Method and device for processing medical inquiry text information and electronic equipment

Abstract

The invention belongs to the technical field of computers and text data processing, and particularly discloses a medical inquiry text information processing method and device and electronic device.The method comprises the steps that a target medical inquiry text of a target object is obtained; key terms are matched in the target medical inquiry text through term knowledge bases in different knowledge fields; based on the key terms of the target medical inquiry text in different knowledge fields, constructing a field knowledge graph of the target medical inquiry text; based on the target medical inquiry text and a domain knowledge graph of the target medical inquiry text, determining a physiological obstacle level and a psychological obstacle level of the target object through a pre-trained medical inquiry text information processing model, the disorder level is used for representing the severity of the physiological disorder or the psychological disorder. Through the technical scheme provided by the invention, the accuracy of processing the medical inquiry text information can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of computer and text data processing, and specifically to a method, device and electronic device for processing medical consultation text information. Background Art

[0002] In the medical field, psychological disorders and physiological disorders have a close interaction, that is, psychological disorders may induce physiological disorders, and physiological disorders may also induce or aggravate psychological disorders. However, traditional medical information analysis models usually analyze physiological disorders or psychological disorders from a single perspective, and are unable to combine the intrinsic relationship between psychological disorders and physiological disorders, thereby reducing the accuracy of medical information analysis. Based on this, how to improve the accuracy of medical consultation text information processing is a technical problem that needs to be solved urgently. Summary of the Invention

[0003] The embodiments of the present application provide a method, apparatus, computer program product or computer program, computer-readable storage medium, and electronic device for processing medical consultation text information, thereby improving the accuracy of processing medical consultation text information at least to a certain extent.

[0004] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the present application.

[0005] According to a first aspect of an embodiment of the present application, a method for processing medical inquiry text information is provided, characterized in that the method includes: obtaining a target medical inquiry text of a target object, the target medical inquiry text being a descriptive text of the target object's physical and mental state; matching key terms in the target medical inquiry text through a terminology knowledge base in different knowledge fields, the terminology knowledge base including a plurality of professional terms belonging to its corresponding knowledge field; constructing a domain knowledge graph of the target medical inquiry text based on the key terms in different knowledge fields of the target medical inquiry text; and determining, based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text, a physiological disorder level and a psychological disorder level of the target object through a pre-trained medical inquiry text information processing model, the disorder level being used to characterize the severity of the physiological disorder or the psychological disorder.

[0006] In some embodiments of the present application, based on the aforementioned scheme, matching key terms in the target medical inquiry text through the terminology knowledge bases of different knowledge fields includes: traversing the terminology knowledge bases of each knowledge field, searching for target professional terms belonging to the terminology knowledge bases of each knowledge field from the target medical inquiry text; and determining the target professional terms as the key terms.

[0007] In some embodiments of the present application, based on the aforementioned scheme, the domain knowledge graph is used to represent the target object and each knowledge domain, as well as the correspondence between key terms in each knowledge domain.

[0008] In some embodiments of the present application, based on the aforementioned scheme, the target medical inquiry text and the domain knowledge graph of the target medical inquiry text are used to determine the target object's physiological disorder level and psychological disorder level through a pre-trained medical inquiry text information processing model, including: based on the target medical inquiry text, performing data conversion processing on the target medical inquiry text through a word embedding model and a baseline model to obtain a first feature matrix; constructing an adjacency matrix based on the domain knowledge graph of the target medical inquiry text to obtain a second feature matrix, wherein the adjacency matrix is used to reflect the position distribution characteristics of each key term in the target medical inquiry text; inputting the first feature matrix and the second feature matrix into the medical inquiry text information processing model, and obtaining the target object's physiological disorder level and psychological disorder level output by the medical inquiry text information processing model.

[0009] In some embodiments of the present application, based on the aforementioned scheme, the medical consultation text information processing model includes an attention network.

[0010] In some embodiments of the present application, based on the aforementioned scheme, the target medical inquiry text is subjected to data conversion processing through a word embedding model and a baseline model to obtain a first feature matrix, including: performing word segmentation processing on the target medical inquiry text to obtain multiple character units; converting each character unit into a word vector through a word embedding model, and inputting the word vector corresponding to each character unit into a baseline model, wherein the baseline model is used to associate the contextual relationship between each character unit in the target medical inquiry text; and obtaining the first feature matrix output by the baseline model.

[0011] In some embodiments of the present application, based on the aforementioned scheme, the medical inquiry text information processing model is trained through the following steps: obtaining multiple reference medical inquiry texts corresponding to multiple reference objects, wherein the reference medical inquiry texts are descriptive texts of the physical and mental states of the reference objects; matching reference key terms in each reference medical inquiry text through a terminology knowledge base of different knowledge fields, and constructing a reference domain knowledge graph of each reference medical inquiry text based on the reference key terms in different knowledge fields of each reference medical inquiry text; obtaining the physiological disorder level and psychological disorder level of the reference object corresponding to each reference medical inquiry text; and performing supervised training on a pre-constructed deep learning model framework based on each reference medical inquiry text, the reference domain knowledge graph of each reference medical inquiry text, and the physiological disorder level and psychological disorder level of the reference object corresponding to each reference medical inquiry text to obtain the medical inquiry text information processing model.

