to smallest.

[0083] In this embodiment of the invention, methods such as Glove (Global Vectors for Word Representation) and Embedding Layer can be used to convert the word vectors of each segmented word into a text vector matrix.

[0084] In this embodiment of the invention, the key value of the target word segmentation can be calculated using the following key value algorithm:

[0085]

[0086] Where K is the key value, |W T | represents the modulus of the transpose of the text vector matrix. The modulus of the word vectors of the target word segmentation.

[0087] It is understandable that not every word segment in each target medical data represents the key features of that medical data. By extracting keywords for each target medical data using the above method, the computational workload is reduced and the analysis efficiency is improved compared to directly using each target medical data for subsequent data analysis.

[0088] Specifically, the step of classifying the target medical data based on the keywords to obtain a classification label for each target medical data item includes:

[0089] By utilizing the keywords of each target medical data point, textual features of the target medical data are generated;

[0090] Calculate the relative probability values ​​between the text features and the preset classification labels;

[0091] The score of each classification label is calculated based on the relative probability value, and the classification label with the highest score is taken as the classification label of the target medical data.

[0092] In an embodiment of the present invention, keyword vectors can be obtained by converting keywords into vectors, and keyword vector matrices can be obtained by concatenating the keyword vectors corresponding to each target medical data. The keyword vector matrix can then be used as the text features of the target medical data.

[0093] In this embodiment of the invention, the types of medical data classification can be set according to the purpose of training the actual medical model, i.e., the preset classification labels. For example, to evaluate the effectiveness of multiple treatment methods, the preset classification labels can be electrotherapy, physical rehabilitation, drug therapy, surgical treatment, etc. To assess the severity of the condition, the preset classification labels can be early stage, early-middle stage, middle stage, middle-late stage, late stage, etc.

[0094] In this embodiment of the invention, the relative probability value between the text features and the preset classification label can be calculated using a pre-trained activation function, wherein the pre-trained activation function includes, but is not limited to, the softmax activation function, the sigmoid activation function, and the ReLU activation function.

[0095] In this embodiment of the invention, the purpose of classifying each target medical data is to achieve classified management of medical data. Furthermore, the classification tags can be used to combine medical data of the same or different categories, thereby generating more medical corpora.

[0096] S3. Based on the preset medical corpus template, the target medical data is filled into the preset medical corpus template using the classification tags of each target medical data to obtain the medical corpus.

[0097] In this embodiment of the invention, the preset medical corpus template refers to the structural composition of medical corpus samples set according to the purpose of training actual medical models. For example, a medical model is used to simulate the evolution of a patient's condition. Accordingly, the preset medical corpus template includes, but is not limited to, information units such as patient gender, age, basic medical history, symptoms, severity of condition, and complications.

[0098] In this embodiment of the invention, the information unit in the preset medical corpus template is a structured organization form for classifying medical data, and the classification label of each target medical data is also a classification organization form for medical data. Therefore, the target medical data and the preset medical corpus template have the same data organization logic, and the classification labels can be used to combine medical data into a medical corpus that conforms to the preset medical corpus template.

[0099] For details, please refer to Figure 3 As shown, S3 includes:

[0100] S31. Decompose the preset medical corpus template into template information units;

[0101] S32. Use the classification labels to label each template information unit;

[0102] S33. Using permutation and combination, sequentially select the medical data corresponding to the classification label to fill the template information unit corresponding to the classification label;

[0103] S34. Collect the filled template information units to obtain the medical corpus.

[0104] For example, the preset medical corpus template includes template information units such as underlying disease and disease severity. The classification label corresponding to the underlying disease is the disease name. Accordingly, the medical data under the disease name classification label includes diabetes and rhinitis. Therefore, the content that can be filled in for the underlying disease includes three situations: diabetes, rhinitis, and diabetes + rhinitis. The classification label corresponding to the disease severity is the disease level, and the corresponding medical data includes early stage, early-middle stage, middle stage, middle-late stage, and late stage. Therefore, the disease severity can be filled in with five situations: early stage, early-middle stage, middle stage, middle-late stage, and late stage.

[0105] This invention application obtains target medical data that meets preset medical data characteristics from a preset data source. By classifying each target medical data, classification management of medical data can be achieved. At the same time, based on the data organization form of a preset medical corpus template, target medical data under different classification tags are filled into the preset medical corpus template, completing the transformation from target medical data to medical corpus and improving the generation efficiency of medical corpus.

