Intelligent query method for nursing knowledge base

Abstract

The invention relates to the technical field of medical care, in particular to an intelligent query method for a nursing knowledge base, which comprises the following steps of: performing multi-form standardized processing on a nursing problem input by a user, deeply analyzing a query request by utilizing a semantic understanding model, and obtaining an intelligent query result based on an inference algorithm; the method comprises the following steps: automatically selecting related literatures, clinical data and nursing experience data from a nursing knowledge base, generating a preliminary nursing scheme candidate set in combination with a nursing professional knowledge graph, and performing personalized screening and optimization on the scheme by using an intelligent recommendation model in combination with the current illness state, treatment progress and historical nursing data of a patient; the intelligent recommendation model is self-optimized through user feedback, and the nursing scheme recommendation precision is continuously improved. The nursing knowledge base automatically fuses newest medical research results and clinical nursing practice data. The problems that in the prior art, nursing scheme recommendation accuracy is low, updating is not timely, and the individuation degree is insufficient are effectively solved.

Description

Technical Field

[0001] The present invention relates to the field of medical care technology, and in particular to an intelligent query method for a nursing knowledge base. Background Art

[0002] With the continuous development of the medical industry, nursing work plays an increasingly important role in ensuring patient recovery and improving treatment outcomes. Especially in clinical nursing, nurses are faced with a large amount of information and cases. How to quickly and accurately provide patients with personalized and precise nursing plans has become a key challenge in the nursing field. To this end, the nursing knowledge base came into being. It brings together medical research, clinical cases, and nursing experience to help nurses better provide patients with scientific nursing plans.

[0003] At present, the existing intelligent query methods of nursing knowledge bases usually rely on traditional rule-based reasoning methods or simple keyword matching algorithms. Although these methods can provide some reference, they have obvious limitations when dealing with complex and personalized nursing needs. First, in the processing of query requests, there is often a lack of effective integration of multiple query forms (such as text, voice, images, etc.), which cannot fully understand and analyze the diverse needs of users. Secondly, the accuracy of the semantic parsing module is insufficient, and it fails to deeply explore the core intent, patient-related characteristics, disease background and other key information in the query request, resulting in low accuracy and personalization of nursing plan recommendations. Summary of the invention

[0004] The invention provides a nursing knowledge base intelligent query method.

[0005] A nursing knowledge base intelligent query method comprises the following steps:

[0006] S1: The collection module collects the nursing query requests input by the user, and the query request forms include text, voice or image, and the data fusion technology is used to uniformly convert the query requests in different forms into standardized query requests;

[0007] S2: Use the semantic understanding model to semantically parse the query request, extract the intent of the nursing query, patient-related characteristics, disease background, and potential nursing plan keywords, and refine the query based on contextual information, including the patient's historical medical history, current treatment plan, and environmental factors;

[0008] S3: According to the semantic analysis results, relevant literature, clinical data and nursing experience data are automatically selected from the nursing knowledge base, and the relevant literature, clinical data and nursing experience data are analyzed in multiple dimensions through the inference algorithm in combination with the nursing professional knowledge map to generate a preliminary nursing plan candidate set;

[0009] S4: Based on the situational reasoning module, combined with the patient's current condition, treatment progress, historical nursing data and environmental factors, the intelligent recommendation model is used to contextually adapt the preliminary nursing plan candidate set, screen out the personalized nursing plan that best meets the patient's needs, and generate personalized nursing recommendations;

[0010] S5: Through the dynamic learning module, real-time feedback is provided on the effect of users’ use of the nursing plan. Based on the user feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the nursing plan and the intelligence level of the nursing plan recommendation, so as to improve the recommendation quality of subsequent queries.

[0011] S6: Continuously update and maintain the nursing knowledge base, and automatically integrate the latest medical research results and clinical nursing practice data.

[0012] Optionally, the S1 includes:

[0013] S11: receiving a nursing question query request input by a user through a collection module, wherein the query request includes text, voice or image form;

[0014] S12: When the query request is in text form, perform text preprocessing, including removing irrelevant information, word segmentation, part-of-speech tagging, entity recognition, and keyword extraction to generate text data containing the query intent;

[0015] S13: When the query request is in voice form, a voice recognition unit is used to convert the voice signal into text form, wherein the voice recognition unit includes an acoustic model and a language model, wherein the acoustic model is used to convert the voice signal into phonemes or syllables, and the language model is used to perform semantic inference on the voice based on the context and statistical data of the voice, so as to obtain text data with nursing query intent;

[0016] S14: When the query request is in the form of an image, an image recognition unit is used to process the image, and a convolutional neural network model is used to analyze the image to identify text information, icons or tables in the image and convert them into text data;

[0017] S15: For all forms of query requests (text, voice, image), the text data obtained after processing by S12, S13 and S14 is subjected to data fusion processing, and different types of query requests are uniformly converted into standardized query requests using data fusion technology.