[0012] According to a second aspect of an embodiment of the present application, a device for processing medical inquiry text information is provided, characterized in that the device includes: an acquisition unit for acquiring a target medical inquiry text of a target object, wherein the target medical inquiry text is a descriptive text of the target object's physical and mental state; a matching unit for matching key terms in the target medical inquiry text through a terminology knowledge base in different knowledge fields, wherein the terminology knowledge base includes multiple professional terms belonging to its corresponding knowledge field; a construction unit for constructing a domain knowledge graph of the target medical inquiry text based on the key terms in different knowledge fields of the target medical inquiry text; and a determination unit for determining the target object's physiological disorder level and psychological disorder level through a pre-trained medical inquiry text information processing model based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text, wherein the disorder level is used to characterize the severity of the physiological disorder or the psychological disorder.

[0013] According to a third aspect of the embodiments of the present application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described in any one of the embodiments of the first aspect above.

[0014] According to a fourth aspect of the embodiments of the present application, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in any one of the embodiments of the first aspect above is implemented.

[0015] According to the fifth aspect of the embodiments of the present application, an electronic device is provided, comprising: one or more processors; a storage device for storing one or more programs, which, when the one or more programs are executed by the one or more processors, enables the one or more processors to implement the method described in any one of the embodiments of the first aspect above.

[0016] Based on the technical solution proposed in this application, the comprehensiveness and accuracy of the information processing of medical consultation texts can be improved. Specifically, on the one hand, by combining the dual perspectives of psychological disorders and physiological disorders, using the terminology knowledge base of multiple knowledge fields to match key terms in the medical consultation text, and constructing a domain knowledge graph based on the extracted key terms, it is possible to integrate the fragmented key terms in the medical consultation text, and integrate the information contained in the key terms into a whole, thereby deeply exploring the intrinsic relationship between psychological disorders and physiological disorders, and improving the comprehensiveness of the information processing of medical consultation texts. On the other hand, by inputting the domain knowledge graph and the medical consultation text into the pre-trained medical consultation text information processing model, the attention of the medical consultation text information processing model to the key terms in the medical consultation text can be enhanced, so that the medical consultation text information processing model can accurately identify the severity of the target object's physiological and psychological disorders, thereby achieving a quantitative assessment of the target object's physiological and psychological disorder level and improving the accuracy of text information processing.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, explaining the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can derive other drawings based on these drawings without inventive effort. In the drawings:

[0019] Figure 1 A flowchart showing a method for processing medical consultation text information in one embodiment of the present application is shown;

[0020] Figure 2 A schematic diagram of a domain knowledge graph in one embodiment of the present application is shown;

[0021] Figure 3 A schematic diagram illustrating the processing of medical consultation text information in one embodiment of the present application is shown;

[0022] Figure 4 A block diagram showing a device for processing medical consultation text information in one embodiment of the present application is shown;

[0023] Figure 5 A schematic structural diagram of an electronic device in one embodiment of the present application is shown. DETAILED DESCRIPTION

[0024] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0025] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner.In the following description, many specific details are provided so as to provide a full understanding of the embodiments of the present application. However, it will be appreciated by those skilled in the art that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps etc. can be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0026] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0027] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0028] It should be noted that the term "plurality" used in this document refers to two or more. "And / or" describes a relationship between associated objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A alone, A and B together, or B alone. The character " / " generally indicates an "or" relationship between the associated objects.

[0029] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, such that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described.

[0030] In the medical field, psychological disorders and physiological disorders have a relatively close interaction, that is, psychological disorders may induce physiological disorders, and physiological disorders may also induce or aggravate psychological disorders. However, traditional medical information analysis models usually analyze physiological disorders or psychological disorders from a single perspective, and are unable to combine the intrinsic relationship between psychological disorders and physiological disorders, thereby reducing the accuracy of medical information analysis. Based on this, this application proposes a method for processing medical consultation text information to improve the accuracy of medical information analysis.

[0031] Next, we will combine Figure 1 The method for processing medical consultation text information proposed in this application is described in detail.

[0032] See also Figure 1 , shows a flowchart of a method for processing medical consultation text information in one embodiment of the present application. Specifically, the method can be executed by a device with a computing and processing function. Figure 1 As shown, the method for processing medical consultation text information may include at least steps 110 to 140, which are described in detail as follows:

[0033] In step 110 , a target medical inquiry text of a target subject is obtained, where the target medical inquiry text is a description text of the target subject's physical and mental state.

