Medical term quality monitoring method and related device, electronic device, and storage medium

Abstract

The application discloses a medical term quality monitoring method and related device, electronic equipment and storage medium, wherein the medical term quality monitoring method comprises the following steps: obtaining an initial term library; based on a term element template, respectively performing term splitting on standard terms and synonymous terms in a medical term pair, and standardizing split words to obtain a split standardized result of the medical term pair; the split standardized result comprises a first split standardized combination of the standard terms and a second split standardized combination of the synonymous terms, and the split standardized combination comprises standard words obtained by standardizing split words corresponding to each term element after splitting the term pair according to the term element template; and based on the split standardized result of each medical term pair in the initial term library, performing quality inspection on the initial term library to obtain a target term library. The above scheme can improve the efficiency of term quality inspection as much as possible, thereby improving the quality of term resources.

Description

Technical Field

[0001] This application relates to the field of smart healthcare technology, and in particular to a method for monitoring the quality of medical terminology and related devices, electronic equipment, and storage media. Background Technology

[0002] With the development of artificial intelligence, the level of informatization in the medical field is constantly improving. However, in the context of big data, different medical institutions, and even different medical personnel, may use different expressions for the same surgeries, examinations, tests, and symptoms. This not only increases the difficulty for machines to understand medical information but also makes it difficult to guarantee the quality of medical terminology resources.

[0003] Currently, terminology quality monitoring mainly relies on medical professionals. However, due to the large scale of this process, it requires significant manpower, and different individuals may have varying understandings of the standards. Even the same person may apply different standards at different times, making it impossible to fundamentally guarantee terminology quality. Therefore, how to maximize the efficiency of terminology quality control and thereby improve the quality of terminology resources has become an urgent problem to be solved. Summary of the Invention

[0004] The main technical problem addressed by this application is to provide a method and related devices, electronic equipment, and storage media for monitoring the quality of medical terminology, which can maximize the efficiency of terminology quality inspection and thus improve the quality of terminology resources.

[0005] To address the aforementioned technical problems, the first aspect of this application provides a method for quality monitoring of medical terminology, comprising: obtaining an initial terminology database; wherein the initial terminology database includes several pairs of medical terminology, and each medical terminology pair includes standard terms and synonyms of the standard terms; further, based on a terminology element template, the standard terms and synonyms in the medical terminology pairs are split into terms, and the split terms are standardized to obtain the splitting and standardization results of the medical terminology pairs; wherein the splitting and standardization results include a first splitting and standardization combination of standard terms and a second splitting and standardization combination of synonyms, and the splitting and standardization combination includes the standardized terms of the split terms corresponding to each terminology element after splitting by the terminology reference terminology element template; and performing quality control on the initial terminology database based on the splitting and standardization results of each pair of medical terminology pairs in the initial terminology database to obtain a target terminology database.

[0006] To address the aforementioned technical problems, a second aspect of this application provides a medical terminology quality monitoring device, comprising: a terminology database acquisition module, a splitting and standardization module, and a terminology database quality inspection module. The terminology database acquisition module acquires an initial terminology database, which includes several pairs of medical terminology pairs, each pair containing standard terms and synonyms of the standard terms. The splitting and standardization module performs term splitting and word standardization on the standard terms and synonyms in the medical terminology pairs based on terminology element templates, obtaining the splitting and standardization results for the medical terminology pairs. These results include a first splitting and standardization combination of standard terms and a second splitting and standardization combination of synonyms, whereby the splitting and standardization combinations include the standardized words corresponding to each terminology element after splitting from the terminology element template. The terminology database quality inspection module performs quality inspection on the initial terminology database based on the splitting and standardization results of each pair of medical terminology pairs, obtaining a target terminology database.

[0007] To address the aforementioned technical problems, a third aspect of this application provides an electronic device including a memory and a processor coupled to each other. The memory stores program instructions, and the processor executes the program instructions to implement the medical terminology quality monitoring method of the first aspect described above.

[0008] To address the aforementioned technical problems, a fourth aspect of this application provides a computer-readable storage medium storing program instructions executable by a processor, the program instructions being used to implement the medical terminology quality monitoring method of the first aspect described above.

[0009] The above scheme obtains an initial terminology database, which includes several pairs of medical terminology, each pair containing standard terms and their synonyms. Then, based on terminology element templates, the standard terms and synonyms within each medical terminology pair are split, and the split terms are standardized to obtain the split-standardized results of the medical terminology pairs. These results include a first standardized combination of standard terms and a second standardized combination of synonyms. The standardized combinations include the standardized terms corresponding to the split terms in the terminology element template after splitting. Based on this, the initial terminology database is quality-checked using the split-standardized results of each pair of medical terminology pairs, resulting in the target terminology database. On one hand, by splitting and standardizing standard terms and synonyms within each medical terminology pair based on terminology element templates, fine-grained terminology quality checking can be performed at the element level, improving the accuracy of terminology quality checking. On the other hand, by performing quality checking on the initial terminology database based on the split-standardized results of each pair of medical terminology pairs, without relying on manual quality checking, the efficiency of terminology quality checking is improved. Therefore, it is possible to improve the efficiency of terminology quality inspection as much as possible, thereby improving the quality of terminology resources.

[0010] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this application. Attached Figure Description

[0011] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with this application and, together with the specification, serve to explain the technical solutions of this application.