[0106] like Figure 4 The diagram shown is a functional block diagram of a medical corpus generation device provided in an embodiment of the present invention.

[0107] The medical corpus generation device 100 of the present invention can be installed in an electronic device. Depending on the functions implemented, the medical corpus generation device 100 includes a source data acquisition module 101, a source data classification module 102, and a source data assembly module 103. The module in this invention can also be called a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and which are stored in the memory of the electronic device.

[0108] In this embodiment, the functions of each module / unit are as follows:

[0109] The source data acquisition module 101 is used to acquire target medical data that meets preset medical data characteristics from a preset data source.

[0110] The source data classification module 102 is used to extract keywords for each target medical data, classify the target medical data according to the keywords, and obtain a classification label for each target medical data.

[0111] The source data assembly module 103 is used to fill the target medical data into the preset medical corpus template according to the preset medical corpus template and using the classification tags of each target medical data to obtain medical corpus.

[0112] In detail, the specific implementation methods of each module of the medical corpus generation device 100 are as follows:

[0113] Step 1: Obtain target medical data that meets the preset medical data characteristics from the preset data source;

[0114] In this embodiment of the invention, the preset data source refers to an authorized data area for storing medical-related data, such as a hospital's electronic pathology system, medical paper websites, and medical databases. It should be noted that the preset data source includes, but is not limited to, databases, blockchains, and web caches.

[0115] In this embodiment of the invention, the preset medical data features can be set according to the needs of actual medical model training. For example, if the medical model is used to simulate the evolution of a specific disease, the corresponding medical data features can be the features of that specific disease, including but not limited to the medical terminology features, symptom features, gene features, indicator features, etc. of that specific disease.

[0116] In this embodiment of the invention, the target medical data includes, but is not limited to, patient medical records, medical papers, etc.

[0117] In this embodiment of the invention, a crawler strategy can be designed based on the preset medical data characteristics, and a crawler program can be written to obtain medical data that meets the preset medical data characteristics from the preset data source. Alternatively, API interfaces and data mining techniques can be used to obtain medical data that meets the preset medical data characteristics.

[0118] Specifically, obtaining target medical data that meets preset medical data characteristics from a preset data source includes:

[0119] Identify data units in the preset data source and obtain the term for each data unit;

[0120] Extract the text features of each term;

[0121] The similarity between the text features of each term and the preset medical data features is calculated sequentially.

[0122] Select data units corresponding to terms with similarity greater than or equal to a preset similarity threshold as the target medical data.

[0123] In this embodiment of the invention, the data unit may be a medical paper, a disease diagnosis form, a laboratory test report, etc.

[0124] In this embodiment of the invention, the term for each data unit refers to the text that summarizes the data unit, such as the title of a medical paper or the name of a disease diagnosis report.

[0125] In this embodiment of the invention, the terms are text content expressed in natural language. Directly analyzing each term would consume significant computing resources, resulting in low analysis efficiency. Therefore, NLP (Natural Language Processing) technology can be used to convert the terms into text features. These text features include, but are not limited to, medical terminology and medical treatment features corresponding to each term.

[0126] In this embodiment of the invention, the text features of each term and the preset medical data features can be vectorized, and then the distance between the vectorized text features and the vectorized data features can be calculated using the Euclidean distance formula or the Mahalanobis distance formula. The calculated distance value is used to characterize the similarity between the two.

[0127] In this embodiment of the invention, the preset similarity threshold can be set according to actual business needs.

[0128] In this embodiment of the invention, the target medical data meets preset medical data characteristics and can be used as the data basis for synthesizing medical corpora, thus ensuring the accuracy of the source data for synthesizing medical corpora.

[0129] Step 2: Extract keywords for each target medical data, and classify the target medical data according to the keywords to obtain classification labels for each target medical data;

[0130] In this embodiment of the invention, the keywords of the target medical data refer to the core key information contained in the target medical data, such as treatment method information, disease description information, description information of disease types such as chronic and acute, and description information of complications.

[0131] It should be noted that in this embodiment of the invention, the keywords of each target medical data are different from the terms of each data unit in the preset data source. The scope represented by the terms of each data unit is broader than that of the keywords of each target medical data. For example, if the term corresponding to a certain medical data is acute gastroenteritis, the keywords corresponding to the medical data can be descriptive information, diagnostic information, or descriptive information of related complications of acute gastroenteritis.