[0034] In the present application, the target object may specifically be an individual who needs to undergo medical consultation information analysis, such as a patient with a chronic disease, a worker with high pressure in the workplace, or a patient in the process of postoperative recovery. This application does not make any specific limitations on this.

[0035] In the present application, the target medical consultation text can describe the physical and mental state of the target object. Specifically, it can be a patient's self-reported text, a doctor's consultation record or diagnosis opinion, or a patient's electronic medical record. This application does not make any specific restrictions on this.

[0036] In this application, obtaining a target medical consultation text containing a description of the target patient's physical and mental state helps to more fully explore and capture the target patient's physical and psychological impairments. Subsequent feature extraction, domain knowledge graph construction, and impairment level analysis based on this text not only improve the comprehensiveness and accuracy of impairment identification but also provide a richer reference for medical decision-making.

[0037] In step 120 , key terms are matched in the target medical inquiry text using terminology knowledge bases in different knowledge fields, wherein the terminology knowledge bases include a plurality of professional terms in their corresponding knowledge fields.

[0038] In the present application, the terminology knowledge base of different knowledge fields may specifically include medical terminology knowledge bases such as psychological symptom nouns, physiological symptom nouns, drug names, human tissues and organs, surgical operations, etc., and may also include terminology knowledge bases of social factors that cause physiological or psychological disorders, and may also include time description terminology knowledge bases related to the duration of physical and mental disorders. This application does not make specific limitations on this.

[0039] In the present application, the key terms are matched in the target medical consultation text. For example, in a specific text "After taking antidepressants, hand tremors and dry mouth occur, but I am worried that I will have an emotional breakdown after stopping the medicine", the drug terms "antidepressants", the physiological symptom terms "hand tremors" and "dry mouth", and the psychological state term "emotional breakdown" are matched from the above text from the term knowledge base according to different knowledge fields, and all of them can be used as key terms.

[0040] In this application, by introducing terminology knowledge bases from different knowledge fields and matching key terms in the target medical consultation text, the professionalism and accuracy of text information extraction can be significantly improved. At the same time, the terminology knowledge base has good scalability and continuous updating capabilities, and can adapt to diverse medical text content and evolving medical knowledge, thereby improving the generalization ability and long-term application value of the model.

[0041] Furthermore, the key terms are matched in the target medical inquiry text by using terminology knowledge bases in different knowledge fields, which can be specifically performed according to the following steps 121 to 122:

[0042] Step 121 , traverse the terminology knowledge base of each knowledge field, and search the target professional terminology belonging to the terminology knowledge base of each knowledge field from the target medical inquiry text.

[0043] Step 122: Determine the target professional term as the key term.

[0044] In the present application, the target professional terms belonging to the terminology knowledge base of each knowledge field are searched from the target medical inquiry text. The professional terms that are the same as those in the medical inquiry text can be determined from the terminology knowledge base of each knowledge field, or the terms with similar meanings but different expressions can be used. For example, if there is "breathing hard" in the medical inquiry text, it can be matched with the professional term "breathing" in the terminology knowledge base as one of the key terms of the medical inquiry text. For another example, if there is a description of "not being able to sleep at night" in the medical inquiry text, the professional term "not being able to sleep" can be matched in the terminology knowledge base as the key term of the medical inquiry text. For another example, if there is a statement of "not being able to eat and losing 10 pounds" in the medical inquiry text, the professional terms "not being able to eat" and "thin" can be matched in the terminology knowledge base as the key terms of the medical inquiry text.

[0045] In this application, by traversing the terminology knowledge base of various knowledge fields, searching and extracting key terms from the target medical inquiry text, not only can the completeness and accuracy of key term extraction in the medical inquiry text be improved, avoiding the omission of important medical information, but also providing a solid data foundation for subsequent medical inquiry text analysis and the construction of domain knowledge graphs.

[0046] In step 130 , a domain knowledge graph of the target medical inquiry text is constructed based on key terms in different knowledge fields of the target medical inquiry text.

[0047] In this application, the domain knowledge graph is used to represent the target object and each knowledge domain, as well as the correspondence between key terms in each knowledge domain. For details, please refer to Figure 2 , shows a schematic diagram of the domain knowledge graph in one embodiment of the present application. As shown in the figure, the domain knowledge graph displays the target object's psychological disorder "depression", the duration of the psychological disorder "3 months", the social factor "high work pressure", the physiological disorders "headache" and "insomnia", the psychological disorder medication "antidepressants", and the physiological disorder medication "painkillers" and "melatonin".

[0048] In this application, by constructing a domain knowledge graph, the scattered key terms in the target medical inquiry text can be integrated into a complete organic whole, thereby improving the accuracy of information processing by the medical inquiry text information processing model.