[0012] Figure 1 This is a flowchart illustrating an embodiment of the medical terminology quality control method of this application;

[0013] Figure 2 This is a schematic diagram of an embodiment of the character identifier for each character;

[0014] Figure 3 yes Figure 1 A flowchart illustrating an embodiment of step S31;

[0015] Figure 4 yes Figure 1 A flowchart illustrating another embodiment of step S31;

[0016] Figure 5 This is a schematic diagram of an embodiment of the standard terminology mapping model;

[0017] Figure 6 This is a schematic diagram of the framework of an embodiment of the medical terminology quality monitoring device of this application;

[0018] Figure 7 This is a schematic diagram of the framework of an embodiment of the electronic device of this application;

[0019] Figure 8 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of this application. Detailed Implementation

[0020] The embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0021] In the following description, specific details such as particular system architectures, interfaces, and technologies are presented for illustrative purposes rather than for limiting purposes, in order to provide a thorough understanding of this application.

[0022] In this document, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. Additionally, the character " / " generally indicates that the preceding and following related objects are in an "or" relationship. Furthermore, "many" in this document means two or more. Moreover, the term "at least one" in this document means any combination of at least two of any one or more of a plurality of objects. For example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C. "Several" means at least one. The terms "first," "second," etc., in the specification, claims, and accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0023] In this embodiment, the initial terminology database is constructed by normalizing the expressions of medical terms and converting medical data with different expressions into standard medical terms. Furthermore, the construction of the initial terminology database relies on professionals with medical backgrounds and clinical experience, and also involves training models using big data to statistically analyze similarity or matching based on rules, followed by manual verification of the output results. However, due to differences in work experience among doctors, it is difficult to ensure uniformity of standards across different doctors, and the mapping results also contain errors, making it difficult to ensure consistent standards across different doctors at different times. Therefore, the construction of the initial terminology database results in a large number of synonyms, similar words, and even incorrectly mapped data, seriously affecting the quality of the database. In view of this, further quality checks are needed for the initial terminology database to improve the quality of the terminology resources.

[0024] Please see Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the medical terminology quality control method of this application. Specifically, it may include the following steps:

[0025] Step S11: Obtain the initial terminology database.

[0026] In this embodiment, the initial terminology database includes several pairs of medical terms, each pair comprising a standard term and its synonyms. It should be noted that synonyms of standard terms can be obtained through term matching based on medical knowledge. For example, they can be obtained by professionals matching and dividing the various medical terms within a terminology set (e.g., which may include several symptom terms, several disease terms, etc.), or they can be obtained based on a trained model. During the matching process, mapping errors may occur in medical terminology pairs. That is, synonyms of standard terms are medical terms that may have the same meaning as the standard terms; that is, synonyms of standard terms may have different or the same meaning representation than the standard terms. Therefore, it is necessary to further quality control the medical terminology pairs to improve the quality of the terminology resources. Please refer to Table 1 for an example. Table 1 is a schematic table of several pairs of medical terms in the initial terminology database, as shown in Table 1. "Left upper limb itching" and "left side upper limb itching" form a medical term pair, with "left upper limb itching" being the standard term and "left side upper limb itching" being a synonym. Alternatively, "wrist pain" and "arm pain" form a medical term pair, with "wrist pain" being the standard term and "arm pain" being a suspected synonym with a different meaning. It is understood that the illustrated method is only one possible scenario in the initial terminology database in actual application and does not limit the content of the initial terminology database in actual application. The initial terminology database can be constructed according to actual circumstances, and no specific limitations are made here.

[0027] Table 1. A schematic diagram of several pairs of medical terms in the initial terminology database for one embodiment.

[0028] Standard Terminology Synonyms of standard terminology itching in the left upper limb itching in the left upper limb Itchy skin on left upper limb Itchy skin on the left upper arm Upper limb pain Upper limb pain Arm pain Arm pain wrist pain Arm pain

[0029] In a given implementation scenario, a medical terminology pair may include surgical terms, examination terms, laboratory terms, symptom terms, etc. It may also selectively include symptom and laboratory terms. Of course, a medical terminology pair may also only include symptom terms. The content of the terms included in a medical terminology pair can be selected based on the actual situation, and no specific limitations are imposed here.

[0030] Step S12: Based on the terminology element template, the standard terms and synonyms in the medical terminology pair are split into terms, and the split terms are standardized to obtain the split and standardized results of the medical terminology pair.

[0031] In one implementation scenario, the medical terminology pair involves at least symptom terms, and the terminology element template includes at least symptom terminology templates applicable to the symptom terms. Each terminology element constituting the symptom terminology template includes location, primary location, secondary location, and symptom. It is understood that, based on the terminology element template, the standard terms and synonyms in the medical terminology pair can be split at the finest granularity, and the resulting tuples can be categorized and unified under a fixed module. For example, symptom terms are determined by finely splitting standard words or synonyms in medical terms to establish the symptom terminology template. For example, please refer to Table 2, which is a schematic table of an embodiment of the symptom terminology template. As shown in Table 2, when the symptom term is "itching of the left upper limb," the terminology elements include "left side," "upper limb," "skin," and "itching." It is understood that when the medical terminology pair involves surgical terms, examination terms, or laboratory terms, the terminology element template includes terminology templates applicable to the corresponding terms, which will not be elaborated further here.