[0018] Optionally, S2 includes:

[0019] S21: performing semantic analysis on the standardized query request using a semantic understanding model, wherein the semantic understanding model uses a natural language processing model based on deep learning to analyze the query text;

[0020] S22: Through named entity recognition technology, relevant patient characteristics (such as age, gender, medical history, etc.), disease background information and keywords of care plans are extracted from the query request, and refined according to the query context information to ensure accurate capture of user needs;

[0021] S23: Through sentiment analysis technology, analyze the emotional color and emotional tendency in the query, identify the user's emotional expression in the query (such as anxiety, confusion, etc.), and combine the emotional information with the output results of the semantic understanding model to further adjust the semantic understanding model's interpretation of the query intent.

[0022] Optionally, S2 further includes:

[0023] S24: Refine the query based on context information, where the context information includes the patient's medical history, current treatment plan, and environmental factors, and use a deep learning model to understand the query context to capture specific situations in the query;

[0024] S25: Combine the patient's characteristics and environmental information to make detailed recommendations on the query results to ensure that the generated care plan is personalized to the patient's actual needs and situation.

[0025] Optionally, the S3 includes:

[0026] S31: automatically selecting literature, clinical data and nursing experience data related to the query request from the nursing knowledge base according to the semantic analysis result;

[0027] S32: Integrate the screened relevant literature, clinical data and nursing experience data into a preliminary candidate set.

[0028] Optionally, S3 further includes:

[0029] S33, using the Bayesian inference model to conduct multi-dimensional correlation analysis on the selected relevant literature, clinical data and nursing experience data to generate a preliminary nursing plan candidate set;

[0030] S34: Based on the generated preliminary nursing plan candidate set, the candidate plans are sorted, and the optimal nursing plan is screened out according to the confidence value output by the Bayesian reasoning model.

[0031] Optionally, the S4 includes:

[0032] S41: Based on the situational reasoning module, the preliminary nursing plan candidate set is situationally adapted in combination with the patient's current condition, treatment progress, historical nursing data and environmental factors, including:

[0033] Condition analysis: Based on the patient's current condition (such as the type of disease, severity, symptoms, etc.), the situational reasoning module is used to evaluate the impact of the condition on the nursing plan;

[0034] Treatment progress assessment: Combined with the patient's current treatment plan and treatment progress, evaluate whether the nursing plan is suitable for the patient's treatment status;

[0035] Historical nursing data analysis: By analyzing the patient's historical nursing data (such as past nursing records, effect feedback, complications, etc.), it is used as the key input in the situational reasoning module to help the situational reasoning module adjust the current recommendation plan based on past nursing effects;

[0036] Environmental factor identification: Consider the patient's environmental factors and evaluate the impact of environmental factors on the care plan through the situational reasoning module;

[0037] S42: Use graph neural networks to build an intelligent recommendation model to make personalized recommendations on the candidate set of preliminary care plans that have been adapted through contextual reasoning.

[0038] Optionally, the S5 includes:

[0039] S51: Collecting user feedback data on the nursing plan in real time through a dynamic learning module, wherein the feedback data includes the user's evaluation of the effect of the nursing plan after implementation, the patient's nursing effect feedback, and the user's satisfaction score;

[0040] S52: Based on the collected feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the recommended care plan, including:

[0041] Model update: Update the parameters and weights in the intelligent recommendation model based on user feedback data;

[0042] Error correction: Analyze the deviation between the recommendation results and user feedback, and correct the prediction error in the recommendation algorithm;

[0043] Recommendation strategy optimization: Optimize recommendation strategies based on user feedback to enhance the ability to provide personalized recommendations for nursing plans;

[0044] S53: According to the feedback results of the dynamic learning module, the accuracy of the recommended nursing plan is adjusted in real time.