[0049] In step 140, based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text, the physiological disorder level and the psychological disorder level of the target object are determined through a pre-trained medical inquiry text information processing model, and the disorder level is used to characterize the severity of the physiological disorder or the psychological disorder.

[0050] In the present application, the psychological disorder level can be specifically divided into five levels, namely, the first level "none yet", which means that the target object has no psychological disorder, the second level "mild", which means that the target object has a mild psychological disorder, the third level "moderate", which means that the target object has a moderate psychological disorder and needs treatment, the fourth level "severe", which means that the target object has a severe psychological disorder and seriously affects his or her normal life, and the fifth level "dangerous", which means that the target object has a severe psychological disorder and has a tendency to harm himself or others' life; similarly, the physiological disorder level can also be specifically divided into five levels, namely, the first level "none yet", which means that the target object has no physiological disorder, the second level "mild", which means that the target object has a physiological disorder, the third level "moderate", which means that the target object has a physiological disorder that needs treatment, the fourth level "severe", which means that the target object has a physiological disorder that seriously affects his or her normal life, and the fifth level "dangerous", which means that the target object has a life-threatening physiological disorder or.

[0051] In this application, a pre-trained medical inquiry text information processing model is used to determine the level of the target patient's physical and psychological disorders by combining the target medical inquiry text with its domain knowledge graph. This model fully explores and utilizes key terms and their interrelationships within each knowledge domain within the target medical inquiry text, enabling the model to comprehensively and accurately assess the severity of psychological or physical disorders. Furthermore, the introduction of the domain knowledge graph enables the model to better capture the complex interactions between psychological and physical disorders, providing a more scientific basis for determining the level of the disorder, thereby effectively improving the accuracy of text information processing.

[0052] In the present application, the comprehensiveness and accuracy of the information processing of medical consultation texts can be improved. Specifically, on the one hand, by combining the dual perspectives of psychological disorders and physiological disorders, using the terminology knowledge base of multiple knowledge fields to match key terms in the medical consultation text, and constructing a domain knowledge graph based on the extracted key terms, the fragmented key terms in the medical consultation text can be integrated, and the information contained in the key terms can be integrated into a whole, thereby deeply exploring the intrinsic relationship between psychological disorders and physiological disorders, and improving the comprehensiveness of the information processing of medical consultation texts. On the other hand, by inputting the domain knowledge graph and the medical consultation text into the pre-trained medical consultation text information processing model, the attention of the medical consultation text information processing model to the key terms in the medical consultation text can be enhanced, so that the medical consultation text information processing model can accurately identify the severity of the target object's physical and psychological disorders, thereby achieving a quantitative assessment of the target object's physical and psychological disorder level and improving the accuracy of text information processing.

[0053] Furthermore, in the above step 140, based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text, the physiological impairment level and the psychological impairment level of the target subject are determined by using a pre-trained medical inquiry text information processing model. Specifically, the steps 141 to 143 may be performed as follows:

[0054] In step 141 , based on the target medical inquiry text, data conversion processing is performed on the target medical inquiry text through a word embedding model and a baseline model to obtain a first feature matrix.

[0055] In the present application, the word embedding model may specifically be a BERT model, and the baseline model may specifically be a BiLSTM model.

[0056] Furthermore, based on the target medical inquiry text, data conversion processing is performed on the target medical inquiry text using a word embedding model and a baseline model to obtain a first feature matrix. Specifically, the process may be performed as follows: Steps 1411 to 1413:

[0057] Step 1411 : performing word segmentation processing on the target medical inquiry text to obtain a plurality of character units.

[0058] Step 1412: Convert each character unit into a word vector through a word embedding model, and input the word vector corresponding to each character unit into a baseline model, which is used to associate the contextual relationship between each character unit in the target medical inquiry text.

[0059] Step 1413: Obtain a first feature matrix output by the baseline model.

[0060] In this application, the target medical consultation text is segmented to obtain multiple character units, specifically an OMC text consisting of a word sequence, such as M = (m1, m2, ..., m n ), n represents the length of the word sequence, m1, m2, etc. can represent a character unit. By inputting each character unit into the BERT model, we can get the word vector Z = (z1, z2, ..., z n ), and then input the word vector into the BiLSTM model. The forward LSTM in the model can learn the contextual relationship of each character unit from the beginning to the end, and then the backward LSTM can reversely learn the contextual relationship of each character unit. The two parts are combined to obtain the complete contextual relationship between each character unit, and then the first feature matrix is output.

[0061] In this application, by performing word segmentation on the target medical inquiry text, the problem of common complex expressions in medical inquiry texts can be effectively solved. For example, "palpitation and shortness of breath" can be divided into the psychological symptom of "palpitation" and the physiological symptom of "shortness of breath"; using the baseline model to establish contextual associations between character units can not only retain the long-distance dependency relationship unique to the medical inquiry text, but also enable the first feature matrix to integrate the local semantic features and global context of the medical inquiry text, thereby avoiding the lack of semantic associations between each character unit, and thus effectively improving the accuracy of the medical inquiry text information processing model in the information processing process.