[0032] Table 2. Schematic diagram of an embodiment of the symptom terminology template.

[0033] Symptom terminology position Main part Sub-area symptom Itching of the skin on the left upper limb Left side upper limbs skin itching

[0034] In this embodiment of the disclosure, the split standardization result includes a first split standardization combination of standard terms and a second split standardization combination of synonyms. The split standardization combination includes the standard words obtained after standardizing the split words corresponding to each term element in the term reference template. Specifically, the split words obtained after splitting the standard terms and synonyms from the term element template are first obtained, and then the aforementioned split words are standardized to obtain standard words. For example, please refer to Table 3, which is a schematic table of an embodiment of the split standardization result. As shown in Table 3, the standard term is "left upper limb itching," and the synonym is "left upper limb skin itching." The split standardization result includes a first split standardization combination of standard terms and a second split standardization combination of synonyms. The first split standardization combination includes the standard words for "left upper limb itching," and the second split standardization combination includes the standard words for "left upper limb skin itching."

[0035] Table 3. Schematic diagram of an example of the decomposition standardization results.

[0036]

[0037] In one implementation scenario, term segmentation can begin by encoding the term to be processed, obtaining the encoded representation of each character within the term. It should be noted that the term to be processed is selected from pairs of medical terms. For example, referring to Table 1, based on several pairs of medical terms, the terms to be processed are selected as "left upper limb itching" and "left side upper limb itching." Furthermore, the term to be processed can be encoded using pre-trained BERT series models, such as XLNet and Span Bert network models, to obtain the encoded representation of each character within the term. Then, prediction is performed based on the encoded representation of each character to obtain the character identifier of each character; and the character identifier is used to characterize that the character belongs to any of the following situations: the character alone constitutes a segmented word, the character constitutes the starting character of a segmented word, or the character constitutes the ending character of a segmented word. Based on this, using the term element template and the character identifiers of each character, the segmented words corresponding to each term element after segmentation of the term to be processed according to the term element template are obtained. The above method, by encoding each character in the term to be processed and then predicting the encoded representation of each character, helps to improve the accuracy of the prediction results. Furthermore, based on the term element template and the character identifier of each character, it further improves the accuracy of the segmentation of the term to be processed.

[0038] Please see Figure 2 , Figure 2 This is a schematic diagram of an embodiment of the character identifier for each character. For example... Figure 2 As shown, the term to be processed is "itching of the skin on the left upper limb." First, each character in the term is encoded to obtain its coded representation. Then, prediction is performed based on the coded representations of each character to obtain its character identifier. The character identifier is used to represent that the character belongs to any of the following categories: the character alone constitutes a split word, the character constitutes the starting character of a split word, or the character constitutes the ending character of a split word. Specifically, a character alone constituting a split word can be represented by BE (Begin, End), a character constituting the starting character of a split word can be represented by BO (Begin, Other), and a character constituting the ending character of a split word can be represented by OE (Other, End). It is understood that in the term to be processed, "left" can constitute a split word alone, "upper" can constitute the starting character of a split word, "limb" can constitute the ending character of a split word, and so on for other characters. Furthermore, based on the term element template and the character identifiers of each character, the split words corresponding to each term element after the term to be processed is split according to the term element template are obtained. The split words can be found in Table 3, and will not be elaborated further here.

[0039] In a specific implementation scenario, term segmentation is performed by a term segmentation model, which includes an encoding network and a prediction network. The encoding network encodes the terms to be processed selected from medical term pairs, obtaining the encoded representation of each character in the term. The encoding network can be a pre-trained BERT series model, such as XLNet, Span Bert, etc. Alternatively, the term can be encoded based on a pre-trained BERT series model, such as XLNet, Span Bert, etc., to obtain the character encoding features of each character in the term. Then, a neural network model processes the character encoding features of each character to obtain the encoded representation of each character. The neural network model can include, but is not limited to, LSTM (Long Short-Term Memory) and GRU (Gate Recurrent Unit). Furthermore, the prediction network predicts the encoded representation of each character, obtaining the character identifier of each character. The prediction network can include, but is not limited to, BP (Multi-layer Feedforward) network and CNN (Convolutional Neural Network).

[0040] Step S13: Perform quality checks on the initial terminology database based on the splitting and standardization results of each pair of medical terms in the initial terminology database to obtain the target terminology database.

[0041] In one implementation scenario, to obtain the target terminology database, a first quality check can be performed on the initial terminology database at the level of term expression, based on the decomposition and standardization results of each pair of medical terms in the initial terminology database. The medical terminology pairs that are correctly represented in the initial terminology database are then identified as the first terminology pairs. The target terminology database is then directly constructed based on these first terminology pairs.

[0042] In another implementation scenario, to improve the accuracy of the target terminology database, a first quality check can be performed on the initial terminology database based on the decomposition and standardization results of each pair of medical terms in the initial terminology database. The medical terminology pairs that are represented without error in the initial terminology database are identified as the first terminology pairs. Then, a second quality check is performed on the first terminology pairs in the initial terminology database at the level of terminology expression content based on a standard terminology mapping model, resulting in third terminology pairs. Based on these third terminology pairs, the target terminology database is constructed.