[0132] It is understandable that target medical data obtained from different preset data sources may have inconsistent terminology or symbolism, and may contain some patient privacy information. Therefore, before extracting keywords from each target medical data, the following preprocessing can be performed on each target medical data:

[0133] The target medical data is anonymized according to a preset sensitive word dictionary;

[0134] The anonymized medical data is standardized according to a pre-defined standard medical terminology dictionary.

[0135] In this embodiment of the invention, the preset sensitive word dictionary refers to entries that are considered personal privacy according to relevant business regulations, such as patients' names, genders, contact information, home addresses, social security information, etc. The preset standard medical terminology dictionary refers to a collection of medical terms that conform to international standards.

[0136] Specifically, the extraction of keywords for each of the target medical data includes:

[0137] Each of the target medical data is segmented into words to obtain a word segmentation set, and each word in the word segmentation set is converted into a word vector;

[0138] Using the word segmentation set and the word vector of each word segment, a text vector matrix of the target medical data is constructed;

[0139] Select one of the word segments in sequence as the target word segment, and calculate the key value of the target word segment based on the word vector of the target word segment and the text vector matrix;

[0140] A preset number of words are selected as keywords from the word segmentation set according to the order of the key values ​​from largest to smallest.

[0141] In this embodiment of the invention, methods such as Glove (Global Vectors for Word Representation) and Embedding Layer can be used to convert the word vectors of each segmented word into a text vector matrix.

[0142] In this embodiment of the invention, the key value of the target word segmentation can be calculated using the following key value algorithm:

[0143]

[0144] Where K is the key value, |W T | represents the modulus of the transpose of the text vector matrix. The modulus of the word vectors of the target word segmentation.

[0145] It is understandable that not every word segment in each target medical data represents the key features of that medical data. By extracting keywords for each target medical data using the above method, the computational workload is reduced and the analysis efficiency is improved compared to directly using each target medical data for subsequent data analysis.

[0146] Specifically, the step of classifying the target medical data based on the keywords to obtain a classification label for each target medical data item includes:

[0147] By utilizing the keywords of each target medical data point, textual features of the target medical data are generated;

[0148] Calculate the relative probability values ​​between the text features and the preset classification labels;

[0149] The score of each classification label is calculated based on the relative probability value, and the classification label with the highest score is taken as the classification label of the target medical data.

[0150] In an embodiment of the present invention, keyword vectors can be obtained by converting keywords into vectors, and keyword vector matrices can be obtained by concatenating the keyword vectors corresponding to each target medical data. The keyword vector matrix can then be used as the text features of the target medical data.

[0151] In this embodiment of the invention, the types of medical data classification can be set according to the purpose of training the actual medical model, i.e., the preset classification labels. For example, to evaluate the effectiveness of multiple treatment methods, the preset classification labels can be electrotherapy, physical rehabilitation, drug therapy, surgical treatment, etc. To assess the severity of the condition, the preset classification labels can be early stage, early-middle stage, middle stage, middle-late stage, late stage, etc.

[0152] In this embodiment of the invention, the relative probability value between the text features and the preset classification label can be calculated using a pre-trained activation function, wherein the pre-trained activation function includes, but is not limited to, the softmax activation function, the sigmoid activation function, and the ReLU activation function.

[0153] In this embodiment of the invention, the purpose of classifying each target medical data is to achieve classified management of medical data. Furthermore, the classification tags can be used to combine medical data of the same or different categories, thereby generating more medical corpora.

[0154] Step 3: Based on the preset medical corpus template, use the classification tags of each target medical data to fill the target medical data into the preset medical corpus template to obtain the medical corpus.

[0155] In this embodiment of the invention, the preset medical corpus template refers to the structural composition of medical corpus samples set according to the purpose of training actual medical models. For example, a medical model is used to simulate the evolution of a patient's condition. Accordingly, the preset medical corpus template includes, but is not limited to, information units such as patient gender, age, basic medical history, symptoms, severity of condition, and complications.

[0156] In this embodiment of the invention, the information unit in the preset medical corpus template is a structured organization form for classifying medical data, and the classification label of each target medical data is also a classification organization form for medical data. Therefore, the target medical data and the preset medical corpus template have the same data organization logic, and the classification labels can be used to combine medical data into a medical corpus that conforms to the preset medical corpus template.

[0157] Specifically, the step of filling the target medical data into the preset medical corpus template using the classification tags of each target medical data point, based on a preset medical corpus template, to obtain medical corpus, includes:

[0158] The preset medical corpus template is decomposed into template information units;

[0159] Each template information unit is labeled using the classification tags;

[0160] Using permutation and combination, medical data corresponding to the category label are selected sequentially to fill the template information unit corresponding to the category label;

[0161] The medical corpus is obtained by collecting and filling the template information units.