[0045] Optionally, the S6 includes:

[0046] S61: Continuously update and maintain the nursing knowledge base, automatically integrating the latest medical research results, clinical nursing practice data and relevant patient feedback data;

[0047] S62: Based on the data fusion results, the content in the nursing knowledge base is dynamically adjusted using knowledge graph update technology;

[0048] S63: Regularly perform version control and quality assessment on the nursing knowledge base.

[0049] Beneficial effects of the present invention:

[0050] The present invention integrates query requests in various forms (such as text, voice, and image) through the acquisition module, and converts them into standardized query requests using data fusion technology. The query request is deeply analyzed through the semantic understanding model to extract the core intent, patient characteristics, disease background, and potential nursing plan keywords to ensure that the query results are highly accurate and targeted. Combined with sentiment analysis technology, it can effectively identify user emotional expressions and further optimize the understanding of query intent, so that the system can provide nursing plans that better meet user needs and improve user experience and query results.

[0051] The present invention, by combining nursing professional knowledge maps and reasoning algorithms, can perform multi-dimensional correlation analysis on relevant literature, clinical data and nursing experience data to generate a preliminary nursing plan candidate set. On this basis, the situational reasoning module combines the patient's current condition, treatment progress, historical nursing data and environmental factors, and uses an intelligent recommendation model to situationally adapt the preliminary nursing plan to screen out the personalized nursing plan that best meets the patient's needs. This process can ensure the personalization and precision of the nursing plan, and improve the patient's treatment effect and nursing experience.

[0052] The present invention has the ability to continuously update and self-optimize. Through the dynamic learning module, the user's nursing plan usage feedback is collected in real time, and the intelligent recommendation model is self-optimized based on the user feedback data, continuously improving the accuracy and intelligence level of subsequent query recommendations. In addition, the nursing knowledge base automatically integrates the latest medical research results, clinical nursing practice data and patient feedback to ensure that the nursing plan always has the latest clinical value and practicality, can adapt to the rapidly changing medical and nursing practice needs, and maintain the long-term effectiveness and foresight of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0054] Figure 1 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0055] Figure 2 Schematic diagram of S2 process of an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. At the same time, it is explained here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments, and those skilled in the art may also adopt other alternatives to implement some known technologies; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0057] like Figure 1-Figure 2 As shown, a nursing knowledge base intelligent query method includes the following steps:

[0058] S1: The collection module collects the nursing query requests input by the user, and the query request forms include text, voice or image, and the data fusion technology is used to uniformly convert the query requests in different forms into standardized query requests;

[0059] S2: Use the semantic understanding model to semantically parse the query request, extract the intent of the nursing query, patient-related characteristics, disease background, and potential nursing plan keywords, and refine the query based on contextual information, including the patient's historical medical history, current treatment plan, and environmental factors;

[0060] S3: According to the semantic analysis results, relevant literature, clinical data and nursing experience data are automatically selected from the nursing knowledge base. Combined with the nursing professional knowledge map, the relevant literature, clinical data and nursing experience data are analyzed in multiple dimensions through the inference algorithm to generate a preliminary nursing plan candidate set. The inference algorithm is optimized based on professional rules and historical case data in the nursing field.

[0061] S4: Based on the situational reasoning module, combined with the patient's current condition, treatment progress, historical nursing data and environmental factors, the intelligent recommendation model is used to contextually adapt the preliminary nursing plan candidate set, screen out the personalized nursing plan that best meets the patient's needs, and generate personalized nursing recommendations;

[0062] S5: Through the dynamic learning module, real-time feedback is provided on the effect of users’ use of the nursing plan. Based on the user feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the nursing plan and the intelligence level of the nursing plan recommendation, so as to improve the recommendation quality of subsequent queries.

[0063] S6: Continuously update and maintain the nursing knowledge base, automatically integrate the latest medical research results and clinical nursing practice data, and ensure that the nursing plan always has the latest clinical value and practicality.

[0064] S1 includes:

[0065] S11: receiving a nursing question query request input by a user through a collection module, where the query request includes text, voice or image form;

[0066] S12: When the query request is in text form, perform text preprocessing, including removing irrelevant information, word segmentation, part-of-speech tagging, entity recognition, and keyword extraction to generate text data containing the query intent;

[0067] S13: When the query request is in voice form, a voice recognition unit is used to convert the voice signal into text form. The voice recognition unit includes an acoustic model and a language model. The acoustic model is used to convert the voice signal into phonemes or syllables. The language model is used to perform semantic inference on the voice based on the context and statistical data of the voice, so as to obtain text data with nursing query intent.