[0043] In another implementation scenario, to improve the accuracy of the target terminology database and ensure the efficiency of its acquisition, a first quality check can be performed on the initial terminology database based on the decomposition and standardization results of each pair of medical terminology pairs. This identifies the error-free medical terminology pairs as the first terminology pairs. A second quality check is then performed on the second terminology pairs (excluding the first terminology pairs) based on a standard terminology mapping model, yielding the third terminology pairs. Based on these, the target terminology database is constructed using the first and third terminology pairs. Specifically, several pairs of medical terminology pairs are constructed, including standard terms and their synonyms. This approach, through quality checks at two levels (i.e., the terminology expression form level and the terminology expression content level), helps to maximize the efficiency of terminology quality checks, thereby improving the quality of terminology resources.

[0044] In a specific implementation scenario, the first term pair is a pair of medical terms that are correctly represented in the initial terminology database. The first term pair can be obtained by evaluating the initial terminology database. For details, please refer to [link to relevant documentation]. Figure 3 , Figure 3 yes Figure 1 A flowchart illustrating an embodiment of step S31. Specifically, it may include the following steps:

[0045] Step S31: Obtain the splitting and standardization results of medical term pairs.

[0046] In this embodiment of the disclosure, the method for obtaining the splitting and standardization results of medical terminology pairs can refer to the method described in the foregoing embodiments of the disclosure, and will not be repeated here.

[0047] In one implementation scenario, the decomposition standardization result includes the first decomposition standardization combination of standard terms and the second decomposition standardization combination of synonyms. The decomposition standardization combination includes the standard terms after the term reference term element template is decomposed and the decomposed words corresponding to each term element are standardized.

[0048] Step S32: Determine whether the first standardized combination of the standard term and the second standardized combination of the synonym in the medical terminology pair are completely consistent; if not, proceed to step S33; otherwise, proceed to step S34.

[0049] It is understandable that in the process of standardizing the split words to obtain the standardized results of medical term pairs, it cannot be guaranteed that the first standardized combination of the standard term and the second standardized combination of the synonym in the medical term pair will be completely consistent. For example, please refer to Table 3. In Table 3, the first standardized combination of the standard term "left upper limb itching" and the second standardized combination of the synonym "left upper limb skin itching" are not completely consistent. The method shown is only one possible situation in actual application and does not limit the standardized results of medical terms in actual application. The standardized results of medical term splitting can be determined according to the actual situation, and no specific limitation is made here.

[0050] Step S33: Use the medical term pair as the second term pair.

[0051] In one implementation scenario, in response to the fact that the first split-standardized combination of the standard terms and the second split-standardized combination of the synonyms in a medical term pair are not completely consistent, the medical term pair can be regarded as the second term pair.

[0052] Step S34: Determine whether the synonyms in the medical terminology pair have only one unique standard term in the initial terminology database; if not, proceed to step S35; otherwise, proceed to step S36.

[0053] Understandably, the initial terminology database cannot guarantee that each medical term pair has only one corresponding standard term. When a synonym corresponds to multiple standard terms, the current medical term pair is incorrect. For example, the synonym "itching skin on the left upper limb" may have two corresponding labeled terms: "itching skin on the left upper limb" and "itching skin on the right upper limb." Understandably, the current medical term pair is incorrect.

[0054] Step S35: Use the medical term pair as the second term pair.

[0055] In one implementation scenario, in response to the fact that synonyms in a medical terminology pair have multiple corresponding standard terms in the initial terminology database, the medical terminology pair can be designated as a second terminology pair. It is understood that incorrectly represented medical terminology pairs in the initial terminology database are designated as second terminology pairs.

[0056] Step S36: Determine the medical term pair as the first term pair that is accurate.

[0057] In one implementation scenario, in response to the complete consistency between the first standardized combination of the standard term and the second standardized combination of the synonym in a medical terminology pair, it can be checked whether the synonym in the medical terminology pair corresponds to only one unique standard term in the initial terminology database. And in response to the fact that the synonym in the medical terminology pair corresponds to only one unique standard term in the initial terminology database, the medical terminology pair is determined to be the first terminology pair with accurate representation. This method, by judging the medical terminology pairs in the initial terminology database, further improves the quality of terminology resources by identifying the first terminology pair with accurate representation and the second terminology pair with incorrect representation.

[0058] Please see Figure 4 , Figure 4 yes Figure 1 A flowchart illustrating another embodiment of step S31. Specifically, it may include the following steps:

[0059] Step S41: Obtain the second term pair.

[0060] Specifically, the method for obtaining the second term pair can be referred to in the aforementioned disclosed embodiments, which will not be repeated here.

[0061] Step S42: Based on the standard term mapping model, decode the splitting results, reference characters, and reference states of synonyms in the second term pair to obtain the decoded characters corresponding to this decoding.

[0062] In this embodiment, the splitting result includes the split words corresponding to each term element after the synonym term element template is split, the reference character is the decoded character obtained before this decoding, and the reference state is the hidden state of the standard term mapping model before this decoding. Furthermore, the standard term mapping model can be a pre-trained transformer series model, and the standard term mapping model can be selected according to the actual situation; no specific limitation is made here.