[0062] In step 142, an adjacency matrix is constructed based on the domain knowledge graph of the target medical inquiry text to obtain a second feature matrix. The adjacency matrix is used to reflect the position distribution characteristics of each key term in the target medical inquiry text.

[0063] In this application, in order to enable those skilled in the art to better understand the logic of constructing an adjacency matrix based on a domain knowledge graph, a specific embodiment is described below.

[0064] For example, the target medical consultation text of the target subject is "I always feel anxious and short of breath." Among them, there are key terms "palpitation" and "shortness of breath", and their corresponding domain knowledge graphs are: target subject-psychological disorder-palpitation, target subject-physiological disorder-shortness of breath.

[0065] By segmenting the target medical consultation text, we can obtain "I", "always", "palpitation", and "shortness of breath".

[0066] Based on the above, in one embodiment, the following adjacency matrix (1) can be constructed:

[0067]

[0068] The first to fourth rows of matrix (1) correspond to "I", "always", "palpitation" and "shortness of breath", respectively, and the first to fourth columns correspond to "I", "always", "palpitation" and "shortness of breath", respectively.

[0069] Based on the above content, in another embodiment, the following adjacency matrix (2) can also be constructed:

[0070]

[0071] The first to fourth rows of matrix (2) correspond to "I", "always", "palpitation" and "shortness of breath", respectively, and the first to fourth columns correspond to "I", "always", "palpitation" and "shortness of breath", respectively.

[0072] Based on the above content, in another embodiment, the following adjacency matrix (3) can also be constructed:

[0073]

[0074] The first to fourth rows of matrix (3) correspond to "I", "always", "palpitation" and "shortness of breath", respectively, and the first to fourth columns correspond to "I", "always", "palpitation" and "shortness of breath", respectively.

[0075] In this application, an adjacency matrix is constructed based on the domain knowledge graph to reflect the position distribution characteristics of each key term in the target medical consultation text, thereby highlighting the distribution position of the key terms in the text, enhancing the expressive ability of the semantic structure, and enriching the feature information in the input model, which helps the model better understand the semantic and logical connections between key terms, thereby significantly improving the accuracy and reliability of the medical consultation text information processing model in disorder level assessment.

[0076] Step 143: Input the first feature matrix and the second feature matrix into the medical consultation text information processing model, and obtain the physiological obstacle level and the psychological obstacle level of the target object output by the medical consultation text information processing model.

[0077] In the present application, the medical consultation text information processing model includes an attention network. Specifically, inputting the first feature matrix and the second feature matrix into the medical consultation text information processing model can be embedding the first feature matrix and the second feature matrix into the attention network.

[0078] In this application, the attention network is embedded in the medical consultation text information processing model. The advantage is that the attention network can dynamically assign weights based on the importance of input features, highlighting key features that are highly relevant to the identification of physical and psychological disorders, effectively capturing different terms in the medical consultation text and their associations in the knowledge graph, thereby improving the model's ability to understand complex semantics and multi-domain information. By integrating text semantic features with knowledge graph structural features, the attention network can more accurately mine key information in the text that reflects the level of the disorder, enhance the model's discriminative and generalization capabilities, and significantly improve the accuracy and robustness of physical and psychological disorder level assessments.

[0079] In this application, please refer to Figure 3, shows a schematic diagram of medical inquiry text information processing in one embodiment of the present application. As shown in the figure, the target medical inquiry text is input into the word embedding model and the baseline model, and data conversion is performed to obtain a first feature matrix. Then, a domain knowledge graph is constructed based on the target medical inquiry text, and then a second feature matrix is constructed based on the domain knowledge graph. Then, the first feature matrix and the second feature matrix are input into the medical inquiry text information processing model, and then the medical inquiry text information processing model outputs the target object's physiological obstacle level and psychological obstacle level.

[0080] In this application, the medical inquiry text information processing model collaboratively processes the first and second feature matrices, effectively combining the semantic feature information of the target medical inquiry text with the knowledge graph information, thereby enhancing the model's focus on key terms in the target medical inquiry text. This collaborative processing effectively improves the model's information processing accuracy, enabling it to more accurately output the target subject's physical and psychological impairment levels.

[0081] In this application, the purpose of determining the target subject's level of physiological and psychological impairments is to quantify and grade the target subject's physical and mental health status, thereby providing clinicians with a scientific and intuitive reference basis to assist in formulating personalized diagnosis and treatment plans and intervention measures; at the same time, this grading result also helps in early warning of disease risks, dynamic tracking of the rehabilitation process, and rational allocation of medical resources, thereby improving the overall diagnosis and treatment efficiency and the patient's health management level.