[0068] S14: When the query request is in the form of an image, an image recognition unit is used to process the image, a convolutional neural network model is used to analyze the image, text information, icons or tables in the image are identified, and converted into text data, including the following steps:

[0069] S141: preprocessing the input image, including denoising, grayscale, binarization and other operations, to improve image quality and reduce noise interference, and ensure the accuracy of subsequent processing;

[0070] S142: using a convolutional neural network model to extract features from the image. The structure of the convolutional neural network model includes multiple convolutional layers, pooling layers, and fully connected layers, which are used to extract local features and global features of the image to identify key information in the image.

[0071] S143: Classify and annotate images through a convolutional neural network model to identify text information, icons, tables and other elements in the image. The text information is further converted into standardized text through optical character recognition technology;

[0072] S144: using the table and icon information in the image, segmenting each cell area of ​​the table through image segmentation technology, and recognizing the text in each cell and converting it into a standard text format;

[0073] S145: Unifying and arranging the recognized text information, icons and table data, converting them into standardized text data, and transmitting them to subsequent steps for processing;

[0074] S15: For all forms of query requests (text, voice, image), the text data obtained after processing by S12, S13 and S14 is subjected to data fusion processing. The data fusion technology is used to uniformly convert different types of query requests into standardized query requests to ensure the uniformity and consistency of data and eliminate the deviation of different input forms.

[0075] Through steps S11 to S15, query requests in different forms (text, voice, image) can be effectively processed, and these different forms of input can be converted into unified standardized query requests through data fusion technology, ensuring that subsequent steps can efficiently process and analyze these query requests.

[0076] S2 includes:

[0077] S21: The semantic understanding model is used to perform semantic analysis on the standardized query request. The semantic understanding model uses a natural language processing model based on deep learning to analyze the query text, where:

[0078] Lexical analysis: The pre-trained long short-term memory network model is used to segment the input query text and tag each word with a part-of-speech tag to ensure that each word in the text can be accurately classified (such as noun, verb, adjective, etc.);

[0079] Syntactic analysis: Analyze the syntactic structure of the query text through the long short-term memory network model, identify the dependency relationship between words (such as subject-predicate relationship, modification relationship, etc.), and build a grammar tree or dependency tree to help understand the sentence structure in the query;

[0080] Semantic analysis: Use the long short-term memory network model to perform deep semantic analysis on the query text, combine contextual information (such as the patient's medical history, current treatment plan, and environmental factors), extract the semantic roles and key elements in the query, and identify potential nursing issues, patient characteristics, and disease background;

[0081] S22: Through named entity recognition technology, relevant patient characteristics (such as age, gender, medical history, etc.), disease background information and keywords of care plans are extracted from the query request, and refined according to the query context information to ensure accurate capture of user needs;

[0082] S23: Use sentiment analysis technology to analyze the emotional color and emotional tendency in the query, identify the user's emotional expression in the query (such as anxiety, confusion, etc.), and combine the emotional information with the output results of the semantic understanding model to further adjust the semantic understanding model's interpretation of the query intent. According to the results of sentiment analysis, optimize the understanding of the query requirements, especially when emotions are expressed in the query (such as anxiety or nervousness), to help improve the model's accurate interpretation of the query;

[0083] Through steps S21 to S23, a deep learning model is used to perform comprehensive semantic analysis of the user's query request, which can not only extract the core intent in the query, but also extract patient characteristics, disease background, and nursing keywords, and combine sentiment analysis to further optimize the understanding of user needs.

[0084] S2 also includes:

[0085] S24: Refine the query based on contextual information, which includes the patient's medical history, current treatment plan, and environmental factors. Use a deep learning model to understand the context of the query and capture the specific situation in the query. The specific steps include:

[0086] Historical medical record analysis: By extracting the patient's historical medical record data, the patient's past medical records, disease development and other information are used as context, combined with the current query request to provide the model with more background information to help more accurately understand the query content;

[0087] Current treatment plan analysis: Based on the patient's current treatment plan (such as drug treatment, surgical plan, etc.), combined with relevant information in the query request, the model can determine whether the current query request is related to the ongoing treatment plan, and adjust the nursing plan based on this information;

[0088] Environmental factor analysis: Analyze the patient’s environment (such as hospitalization environment, home care, etc.), and combine it with the environmental description in the query request to ensure that the model can take these factors into account when understanding the query and further refine the query context;

[0089] S25: Combine the patient's characteristics and environmental information to make detailed recommendations on the query results to ensure that the generated nursing plan is personalized to the patient's actual needs and situations;

[0090] Through steps S24 to S25, context information is used to refine the query, fully considering the patient's medical history, treatment plan and environmental factors, and further improving the personalization and accuracy of nursing plan recommendations.