[0063] Please see Figure 5 , Figure 5 This is a schematic diagram of an embodiment of the standard terminology mapping model, such as... Figure 5As shown, the standard term mapping model mainly includes feed-forward, encoder-decoder attention layer, residual connections & normalization (Add & Normalize), and masked self-attention mechanism. Feed-forward is used to pre-monitor interference and prevent it from disrupting the signal. In the encoder-decoder attention layer, the encoder embeds the input vector, and the decoder decodes the vector to obtain the output. In residual connections & normalization, Add is a type of residual connection, typically used to solve problems in training multi-layer networks. Normalize transforms the input of each neuron in each layer into values ​​with the same mean and variance, thus accelerating convergence. The masked self-attention mechanism can adjust the data dimension. Furthermore, based on the standard term mapping model, the splitting results, reference characters, and reference states of synonyms in the second term pair are decoded to obtain the decoded characters corresponding to this decoding. This can be represented by an expression, as follows:

[0064] out i =f(enc out ,y i ,q i-1 )

[0065] Among them, out i The decoded character corresponding to this decoding, enc out The result of splitting synonyms in the second term pair, y i For reference character, q i-1 This serves as a reference state. Specifically, after the decoder acquires the input, it first goes through a masked self-attention mechanism to obtain the correlation between the current input and the previous inputs. Then, the vector undergoes batch regularization to effectively mitigate generalization error.

[0066] Step S43: Combine the decoded characters corresponding to each decoding to obtain the mapped terms after standardized mapping of synonyms.

[0067] In one implementation scenario, each decoding operation yields a corresponding decoded character. Combining these decoded characters from each decoding operation results in a standardized mapping terminology for synonyms. For an example, please refer to [link to example]. Figure 5 The second terminology is synonymous with "left upper limb pruritus," and the standardized mapped term is "left upper limb pruritus."

[0068] Step S44: Determine whether the mapping terms corresponding to the standard term and the synonym in the second term pair are consistent; if yes, proceed to step S45; otherwise, proceed to step S46.

[0069] In one implementation scenario, the third term pair after quality inspection of the second term pair is determined based on whether the mapping terms corresponding to the standard term and the synonym in the second term pair are consistent.

[0070] Step S45: Directly identify the second term pair as the third term pair.

[0071] In one implementation scenario, the second term pair can be directly identified as the third term pair in response to the consistency of the mapping terms corresponding to the standard term and the synonym in the second term pair.

[0072] Step S46: Based on the correction feedback of the second term pair, obtain the third term pair.

[0073] In one implementation scenario, in response to the inconsistency between the mapped terms corresponding to the standard terms and synonyms in the second term pair, a third term pair is obtained based on the correction feedback of the second term pair. Specifically, the correction feedback can be determined directly based on the output of the standard term mapping model, that is, the synonyms in the second term pair and the mapped terms after standardization mapping through the standard term mapping model are determined as the third term pair. For example, please refer to [link to relevant documentation]. Figure 5 In the second terminology pair, the synonym "itching of the skin on the left upper limb" and the standardized mapped term "itching of the skin on the left upper limb" can be directly identified as the third terminology pair. Alternatively, the standardized mapped term can be fed back to a professional for evaluation and correction to obtain the third terminology pair. The method of feedback and correction can be chosen based on the actual situation and is not specifically limited here.

[0074] Please refer to Table 4, which is an illustration of an embodiment of several pairs of medical terms in the target terminology database. As shown in Table 4, by performing the finest granular segmentation of medical terms and then conducting the first and second quality checks, the quality of terminology resource construction can be greatly improved, which helps to improve its application in the medical record structuring scenario and can maximize the efficiency of terminology quality checks, thereby improving the quality of terminology resources.

[0075] Table 4. A schematic diagram of several pairs of medical terms in the target terminology database for one embodiment.

[0076]

[0077]

[0078] The above scheme obtains an initial terminology database, which includes several pairs of medical terminology, each pair containing standard terms and their synonyms. Then, based on terminology element templates, the standard terms and synonyms within each medical terminology pair are split, and the split terms are standardized to obtain the split-standardized results of the medical terminology pairs. These results include a first standardized combination of standard terms and a second standardized combination of synonyms. The standardized combinations include the standardized terms corresponding to the split terms in the terminology element template after splitting. Based on this, the initial terminology database is quality-checked using the split-standardized results of each pair of medical terminology pairs, resulting in the target terminology database. On one hand, by splitting and standardizing standard terms and synonyms within each medical terminology pair based on terminology element templates, fine-grained terminology quality checking can be performed at the element level, improving the accuracy of terminology quality checking. On the other hand, by performing quality checking on the initial terminology database based on the split-standardized results of each pair of medical terminology pairs, without relying on manual quality checking, the efficiency of terminology quality checking is improved. Therefore, it is possible to improve the efficiency of terminology quality inspection as much as possible, thereby improving the quality of terminology resources.

[0079] Those skilled in the art will understand that, in the above-described method of the specific implementation, the order in which each step is written does not imply a strict execution order and does not constitute any limitation on the implementation process. The specific execution order of each step should be determined by its function and possible internal logic.