[0162] For example, the preset medical corpus template includes template information units such as underlying disease and disease severity. The classification label corresponding to the underlying disease is the disease name. Accordingly, the medical data under the disease name classification label includes diabetes and rhinitis. Therefore, the content that can be filled in for the underlying disease includes three situations: diabetes, rhinitis, and diabetes + rhinitis. The classification label corresponding to the disease severity is the disease level, and the corresponding medical data includes early stage, early-middle stage, middle stage, middle-late stage, and late stage. Therefore, the disease severity can be filled in with five situations: early stage, early-middle stage, middle stage, middle-late stage, and late stage.

[0163] The medical corpus generation device provided in this invention application obtains target medical data that meets preset medical data characteristics from a preset data source. By classifying each target medical data, the device can achieve classified management of medical data. At the same time, based on the data organization form of a preset medical corpus template, the device fills the target medical data under different classification tags into the preset medical corpus template, thus completing the transformation from target medical data to medical corpus and improving the generation efficiency of medical corpus.

[0164] like Figure 5 The diagram shown is a schematic representation of an electronic device for implementing a medical corpus generation method according to an embodiment of the present invention.

[0165] The electronic device 1 may include a processor 10, a memory 11 and a bus, and may also include a computer program stored in the memory 11 and executable on the processor 10, such as medical corpus generation.

[0166] The memory 11 includes at least one type of readable storage medium, such as flash memory, portable hard drive, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 11 can be an internal storage unit of the electronic device 1, such as a portable hard drive. In other embodiments, the memory 11 can be an external storage device of the electronic device 1, such as a plug-in portable hard drive, smart media card (SMC), secure digital card (SD), flash card, etc., equipped on the electronic device 1. Furthermore, the memory 11 can include both internal and external storage units of the electronic device 1. The memory 11 can be used not only to store application software and various types of data installed on the electronic device 1, such as code generated from medical corpora, but also to temporarily store data that has been output or will be output.

[0167] In some embodiments, the processor 10 may be composed of integrated circuits, such as a single packaged integrated circuit or multiple integrated circuits with the same or different functions, including combinations of one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and various control chips. The processor 10 is the control unit of the electronic device, connecting various components of the entire electronic device through various interfaces and lines. It executes programs or modules (e.g., medical corpus generation) stored in the memory 11, and calls data stored in the memory 11 to perform various functions of the electronic device 1 and process data.

[0168] The bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into an address bus, a data bus, a control bus, etc. The bus is configured to enable communication between the memory 11 and at least one processor 10, etc.

[0169] Figure 5Only electronic devices with components are shown; it will be understood by those skilled in the art that... Figure 5 The structure shown does not constitute a limitation on the electronic device 1, and may include fewer or more components than shown, or combine certain components, or have different component arrangements.

[0170] For example, although not shown, the electronic device 1 may also include a power supply (such as a battery) to power various components. Preferably, the power supply can be logically connected to the at least one processor 10 through a power management device, thereby enabling functions such as charging management, discharging management, and power consumption management. The power supply may also include one or more DC or AC power supplies, recharging devices, power fault detection circuits, power converters or inverters, power status indicators, and other arbitrary components. The electronic device 1 may also include various sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be described in detail here.

[0171] Furthermore, the electronic device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a Wi-Fi interface, a Bluetooth interface, etc.), which is typically used to establish communication connections between the electronic device 1 and other electronic devices.

[0172] Optionally, the electronic device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired or wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display may also be appropriately referred to as a screen or display unit, used to display information processed in the electronic device 1 and to display a visual user interface.

[0173] It should be understood that the embodiments described are for illustrative purposes only and are not limited to this structure in the scope of the patent application.

[0174] The medical corpus stored in the memory 11 of the electronic device 1 is generated by a combination of multiple instructions, which, when run in the processor 10, can achieve the following:

[0175] Retrieve target medical data that meets preset medical data characteristics from preset data sources;

[0176] Extract keywords from each target medical data, classify the target medical data according to the keywords, and obtain classification labels for each target medical data;

[0177] Based on a preset medical corpus template, the target medical data is populated into the preset medical corpus template using the classification tags of each target medical data data, thus obtaining a medical corpus.