[0082] Next, in order to enable those skilled in the art to better understand, this application will provide a detailed description of the medical consultation text information processing model.

[0083] Specifically, training the medical inquiry text information processing model may include steps 210 to 240:

[0084] Step 210 : Acquire a plurality of reference medical inquiry texts corresponding to a plurality of reference subjects, wherein the reference medical inquiry texts are description texts of the physical and mental states of the reference subjects.

[0085] Step 220, matching reference key terms in each reference medical inquiry text through terminology knowledge bases in different knowledge fields, and constructing a reference domain knowledge graph for each reference medical inquiry text based on the reference key terms in different knowledge fields of each reference medical inquiry text.

[0086] Step 230 : Obtain the physiological impairment level and the psychological impairment level of the reference subject corresponding to each reference medical inquiry text.

[0087] Step 240 , based on each reference medical inquiry text, the reference domain knowledge graph of each reference medical inquiry text, and the physiological and psychological impairment levels of the reference subjects corresponding to each reference medical inquiry text, supervised training is performed on the pre-built deep learning model framework to obtain the medical inquiry text information processing model.

[0088] In the present application, the obtaining of multiple reference medical inquiry texts corresponding to the multiple reference objects may specifically involve crawling relevant medical inquiry texts from Internet resources, for example, descriptions of users' own physical and mental states on an online medical diagnosis platform.

[0089] In this application, the physiological and psychological impairment levels of the reference subjects corresponding to each reference medical consultation text are obtained. Specifically, it can be judged manually, which can combine the clinical experience and professional knowledge of medical experts to ensure high accuracy and professionalism, especially when faced with complex or ambiguous medical consultation texts, so that more detailed and accurate judgments can be made. The flexibility of manual annotation enables it to effectively deal with semantic ambiguity and implicit information in the text, thereby reducing the possibility of misjudgment. In addition, manual annotation can also integrate multi-dimensional information, comprehensively evaluate the impairment level, and continuously optimize and standardize annotation standards in practice to further improve the accuracy of impairment level judgment.

[0090] In this application, by utilizing terminology knowledge bases from different knowledge domains to match key terms in reference medical consultation texts and constructing domain knowledge graphs, the integration of multi-domain medical and psychological knowledge is achieved, which can enhance the model's ability to understand descriptions of complex physical and mental states; at the same time, the structured domain knowledge graph can effectively mine deep semantic associations in the text, thereby enhancing the model's attention to key term features; supervised training is performed in combination with the reference objects' physiological and psychological disorder levels, so that the model can learn the mapping relationship between text features and disorder levels, thereby improving the accuracy and interpretability of the output results; in addition, collecting multiple reference objects and their corresponding text and knowledge graph information can increase the diversity of training samples, help improve the generalization ability of the model, and ultimately improve the accuracy of the model's processing of text information.

[0091] Based on the technical solution proposed in this application, the comprehensiveness and accuracy of the information processing of medical consultation texts can be improved. Specifically, on the one hand, by combining the dual perspectives of psychological disorders and physiological disorders, using the terminology knowledge base of multiple knowledge fields to match key terms in the medical consultation text, and constructing a domain knowledge graph based on the extracted key terms, it is possible to integrate the fragmented key terms in the medical consultation text, and integrate the information contained in the key terms into a whole, thereby deeply exploring the intrinsic relationship between psychological disorders and physiological disorders, and improving the comprehensiveness of the information processing of medical consultation texts. On the other hand, by inputting the domain knowledge graph and the medical consultation text into the pre-trained medical consultation text information processing model, the attention of the medical consultation text information processing model to the key terms in the medical consultation text can be enhanced, so that the medical consultation text information processing model can accurately identify the severity of the target object's physiological and psychological disorders, thereby achieving a quantitative assessment of the target object's physiological and psychological disorder level and improving the accuracy of text information processing.

[0092] The following describes an embodiment of the device of the present application, which can be used to implement the method for processing medical consultation text information in the above-mentioned embodiment of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method for processing medical consultation text information in the above-mentioned embodiment of the present application.

[0093] Figure 4 A block diagram of a device for processing medical inquiry text information in one embodiment of the present application is shown.

[0094] Reference Figure 4 As shown, according to an embodiment of the present application, a device 400 for processing medical inquiry text information includes: an acquisition unit 401 , a matching unit 402 , a construction unit 403 and a determination unit 404 .

[0095] Among them, the acquisition unit 401 is used to obtain the target medical inquiry text of the target object, and the target medical inquiry text is a descriptive text of the physical and mental state of the target object; the matching unit 402 is used to match key terms in the target medical inquiry text through a terminology knowledge base in different knowledge fields, and the terminology knowledge base includes multiple professional terms belonging to its corresponding knowledge field; the construction unit 403 is used to construct a domain knowledge graph of the target medical inquiry text based on the key terms in different knowledge fields of the target medical inquiry text; the determination unit 404 is used to determine the physiological and psychological disorder levels of the target object based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text through a pre-trained medical inquiry text information processing model, and the disorder level is used to characterize the severity of the physiological disorder or psychological disorder.