[0091] S3 includes:

[0092] S31: According to the semantic analysis results, the literature, clinical data and nursing experience data related to the query request are automatically selected from the nursing knowledge base. The process includes the following steps:

[0093] Literature selection: Through keyword matching, semantic similarity calculation and domain expert system rules, medical literature related to the user's query needs is screened from the nursing knowledge base. Natural language processing technology (such as the BERT model) is used to semantically vectorize the literature, and similarity is calculated with the query text to select the literature content that best matches the query intent.

[0094] Clinical data selection: Based on the patient's characteristic information (such as age, gender, medical history, etc.) and disease background, clinical data that matches the patient's condition is retrieved from the database. This step uses rule-based query optimization technology combined with multi-level labels and attribute information of the nursing database to ensure that the selected clinical data is highly matched with the patient's needs;

[0095] Nursing experience data selection: Based on the specific situation of the patient, highly practical nursing experience data is selected from the nursing knowledge base. This process selects the most targeted nursing experience data through labeling analysis of nursing cases and intelligent matching of nursing log data;

[0096] S32: Integrate the relevant literature, clinical data and nursing experience data obtained through screening into a preliminary candidate set to provide input data for subsequent reasoning analysis;

[0097] Through steps S31 and S32, literature, clinical data, and nursing experience data related to the query are automatically selected based on the semantic parsing results, effectively providing accurate data support for subsequent reasoning analysis and nursing plan generation.

[0098] S3 also includes:

[0099] S33, using the Bayesian inference model to conduct multi-dimensional correlation analysis on the selected relevant literature, clinical data and nursing experience data to generate a preliminary nursing plan candidate set. The specific steps include:

[0100] Data preprocessing: Standardize the selected literature, clinical data, and nursing experience data and convert them into probability distribution form to ensure that the format of input data is suitable for processing by the Bayesian inference model;

[0101] Bayesian reasoning analysis: The Bayesian reasoning model is used to perform joint probability calculations on multiple input data sources (including literature, clinical data, and nursing experience data) to derive the relationship between different data. Specifically, the Bayesian reasoning model is based on prior probability and likelihood function, combined with historical case data and professional rules in the nursing field to update parameters and derive the probability distribution of each nursing plan;

[0102] Generate candidate set: Generate a preliminary nursing plan candidate set based on the output of the Bayesian inference model. The recommendation priority of each candidate nursing plan is based on its posterior probability value. Nursing plans with higher posterior probabilities are selected first and used as candidate plans for subsequent recommendations.

[0103] S34: Based on the generated preliminary nursing plan candidate set, the candidate plans are sorted, and the optimal nursing plan is selected according to the confidence value output by the Bayesian reasoning model. This step ensures that the recommended nursing plan best meets the patient's specific situation and treatment needs by comprehensively considering the rationality of each plan, the patient's personalized needs, and the best practices in the nursing field;

[0104] The Bayesian reasoning model can combine multi-dimensional literature, clinical data and nursing experience data to conduct effective association analysis and generate a preliminary nursing plan candidate set that best matches the patient's needs. At the same time, the Bayesian reasoning model can optimize the recommendation process and improve the intelligence and accuracy of the nursing plan through parameter updating and probability calculation.

[0105] S4 includes:

[0106] S41: Based on the situational reasoning module, the preliminary nursing plan candidate set is situationally adapted in combination with the patient's current condition, treatment progress, historical nursing data and environmental factors, including:

[0107] Condition analysis: Based on the patient's current condition (such as the type of disease, severity, symptoms, etc.), the situational reasoning module is used to evaluate the impact of the condition on the nursing plan. For example, the selection criteria and recommended strategies for nursing plans may be different for patients with acute diseases and those with chronic diseases.

[0108] Treatment progress assessment: Combined with the patient's current treatment plan and treatment progress, evaluate whether the care plan is suitable for the patient's treatment status. For example, if the patient is receiving antibiotics, relevant care plans may need to be selected to avoid drug interactions.