[0080] Please see Figure 6 , Figure 6 This is a schematic diagram of an embodiment of the medical terminology quality monitoring device of this application. The medical terminology quality monitoring device 60 includes: a terminology database acquisition module 61, a splitting and standardization module 62, and a terminology database quality inspection module 63. The terminology database acquisition module 61 is used to acquire an initial terminology database, which includes several pairs of medical terminology pairs, and each medical terminology pair includes standard terms and synonyms of the standard terms. The splitting and standardization module 62 is used to perform term splitting and word standardization on the standard terms and synonyms in the medical terminology pairs based on terminology element templates, obtaining the splitting and standardization results of the medical terminology pairs. The splitting and standardization results include a first splitting and standardization combination of standard terms and a second splitting and standardization combination of synonyms. The splitting and standardization combination includes the standardized words corresponding to the splitting words of each terminology element after splitting the terminology reference terminology element template. The terminology database quality inspection module 63 is used to perform quality inspection on the initial terminology database based on the splitting and standardization results of each pair of medical terminology pairs in the initial terminology database, obtaining a target terminology database.

[0081] The above-described scheme, on the one hand, performs terminology segmentation and standardization of standard and synonymous terms in medical terminology pairs based on terminology element templates, enabling fine-grained terminology quality control at the element level and improving the accuracy of terminology quality control. On the other hand, it performs quality control on the initial terminology database based on the segmentation and standardization results of each pair of medical terminology pairs, eliminating the need for manual quality control and improving the efficiency of terminology quality control. Therefore, it can maximize the efficiency of terminology quality control, thereby improving the quality of terminology resources.

[0082] In some disclosed embodiments, the terminology database quality inspection module 63 includes a determination submodule, which is used to perform a first quality inspection on the initial terminology database based on the splitting and standardization results of each pair of medical terminology pairs in the initial terminology database, and determine the medical terminology pairs in the initial terminology database that are represented without error as the first terminology pair; the terminology database quality inspection module 63 includes a quality inspection submodule, which is used to perform a second quality inspection on the second terminology pairs other than the first terminology pairs in the initial terminology database based on a standard terminology mapping model, and obtain a third terminology pair; the terminology database quality inspection module 63 also includes a construction submodule, which is used to construct a target terminology database based on the first terminology pair and the third terminology pair.

[0083] Therefore, conducting quality inspections at both the first and second levels helps to maximize the efficiency of terminology quality inspection, thereby improving the quality of terminology resources.

[0084] In some disclosed embodiments, the determining submodule includes a first response unit, which is configured to check whether the synonyms in the medical terminology pair have only one unique standard term in the initial terminology database in response to the complete consistency between the first split-standard combination of the standard term and the second split-standard combination of the synonym in the medical terminology pair; the determining submodule also includes a second response unit, which is configured to determine the medical terminology pair as an error-free first terminology pair in response to the fact that the synonyms in the medical terminology pair have only one unique standard term in the initial terminology database.

[0085] In some disclosed embodiments, the determining submodule includes a third response unit, which is used to determine the medical term pair as a second term pair in response to the fact that the first split-standardized combination of the standard term and the second split-standardized combination of the synonym in the medical term pair are not completely consistent; the determining submodule includes a fourth response unit, which is used to determine the medical term pair as a second term pair in response to the fact that the synonym in the medical term pair corresponds to multiple standard terms in the initial terminology database.

[0086] Therefore, by judging the medical term pairs in the initial terminology database, the first term pair that is correct and the second term pair that is incorrect can be determined, thereby further improving the quality of terminology resources.

[0087] In some disclosed embodiments, the quality inspection submodule includes a decoding unit, which is used to decode the splitting results, reference characters, and reference states of synonyms in the second term pair based on a standard term mapping model to obtain the decoded characters corresponding to this decoding. The splitting results include the split words corresponding to each term element after the synonym comparison term element template is split, the reference characters are the decoded characters obtained before this decoding, and the reference states are the hidden state of the standard term mapping model before this decoding. The quality inspection submodule includes a combination unit, which is used to combine the decoded characters corresponding to each previous decoding to obtain the mapped terms after the standardized mapping of synonyms. The quality inspection submodule also includes a determination unit, which is used to determine the third term pair after the quality inspection of the second term pair based on whether the mapped terms corresponding to the standard terms and synonyms in the second term pair are consistent.

[0088] In some disclosed embodiments, the determining unit includes a first response subunit, which is used to directly determine the second term pair as a third term pair in response to the consistency between the mapping terms corresponding to the standard term and the synonym in the second term pair; the determining unit also includes a second response subunit, which is used to obtain the third term pair based on the correction feedback of the second term pair in response to the inconsistency between the mapping terms corresponding to the standard term and the synonym in the second term pair.

[0089] In some disclosed embodiments, the splitting standardization module 62 includes an encoding submodule, which is used to encode based on the term to be processed to obtain the encoded representation of each character in the term to be processed, and the term to be processed is selected from the medical term pair; the splitting standardization module 62 includes a prediction submodule, which is used to predict based on the encoded representation of each character to obtain the character identifier of each character, and the character identifier of the character is used to characterize that the character belongs to any of the following: the character alone constitutes a split word, the character constitutes the starting character of a split word, or the character constitutes the ending character of a split word; the splitting standardization module 62 also includes an acquisition submodule, which is used to obtain the split words corresponding to each term element after the term to be processed is split according to the term element template based on the term element template and the character identifier of each character.

[0090] Therefore, by encoding each character in the term to be processed and then predicting the encoded representation of each character, the accuracy of the prediction results can be improved. Furthermore, based on the term element template and the character identifier of each character, the accuracy of the segmentation of the term to be processed can be further improved.