[0096] In some embodiments of the present application, based on the aforementioned scheme, the matching unit 402 is configured to: traverse the terminology knowledge base of each knowledge field, search for target professional terms belonging to the terminology knowledge base of each knowledge field from the target medical inquiry text; and determine the target professional terms as the key terms.

[0097] In some embodiments of the present application, based on the aforementioned solution, the construction unit 403 is configured as follows: the domain knowledge graph is used to represent the target object and each knowledge domain, as well as the correspondence between key terms in each knowledge domain.

[0098] In some embodiments of the present application, based on the aforementioned scheme, the determination unit 404 is configured to: based on the target medical inquiry text, perform data conversion processing on the target medical inquiry text through a word embedding model and a baseline model to obtain a first feature matrix; construct an adjacency matrix based on the domain knowledge graph of the target medical inquiry text to obtain a second feature matrix, wherein the adjacency matrix is used to reflect the position distribution characteristics of each key term in the target medical inquiry text; input the first feature matrix and the second feature matrix into the medical inquiry text information processing model, and obtain the physiological disorder level and the psychological disorder level of the target object output by the medical inquiry text information processing model.

[0099] In some embodiments of the present application, based on the aforementioned scheme, the medical consultation text information processing model includes an attention network.

[0100] In some embodiments of the present application, based on the aforementioned scheme, the determination unit 404 is further configured to: perform word segmentation processing on the target medical inquiry text to obtain multiple character units; convert each character unit into a word vector through a word embedding model, and input the word vector corresponding to each character unit into a baseline model, wherein the baseline model is used to associate the contextual relationship between each character unit in the target medical inquiry text; and obtain a first feature matrix output by the baseline model.

[0101] In some embodiments of the present application, based on the aforementioned scheme, the device further includes a training unit, which is configured to: obtain multiple reference medical inquiry texts corresponding to multiple reference objects, wherein the reference medical inquiry texts are descriptive texts of the physical and mental states of the reference objects; match reference key terms in each of the reference medical inquiry texts through a terminology knowledge base in different knowledge fields, and construct a reference domain knowledge graph of each reference medical inquiry text based on the reference key terms in different knowledge fields of each reference medical inquiry text; obtain the physiological disorder level and psychological disorder level of the reference object corresponding to each reference medical inquiry text; and perform supervised training on a pre-constructed deep learning model framework based on each reference medical inquiry text, the reference domain knowledge graph of each reference medical inquiry text, and the physiological disorder level and psychological disorder level of the reference object corresponding to each reference medical inquiry text to obtain the medical inquiry text information processing model.

[0102] As another embodiment of the present application, a computer program product or computer program is further provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method described in the above embodiment.

[0103] As another embodiment of the present application, a computer-readable storage medium is provided. The computer-readable storage medium may be included in the electronic device described in the above embodiments, or may exist independently without being incorporated into the electronic device. The computer-readable storage medium carries one or more programs. When the one or more programs are executed by the electronic device, the electronic device implements the method described in the above embodiments.

[0104] Based on the same inventive concept, the embodiment of the present application also provides an electronic device. Figure 5 , which shows a schematic diagram of the structure of an electronic device in one embodiment of the present application. The electronic device includes one or more memories 504, one or more processors 502, and at least one computer program (program code) stored in the memories 504 and executable on the processors 502. When the processors 502 execute the computer program, the aforementioned method is implemented.

[0105] Among them, Figure 5In the embodiment of the present invention, a bus architecture (represented by bus 500) is shown. Bus 500 may include any number of interconnected buses and bridges, and bus 500 links various circuits including one or more processors represented by processor 502 and memory represented by memory 504. Bus 500 may also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 505 provides an interface between bus 500 and receiver 501 and transmitter 503. Receiver 501 and transmitter 503 may be the same component, namely a transceiver, which provides a unit for communicating with various other devices over a transmission medium. Processor 502 is responsible for managing bus 500 and general processing, while memory 504 may be used to store data used by processor 502 when performing operations.

[0106] The functions described herein may be implemented in hardware, software executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, the functions may be stored as one or more instructions or codes on or transmitted via a computer-readable medium. Other examples and implementations are within the scope and spirit of this application and the appended claims. For example, due to the nature of software, the functions described above may be implemented using software executed by a processor, hardware, firmware, hardwiring, or a combination of any of these. Furthermore, the functional units may be integrated into a single processing unit, each unit may exist physically separately, or two or more units may be integrated into a single unit.