[0109] Historical nursing data analysis: By analyzing the patient's historical nursing data (such as past nursing records, effect feedback, complications, etc.), it is used as the key input in the situational reasoning module to help the situational reasoning module adjust the current recommendation plan based on past nursing effects;

[0110] Identification of environmental factors: Considering the patient's environmental factors (such as hospitalization environment, home care situation, professional level of nursing staff, etc.), the situational reasoning module is used to evaluate the impact of these environmental factors on the nursing plan to further adjust the adaptability of the recommended plan;

[0111] S42: Use graph neural networks to build an intelligent recommendation model to make personalized recommendations for the candidate set of preliminary care plans adapted through contextual reasoning, including:

[0112] Graph construction: Model the patient's medical history, current treatment plan, and environmental factors as graph-structured data, where nodes represent the patient's various characteristics and care plans, and edges represent the relationship between characteristics;

[0113] Graph convolutional network application: Use the graph convolutional network in the graph neural network to perform convolution operations on the patient's feature map and obtain information from neighboring nodes to learn the complex relationship between the patient's personalized needs and care plans;

[0114] Feature aggregation and recommendation: The intelligent recommendation model built through the graph neural network aggregates the patient's individual characteristics, historical nursing data, and treatment progress, and combines the contextual information of the candidate nursing plans to generate the final personalized nursing recommendations. When outputting the recommendation results, the nursing plans are sorted according to the patient's characteristics and the most suitable plan is recommended;

[0115] Through steps S41 to S42, the situational reasoning module is combined with the intelligent recommendation model, so that the care plan can be dynamically adapted according to the patient's current status, historical data and environmental changes, ensuring that the recommended care plan is highly personalized and accurate, thereby significantly improving the quality of care and the patient's care experience.

[0116] S5 includes:

[0117] S51: Through the dynamic learning module, real-time user feedback data on the nursing plan is collected. The feedback data includes the user's evaluation of the effect of the nursing plan after implementation, the patient's nursing effect feedback, and the user's satisfaction score. The process includes the following steps:

[0118] Effect feedback collection: Through the user interface or nursing log system, collect the effect data of patients after using the nursing plan (such as changes in condition, symptom relief, etc.), and record the user's regular evaluation of the implementation effect of the nursing plan;

[0119] User satisfaction rating: Collect user satisfaction ratings on the nursing plan through questionnaires, scoring systems, etc. The feedback includes evaluations of the practicality, comfort, and ease of operation of the nursing plan;

[0120] Nursing effect record: The system records the effect of the nursing plan after implementation based on the patient's feedback data, including treatment progress, changes in the condition, etc., and provides comprehensive feedback information;

[0121] S52: Based on the collected feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the recommended care plan, including:

[0122] Model update: Update the parameters and weights in the intelligent recommendation model based on user feedback data, and adjust the model's response strategy to different types of queries and patient characteristics;

[0123] Error correction: Analyze the deviation between the recommendation results and user feedback, correct the prediction error in the recommendation algorithm, and use incremental learning or reinforcement learning methods to enable the model to make subsequent recommendations more accurately;

[0124] Recommendation strategy optimization: Based on user feedback, optimize the recommendation strategy and enhance the personalized recommendation capability of nursing plans. The optimization process includes adjusting the nursing plan recommendation strategy for similar user groups, thereby improving the quality and intelligence of recommendations.

[0125] S53: According to the feedback results of the dynamic learning module, the accuracy of the nursing plan recommendation is adjusted in real time to improve the quality of the nursing plan recommendation in subsequent queries. This process is carried out in the following ways:

[0126] Improved accuracy: The dynamic learning module adjusts the recommendation strategy in real time based on user feedback. Through continuous iteration and learning, the recommended nursing plans are more accurate and personalized.

[0127] Learning feedback loop: Establish a feedback learning loop to gradually optimize the recommendation effect based on the user's acceptance of the recommended care plan, and continuously enhance the adaptive ability of the intelligent recommendation system.

[0128] S6 includes:

[0129] S61: Continuously update and maintain the nursing knowledge base, automatically integrating the latest medical research results, clinical nursing practice data and related patient feedback data. The specific steps are as follows:

[0130] Integration of medical research results: Through automated literature retrieval and data mining technology, the latest medical research results are obtained from medical journals, clinical trial databases and other sources. Natural language processing (NLP) technology is used to extract key knowledge points from the research results and store them in a structured manner in the nursing knowledge base to ensure that the new medical research results are seamlessly connected with the existing knowledge system.