[0091] In some disclosed embodiments, term segmentation is performed by a term segmentation model, which includes an encoding network and a prediction network. The encoding network is used to encode the terms to be processed selected from medical term pairs to obtain the encoded representation of each character in the term to be processed. The prediction network is used to predict the encoded representation of each character to obtain the character identifier of each character.

[0092] In some disclosed embodiments, the medical terminology pair involves at least symptom terms, and the terminology element template includes at least a symptom terminology template applicable to symptom terms. The various terminology elements constituting the symptom terminology template include location, primary location, secondary location, and symptom.

[0093] Please see Figure 7 , Figure 7 This is a schematic diagram of a framework of an embodiment of the electronic device of this application. The electronic device 70 includes a memory 71 and a processor 72 coupled to each other. The memory 71 stores program instructions, and the processor 72 is used to execute the program instructions to implement the steps in any of the above embodiments of the medical terminology quality monitoring method. Specifically, the electronic device 70 may include, but is not limited to, desktop computers, laptops, servers, mobile phones, tablets, etc., and is not limited thereto.

[0094] Specifically, processor 72 controls itself and memory 71 to implement the steps in any of the above-described embodiments of the medical terminology quality monitoring method. Processor 72 may also be referred to as a CPU (Central Processing Unit). Processor 72 may be an integrated circuit chip with signal processing capabilities. Processor 72 may also be a general-purpose processor, digital signal processor (DSP), application-specific integrated circuit (ASIC), field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. A general-purpose processor may be a microprocessor or any conventional processor. Furthermore, processor 72 may be implemented using integrated circuit chips.

[0095] The above-described scheme, on the one hand, performs terminology segmentation and standardization of standard and synonymous terms in medical terminology pairs based on terminology element templates, enabling fine-grained terminology quality control at the element level and improving the accuracy of terminology quality control. On the other hand, it performs quality control on the initial terminology database based on the segmentation and standardization results of each pair of medical terminology pairs, eliminating the need for manual quality control and improving the efficiency of terminology quality control. Therefore, it can maximize the efficiency of terminology quality control, thereby improving the quality of terminology resources.

[0096] Please see Figure 8 , Figure 8 This is a schematic diagram of a framework of an embodiment of the computer-readable storage medium of this application. The computer-readable storage medium 80 stores program instructions 81 that can be executed by a processor. The program instructions 81 are used to implement the steps in any of the above embodiments of the medical terminology quality monitoring method.

[0097] The above-described scheme, on the one hand, performs terminology segmentation and standardization of standard and synonymous terms in medical terminology pairs based on terminology element templates, enabling fine-grained terminology quality control at the element level and improving the accuracy of terminology quality control. On the other hand, it performs quality control on the initial terminology database based on the segmentation and standardization results of each pair of medical terminology pairs, eliminating the need for manual quality control and improving the efficiency of terminology quality control. Therefore, it can maximize the efficiency of terminology quality control, thereby improving the quality of terminology resources.

[0098] In some embodiments, the functions or modules of the apparatus provided in this disclosure can be used to perform the methods described in the above method embodiments. The specific implementation can be referred to the description of the above method embodiments, and for the sake of brevity, it will not be repeated here.

[0099] The description of the various embodiments above tends to emphasize the differences between the various embodiments. The similarities or similarities between them can be referred to, and for the sake of brevity, they will not be repeated here.

[0100] In the several embodiments provided in this application, it should be understood that the disclosed methods and apparatus can be implemented in other ways. For example, the apparatus implementations described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0101] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0102] Furthermore, the functional units in the various embodiments of this application 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 as a software functional unit.

[0103] If the integrated unit is implemented as 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 this application, in essence, 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. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute all or part of the steps of the methods of various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0104] If the technical solution of this application involves personal information, the product using this technical solution has clearly informed the user of the personal information processing rules and obtained the user's voluntary consent before processing the personal information. If the technical solution of this application involves sensitive personal information, the product using this technical solution has obtained the user's separate consent before processing the sensitive personal information, and also meets the requirement of "express consent". For example, at personal information collection devices such as cameras, clear and prominent signs are set up to inform users that they have entered the scope of personal information collection and that personal information will be collected. If an individual voluntarily enters the collection scope, it is deemed that they have agreed to the collection of their personal information; or on the personal information processing device, with clear signs / information informing users of the personal information processing rules, authorization is obtained from the individual through pop-up information or by asking the individual to upload their personal information; wherein, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the types of personal information processed.

Claims

1. A method for quality control of medical terminology, characterized in that, include: Obtain an initial terminology database; wherein the initial terminology database includes several pairs of medical terms, and the medical terminology pairs include standard terms and synonyms of the standard terms; Based on the terminology element template, the standard terms and synonyms in the medical terminology pair are split into terms, and the split terms are standardized to obtain the split and standardized results of the medical terminology pair. The split and standardized results include a first split and standardized combination of the standard terms and a second split and standardized combination of the synonyms. The split and standardized combination includes the standardized terms of the split terms corresponding to each terminology element after being split by the terminology element template. Based on the decomposition and standardization results of each pair of medical terms in the initial terminology database, a first quality check is performed on the initial terminology database to determine the medical terminology pairs that are expressed correctly, which are then used as the first terminology pairs. Based on the standard terminology mapping model, the splitting results, reference characters, and reference states of the synonyms in the second terminology pair are decoded to obtain the decoded characters corresponding to this decoding. The second terminology pair is a medical terminology pair other than the first terminology pair in the initial terminology library. The splitting results include the split words corresponding to each terminology element after the synonyms are split according to the terminology element template. The reference characters are the decoded characters obtained before this decoding, and the reference states are the hidden state of the standard terminology mapping model before this decoding. Based on the combination of the decoded characters corresponding to each decoding, the standardized mapping terms of the synonyms are obtained. Based on whether the mapping terms corresponding to the standard terms and the synonyms in the second term pair are consistent, determine the third term pair after quality inspection of the second term pair; Based on the first term pair and the third term pair, a target term library is constructed.