[0107] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only exemplary. For example, the division of the units can be a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0108] The units described as separate components may or may not be physically separate, and the components of the control device may or may not be physical units, that is, they may be located in one place or distributed across multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0109] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including instructions for enabling a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk.

[0110] The foregoing is merely an embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of the claims of the present application.

Claims

1. A method for processing medical consultation text information, characterized in that: The method comprises: Obtaining a target medical consultation text of a target subject, wherein the target medical consultation text is a text describing the physical and mental state of the target subject; Matching key terms in the target medical inquiry text using terminology knowledge bases in different knowledge fields, wherein the terminology knowledge bases include multiple professional terms belonging to their corresponding knowledge fields; Constructing a domain knowledge graph of the target medical inquiry text based on key terms in different knowledge fields of the target medical inquiry text; Based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text, the physiological and psychological impairment levels of the target object are determined through a pre-trained medical inquiry text information processing model, and the impairment level is used to characterize the severity of the physiological or psychological impairment.

2. The method according to claim 1, characterized in that The method of matching key terms in the target medical inquiry text using terminology knowledge bases in different knowledge fields includes: Traversing the terminology knowledge base of each knowledge field, searching for target professional terms belonging to the terminology knowledge base of each knowledge field from the target medical inquiry text; The target professional term is determined as the key term.

3. The method according to claim 1, characterized in that The domain knowledge graph is used to represent the target object and each knowledge domain, as well as the correspondence between key terms in each knowledge domain.

4. The method according to claim 1, wherein The determining of the target subject's physiological impairment level and psychological impairment level based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text by using a pre-trained medical inquiry text information processing model includes: Based on the target medical inquiry text, performing data conversion processing on the target medical inquiry text using a word embedding model and a baseline model to obtain a first feature matrix; Constructing an adjacency matrix based on the domain knowledge graph of the target medical inquiry text to obtain a second feature matrix, wherein the adjacency matrix is used to reflect the position distribution characteristics of each key term in the target medical inquiry text; The first feature matrix and the second feature matrix are input into the medical inquiry text information processing model, and the physiological obstacle level and the psychological obstacle level of the target object output by the medical inquiry text information processing model are obtained.

5. The method according to claim 4, characterized in that The medical inquiry text information processing model includes an attention network.

6. The method according to claim 4, characterized in that The target medical inquiry text is subjected to data conversion processing by using a word embedding model and a baseline model based on the target medical inquiry text to obtain a first feature matrix, including: Performing word segmentation processing on the target medical consultation text to obtain a plurality of character units; Converting each character unit into a word vector using a word embedding model, and inputting the word vector corresponding to each character unit into a baseline model, wherein the baseline model is used to associate the contextual relationship between each character unit in the target medical consultation text; A first feature matrix output by the baseline model is obtained.

7. The method according to claim 1, characterized in that The medical consultation text information processing model is trained through the following steps: Acquire a plurality of reference medical inquiry texts corresponding to a plurality of reference subjects, wherein the reference medical inquiry texts are texts describing the physical and mental conditions of the reference subjects; Matching reference key terms in each reference medical inquiry text through terminology knowledge bases in different knowledge fields, and constructing a reference domain knowledge graph for each reference medical inquiry text based on the reference key terms in different knowledge fields of each reference medical inquiry text; Obtaining the physiological impairment level and the psychological impairment level of the reference subject corresponding to each reference medical consultation text; Based on each reference medical inquiry text, the reference domain knowledge graph of each reference medical inquiry text, and the physiological and psychological impairment levels of the reference objects corresponding to each reference medical inquiry text, a pre-constructed deep learning model framework is supervised trained to obtain the medical inquiry text information processing model.

8. A device for processing medical consultation text information, characterized in that: The device comprises: an acquiring unit, configured to acquire a target medical consultation text of a target subject, wherein the target medical consultation text is a description text of the target subject's physical and mental state; a matching unit, configured to match key terms in the target medical inquiry text using terminology knowledge bases in different knowledge fields, wherein the terminology knowledge bases include a plurality of professional terms belonging to their corresponding knowledge fields; A construction unit, configured to construct a domain knowledge graph of the target medical inquiry text based on key terms in different knowledge fields of the target medical inquiry text; A determination unit is used to determine the physiological and psychological impairment levels of the target subject based on the target medical inquiry text and the domain knowledge graph of the target medical inquiry text through a pre-trained medical inquiry text information processing model, wherein the impairment level is used to characterize the severity of the physiological or psychological impairment.

9. A computer program product, characterized in that The computer program product includes computer instructions stored in a computer-readable storage medium and adapted to be read and executed by a processor, so as to enable a computer device having the processor to perform the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: The electronic device includes one or more processors and one or more memories, wherein at least one program code is stored in the one or more memories, and the at least one program code is loaded and executed by the one or more processors to implement the method according to any one of claims 1 to 7.

Images