[0131] Clinical nursing practice data integration: real-time access to the practice data of hospitals and clinical nursing departments, such as clinical nursing logs, nursing case libraries, etc., automatically extracting and organizing key data in the implementation of nursing plans, and using data cleaning and standardization technologies to ensure that nursing data from different sources can be unified in format and integrated into the knowledge base;

[0132] Patient feedback data integration: Extract feedback data from patients' usage feedback, nursing effect evaluation and patient satisfaction scores, and automatically integrate them into the nursing knowledge base. By analyzing patient feedback and the effects of nursing plans, potential improvement points can be found, and the practicality and adaptability of nursing plans can be continuously improved;

[0133] S62: Based on the data fusion results, the content in the nursing knowledge base is dynamically adjusted using the knowledge graph update technology. The specific steps are as follows:

[0134] Knowledge graph update: Automatically update the knowledge graph in the nursing knowledge base based on newly integrated medical research results, clinical nursing data, and patient feedback data, and expand the nodes and edges in the graph to more accurately reflect the latest information and nursing practice trends in the field of clinical nursing;

[0135] Knowledge reasoning optimization: Use the Bayesian reasoning model to reason about the updated knowledge graph, optimize the accuracy and adaptability of nursing plan recommendations, and continuously improve the intelligence level of the nursing knowledge base for personalized nursing plan recommendations through reasoning algorithms and deep learning technology;

[0136] S63: Regularly conduct version control and quality assessment of the nursing knowledge base. Specific steps include:

[0137] Version management: Manage the data in the nursing knowledge base through the version control system to ensure that the updated version of each data can be traced back to avoid confusion of knowledge in the knowledge base due to update errors;

[0138] Quality assessment: By regularly evaluating the effectiveness and accuracy of each nursing plan in the nursing knowledge base, the nursing plan content in the knowledge base is continuously adjusted based on expert review and user feedback to ensure that the nursing plan always meets the latest clinical needs and standards.

[0139] The present invention covers any substitution, modification, equivalent method and scheme made on the essence and scope of the present invention. In order to make the public have a thorough understanding of the present invention, specific details are described in detail in the following preferred embodiments of the present invention, but those skilled in the art can fully understand the present invention without the description of these details. In addition, in order to avoid unnecessary confusion about the essence of the present invention, well-known methods, processes, procedures, components and circuits are not described in detail.

[0140] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A nursing knowledge base intelligent query method, characterized in that: The following steps are involved: S1: The collection module collects the nursing query requests input by the user, and the query request forms include text, voice or image, and the data fusion technology is used to uniformly convert the query requests in different forms into standardized query requests; S2: Use the semantic understanding model to semantically parse the query request, extract the intent of the nursing query, patient-related characteristics, disease background, and potential nursing plan keywords, and refine the query based on contextual information, including the patient's historical medical history, current treatment plan, and environmental factors; S3: According to the semantic analysis results, relevant literature, clinical data and nursing experience data are automatically selected from the nursing knowledge base, and the relevant literature, clinical data and nursing experience data are analyzed in multiple dimensions through the inference algorithm in combination with the nursing professional knowledge map to generate a preliminary nursing plan candidate set; S4: Based on the situational reasoning module, combined with the patient's current condition, treatment progress, historical nursing data and environmental factors, the intelligent recommendation model is used to contextually adapt the preliminary nursing plan candidate set, screen out the personalized nursing plan that best meets the patient's needs, and generate personalized nursing recommendations; S5: Through the dynamic learning module, real-time feedback is provided on the effect of users’ use of the nursing plan. Based on the user feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the nursing plan and the intelligence level of the nursing plan recommendation, so as to improve the recommendation quality of subsequent queries. S6: Continuously update and maintain the nursing knowledge base, and automatically integrate the latest medical research results and clinical nursing practice data.

2. A nursing knowledge base intelligent query method according to claim 1, characterized in that: The S1 includes: S11: receiving a nursing question query request input by a user through a collection module, wherein the query request includes text, voice or image form; S12: When the query request is in text form, perform text preprocessing, including removing irrelevant information, word segmentation, part-of-speech tagging, entity recognition, and keyword extraction to generate text data containing the query intent; S13: When the query request is in voice form, a voice recognition unit is used to convert the voice signal into text form, wherein the voice recognition unit includes an acoustic model and a language model, wherein the acoustic model is used to convert the voice signal into phonemes or syllables, and the language model is used to perform semantic inference on the voice based on the context and statistical data of the voice, so as to obtain text data with nursing query intent; S14: When the query request is in the form of an image, an image recognition unit is used to process the image, and a convolutional neural network model is used to analyze the image to identify text information, icons or tables in the image and convert them into text data; S15: For all forms of query requests, the text data obtained after processing by S12, S13 and S14 is subjected to data fusion processing, and different types of query requests are uniformly converted into standardized query requests using data fusion technology.