2. The method according to claim 1, characterized in that, The first quality check of the initial terminology database is performed based on the decomposition and standardization results of each pair of medical terms in the initial terminology database, and the medical terminology pairs in the initial terminology database that are expressed without error are determined as the first term pairs, including: In response to the fact that the first split-standardized combination of the standard term in the medical term pair and the second split-standardized combination of the synonym are completely consistent, check whether the synonym in the medical term pair has only one corresponding standard term in the initial term database. In response to the fact that the synonyms in the medical term pair have only one unique standard term in the initial terminology database, the medical term pair is determined to be the first term pair that is accurate.

3. The method according to claim 2, characterized in that, The method further includes at least one of the following: In response to the fact that the first split-standardized combination of the standard terms in the medical term pair and the second split-standardized combination of the synonyms are not completely consistent, the medical term pair is taken as the second term pair. In response to the fact that the synonyms in the medical term pair correspond to multiple standard terms in the initial terminology database, the medical term pair is used as the second term pair.

4. The method according to claim 1, characterized in that, The step of determining the third term pair after quality inspection of the second term pair based on whether the mapping terms corresponding to the standard terms and the synonyms in the second term pair are consistent includes at least one of the following: In response to the fact that the standard term and the mapped term corresponding to the synonym in the second term pair are consistent, the second term pair is directly determined as the third term pair; In response to the inconsistency between the standard term and the mapped term corresponding to the synonym in the second term pair, the third term pair is obtained based on the correction feedback of the second term pair.

5. The method according to claim 1, characterized in that, The terminology splitting steps include: Encoding is performed based on the term to be processed to obtain the encoded representation of each character in the term to be processed; wherein, the term to be processed is selected from the medical term pair; Based on the encoded representation of each character, a prediction is made to obtain the character identifier of each character; wherein, the character identifier of the character is used to indicate that the character belongs to any of the following situations: the character alone constitutes the split word, the character constitutes the starting character of the split word, or the character constitutes the ending character of the split word; Based on the terminology element template and the character identifiers of each character, the decomposed words corresponding to each terminology element after the term to be processed is decomposed according to the terminology element template are obtained.

6. The method according to claim 1 or 5, characterized in that, The term segmentation is performed by a term segmentation model, which includes an encoding network and a prediction network. The encoding network is used to encode the terms to be processed selected from the medical term pairs to obtain the encoded representation of each character in the terms to be processed. The prediction network is used to predict the encoded representation of each character to obtain the character identifier of each character.

7. The method according to claim 1, characterized in that, The medical terminology refers to at least symptom terms, and the terminology element template includes at least a symptom terminology template applicable to the symptom terms. Each terminology element constituting the symptom terminology template includes location, primary location, secondary location, and symptom.

8. A medical terminology quality monitoring device, characterized in that, include: A terminology database acquisition module is used to acquire an initial terminology database; wherein, the initial terminology database includes several pairs of medical terminology, and the medical terminology pairs include standard terms and synonyms of the standard terms; The splitting and standardization module is used to perform term splitting and word standardization on the standard terms and synonyms in the medical term pair based on the term element template, so as to obtain the splitting and standardization result of the medical term pair; wherein, the splitting and standardization result includes a first splitting and standardization combination of the standard terms and a second splitting and standardization combination of the synonyms, and the splitting and standardization combination includes the standard words after the terms are split by referring to the term element template and the words corresponding to each term element after standardization. The determination submodule is used to perform a first quality check on the initial terminology database based on the splitting and standardization results of each pair of medical terminology pairs in the initial terminology database, and to determine the medical terminology pairs in the initial terminology database that are expressed correctly as the first terminology pairs; The decoding unit is used to decode the splitting results, reference characters, and reference states of the synonyms in the second term pair based on the standard term mapping model, so as to obtain the decoded characters corresponding to this decoding. The second term pair is a pair of medical terms other than the first term pair in the initial term library. The splitting results include the split words corresponding to each term element after the synonyms are split according to the term element template. The reference characters are the decoded characters obtained before this decoding, and the reference states are the hidden state of the standard term mapping model before this decoding. The combination unit is used to combine the decoded characters corresponding to each decoding to obtain the standardized mapping term of the synonym term; The determining unit is configured to determine a third term pair after quality inspection of the second term pair based on whether the mapping terms corresponding to the standard terms and the synonyms in the second term pair are consistent. A construction submodule is used to construct a target terminology library based on the first term pair and the third term pair.

9. An electronic device, characterized in that, The method includes a memory and a processor coupled to each other, the memory storing program instructions, and the processor executing the program instructions to implement the medical terminology quality monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The device stores program instructions that can be executed by a processor, the program instructions being used to implement the medical terminology quality monitoring method according to any one of claims 1 to 7.

Images