[0178] Furthermore, if the modules / units integrated in the electronic device 1 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. The computer-readable storage medium can be volatile or non-volatile. For example, the computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, or a read-only memory (ROM).

[0179] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor of an electronic device, can perform the following:

[0180] Retrieve target medical data that meets preset medical data characteristics from preset data sources;

[0181] Extract keywords from each target medical data, classify the target medical data according to the keywords, and obtain classification labels for each target medical data;

[0182] Based on a preset medical corpus template, the target medical data is populated into the preset medical corpus template using the classification tags of each target medical data data, thus obtaining a medical corpus.

[0183] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0184] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0185] Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be embraced within the invention. No appended diagram markings in the claims should be construed as limiting the scope of the claims.

[0186] The blockchain referred to in this invention is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0187] The embodiments of this application can acquire and process relevant data based on holographic projection technology. Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0188] Furthermore, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices recited in a system claim may also be implemented by a single unit or device through software or hardware. The term "second class" is used to indicate names and does not indicate any specific order.

[0189] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for generating medical corpus based on a search engine, characterized in that, The method includes: Retrieve target medical data that meets preset medical data characteristics from preset data sources; Extracting keywords for each target medical data includes: segmenting each target medical data into words to obtain a word segmentation set; constructing a text vector matrix of the target medical data based on the word vectors corresponding to each word in the word segmentation set; sequentially selecting a word from the word segmentation set as a target word; calculating the key value of the target word based on the word vector of the target word and the text vector matrix; and selecting a preset number of words from the word segmentation set as keywords in descending order of the key values. Using the keywords of each target medical data, text features of the target medical data are generated, the relative probability value between the text features and the preset classification label is calculated, the score of each classification label is calculated based on the relative probability value, and the classification label with the highest score is taken as the classification label of the target medical data. The preset medical corpus template is decomposed into template information units. Each template information unit is labeled using the classification label of each target medical data. According to the preset arrangement and combination method, the medical data corresponding to the classification label is selected in sequence and filled into the corresponding template information unit in the preset medical corpus template to obtain medical corpus. The template information unit is a structured organization form for classifying medical data.

2. The medical corpus generation method as described in claim 1, characterized in that, The step of obtaining target medical data that meets preset medical data characteristics from a preset data source includes: Identify data units in the preset data source and obtain the term for each data unit; Extract the text features of each term; The similarity between the text features of each term and the preset medical data features is calculated sequentially. Select data units corresponding to terms with similarity greater than or equal to a preset similarity threshold as the target medical data.

3. The medical corpus generation method as described in claim 1, characterized in that, Before extracting keywords from each of the target medical data, the method further includes: The target medical data is anonymized according to a preset sensitive word dictionary; The anonymized medical data is standardized according to a pre-defined standard medical terminology dictionary.

4. A medical corpus generation device, characterized in that, The device includes: The source data acquisition module is used to acquire target medical data that meets the preset medical data characteristics from the preset data source. The source data classification module is used to extract keywords for each target medical data, including: segmenting each target medical data into words to obtain a word segmentation set; constructing a text vector matrix of the target medical data based on the word vectors corresponding to each word in the word segmentation set; sequentially selecting a word from the word segmentation set as a target word; calculating the key value of the target word based on the word vector of the target word and the text vector matrix; and selecting a preset number of words from the word segmentation set as keywords in descending order of the key values. The source data classification module is further configured to generate text features of the target medical data using keywords of each target medical data, calculate the relative probability value between the text features and the preset classification label, calculate the score of each classification label based on the relative probability value, and take the classification label with the highest score as the classification label of the target medical data. The source data assembly module is used to decompose a preset medical corpus template into template information units, label each template information unit with a classification label of each target medical data, and select the medical data corresponding to the classification label in sequence according to a preset arrangement and combination method to fill the corresponding template information unit in the preset medical corpus template to obtain medical corpus. The template information unit is a structured organization form for classifying medical data.

5. The medical corpus generation device as described in claim 4, characterized in that, The source data acquisition module obtains target medical data that meets preset medical data characteristics from a preset data source using the following method: Identify data units in the preset data source and obtain the term for each data unit; Extract the text features of each term; The similarity between the text features of each term and the preset medical data features is calculated sequentially. Select data units corresponding to terms with similarity greater than or equal to a preset similarity threshold as the target medical data.

6. An electronic device, characterized in that, The electronic device includes: At least one processor; and, A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the medical corpus generation method as described in any one of claims 1 to 3.

7. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the medical corpus generation method as described in any one of claims 1 to 3.

Images