3. A nursing knowledge base intelligent query method according to claim 2, characterized in that: The S2 includes: S21: performing semantic analysis on the standardized query request using a semantic understanding model, wherein the semantic understanding model uses a natural language processing model based on deep learning to analyze the query text; S22: Through named entity recognition technology, relevant patient characteristics, disease background information and nursing plan keywords are extracted from the query request, and refined according to the query context information; S23: Through sentiment analysis technology, analyze the emotional color and emotional tendency in the query, identify the user's emotional expression in the query, and combine the emotional information with the output results of the semantic understanding model to further adjust the semantic understanding model's interpretation of the query intent.

4. A nursing knowledge base intelligent query method according to claim 3, characterized in that: The S2 further includes: S24: Refine the query based on context information, where the context information includes the patient's medical history, current treatment plan, and environmental factors, and use a deep learning model to understand the query context to capture specific situations in the query; S25: Based on the patient's characteristics and environmental information, make refined recommendations on the query results.

5. A nursing knowledge base intelligent query method according to claim 4, characterized in that: The S3 includes: S31: automatically selecting literature, clinical data and nursing experience data related to the query request from the nursing knowledge base according to the semantic analysis result; S32: Integrate the screened relevant literature, clinical data and nursing experience data into a preliminary candidate set.

6. A nursing knowledge base intelligent query method according to claim 5, characterized in that: The S3 further includes: S33, using the Bayesian inference model to conduct multi-dimensional correlation analysis on the selected relevant literature, clinical data and nursing experience data to generate a preliminary nursing plan candidate set; S34: Based on the generated preliminary nursing plan candidate set, the candidate plans are sorted, and the optimal nursing plan is screened out according to the confidence value output by the Bayesian reasoning model.

7. A nursing knowledge base intelligent query method according to claim 6, characterized in that: The S4 includes: S41: Based on the situational reasoning module, the preliminary nursing plan candidate set is situationally adapted in combination with the patient's current condition, treatment progress, historical nursing data and environmental factors, including: Condition analysis: Based on the patient's current condition, the situational reasoning module is used to assess the impact of the condition on the nursing plan; Treatment progress assessment: Combined with the patient's current treatment plan and treatment progress, evaluate whether the nursing plan is suitable for the patient's treatment status; Historical nursing data analysis: By analyzing the patient's historical nursing data and using it as input in the contextual reasoning module, the contextual reasoning module can adjust the current recommendation plan based on past nursing results; Environmental factor identification: Consider the patient's environmental factors and evaluate the impact of environmental factors on the care plan through the situational reasoning module; S42: Use graph neural networks to build an intelligent recommendation model to make personalized recommendations on the candidate set of preliminary care plans that have been adapted through contextual reasoning.

8. A nursing knowledge base intelligent query method according to claim 7, characterized in that: The S5 includes: S51: Collecting user feedback data on the nursing plan in real time through a dynamic learning module, wherein the feedback data includes the user's evaluation of the effect of the nursing plan after implementation, the patient's nursing effect feedback, and the user's satisfaction score; S52: Based on the collected feedback data, the intelligent recommendation model is self-optimized to improve the accuracy of the recommended care plan, including: Model update: Update the parameters and weights in the intelligent recommendation model based on user feedback data; Error correction: Analyze the deviation between the recommendation results and user feedback, and correct the prediction error in the recommendation algorithm; Recommendation strategy optimization: Optimize recommendation strategies based on user feedback to enhance the ability to provide personalized recommendations for nursing plans; S53: According to the feedback results of the dynamic learning module, the accuracy of the recommended nursing plan is adjusted in real time.

9. A nursing knowledge base intelligent query method according to claim 8, characterized in that: The S6 includes: S61: Continuously update and maintain the nursing knowledge base, automatically integrating the latest medical research results, clinical nursing practice data and relevant patient feedback data; S62: Based on the data fusion results, the content in the nursing knowledge base is dynamically adjusted using knowledge graph update technology; S63: Regularly perform version control and quality assessment on the nursing knowledge base.

Images