A hypertension management system based on cloud platform

Abstract

The present invention relates to the field of hypertension management, and in particular to a hypertension management system based on a cloud platform. The present invention provides a preprocessing module, a data pool module, a label extraction module, and a labeling analysis module. The preprocessing module determines the semantic specificity characterization parameters corresponding to each disease type by extracting the semantic difference between the symptom description information and the symptom description information in the standard symptom description information library, and divides the semantic specificity category of the disease type. The data pool module determines whether it is necessary to build a symptom description information data pool for the disease type based on the specific category of the disease type. The label extraction module sets the disease type label for the user end. The labeling analysis module determines the semantic specificity category based on the disease type label corresponding to the user end and processes the symptom description information, including sending the corrected symptom description information and / or symptom description information to the receiving end. The present invention improves diagnostic efficiency and accuracy by dividing the semantic specificity category of the disease type and correcting the symptom description information.

Description

Technical Field

[0001] The present invention relates to the field of hypertension management, and in particular to a hypertension management system based on a cloud platform. Background Art

[0002] Hypertension is a common chronic disease that can be triggered by a variety of conditions. It not only increases the risk of cardiovascular disease but can also lead to a variety of postoperative complications. For example, cerebral hyperperfusion syndrome (CHS) is a specific complication of moyamoya disease (MD) surgery, associated with high mortality and disability rates. Hypertension is a major contributing factor to CHS, making continuous postoperative monitoring of MD patients crucial. With the rapid development of the "Internet + Healthcare" model, many innovative applications have emerged in modern healthcare, providing a convenient platform for communication between doctors and patients. These applications allow MD patients to easily describe their symptoms, postoperative recovery, and recent medications on their smartphones or computers, eliminating the need for hospital stays. Doctors can then view this information in real time and provide professional medical advice and guidance. This online communication approach not only reduces the strain on hospitals' physical resources and congestion, but also improves the efficiency of medical services.

[0003] Chinese Patent Publication No.: CN116737911A discloses a deep learning-based hypertension question-and-answer method and system, which includes obtaining user voice or text question-and-answer information, performing semantic analysis on the question-and-answer information through a semantic analysis model, and classifying the question-and-answer information based on the results of the semantic analysis to determine the category to which the question-and-answer information belongs; based on the question-and-answer information and the category to which the question-and-answer information belongs, retrieving candidate answer information that matches the question-and-answer information from a preset knowledge graph or question-and-answer knowledge base; extracting the candidate answer information, calculating the confidence of the candidate answer information, and filtering out target answer information from the candidate answer information based on the confidence, wherein the target answer information includes the candidate answer information with the highest confidence. The method of this invention can accurately identify user intent and provide professional hypertension medical solutions.

[0004] Chinese Patent Publication No.: CN116805519A discloses an artificial intelligence-based hypertension data transmission warning system and method, including: a transmission information acquisition module, a database, a reminder target screening module, a measurement reminder management module and a transmission data management module. The transmission information acquisition module collects historical uploaded information of users after logging into the health management platform, transmits the historical uploaded information to the database, stores the historical uploaded information through the database, analyzes the historical uploaded information through the reminder target screening module, screens out target users who need to be reminded to adjust their blood pressure measurement time, adjusts the blood pressure measurement time of the target users through the measurement reminder management module and reminds the users to measure their blood pressure, and stores and manages the blood pressure data uploaded by the users through the transmission data management module, reminds the users to transmit the blood pressure data in a timely manner, and ensures to the greatest extent that the users measure and upload their blood pressure within the optimal measurement time period, thereby improving the reference value of the uploaded data.

[0005] However, there are still the following problems in the prior art:

[0006] When describing symptoms and postoperative recovery, some diseases are relatively niche and the symptom descriptions are highly professional. Due to patients' lack of understanding of the symptoms, the symptom description information sent through the user end is usually unprofessional, may contain missing words, incorrect words or be ambiguous, affecting the subsequent understanding of the symptom description. Summary of the Invention

[0007] To this end, the present invention provides a hypertension management system based on a cloud platform to overcome the problem that when describing symptoms and postoperative recovery conditions, some diseases are relatively niche and the symptom descriptions are highly professional. Due to patients' lack of understanding of the symptoms, the symptom description information sent through the user terminal is usually non-professional and may contain missing words, wrong words or ambiguity, which affects the subsequent understanding of the symptom description.

[0008] To achieve the above objectives, the present invention provides a hypertension management system based on a cloud platform, which includes:

[0009] A preprocessing module is used to extract symptom description information for various diseases that cause hypertension, determine the semantic specificity representation parameters corresponding to each disease based on the semantic differences between the symptom description information corresponding to each disease and the symptom description information in the standard symptom description information database, and classify the semantic specificity categories of the diseases;

[0010] A data pool module, connected to the pre-processing module, is used to determine whether it is necessary to build a symptom description information data pool for the disease type according to the specific category of the disease type;

[0011] A tag extraction module is used to extract the medical record information of the user terminal and set the disease type tag for the user terminal based on the keywords in the medical record information;

[0012] The labeling analysis module is connected to the data pool module and the label extraction module to receive the symptom description information sent by the user terminal, determine the semantic specificity category according to the disease label corresponding to the user terminal, and process the symptom description information, including:

[0013] Determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembled set to match the sample disassembled sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching results, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end;

[0014] Or, send symptom description information to the receiving end;

[0015] The suspected non-standard keywords are decomposed into a set of single elements.

[0016] Furthermore, the preprocessing module determines the semantic specific characterization parameters corresponding to each disease type based on the semantic difference between the symptom description information corresponding to each disease type and the symptom description information in the standard symptom description information database, including:

[0017] To determine the probability of occurrence of each keyword in the symptom description information corresponding to the disease in the symptom description information database;

[0018] To determine the semantic collocation degree in the symptom description information corresponding to the disease;

[0019] The ratio of the occurrence probability to a preset occurrence probability threshold is used to obtain a first semantic specificity representation factor;

[0020] The ratio of the semantic collocation degree to a preset semantic collocation degree threshold is used to obtain a second semantic specificity representation factor;

[0021] The first semantic-specific representation factor and the second semantic-specific representation factor are weighted and summed to determine a semantic-specific representation parameter.

[0022] Furthermore, the pre-processing module divides the diseases into semantically specific categories, wherein:

[0023] If the semantic specificity representation parameter is greater than or equal to the benchmark threshold, the semantic specificity category is classified as a nonspecific semantic category;

[0024] If the semantic specificity representation parameter is less than the benchmark threshold, the semantic specificity category is classified as a specific semantic category.

[0025] Furthermore, the data pool module determines whether it is necessary to construct a symptom description information data pool for the disease type based on the specific category of the disease type, wherein:

[0026] If the semantic specificity category is a non-specific semantic category, it is determined that there is no need to construct a symptom description information data pool;

[0027] If the semantic specificity category is a specific semantic category, it is determined that a symptom description information data pool needs to be constructed.

[0028] Furthermore, the label extraction module sets disease labels for the user terminal, including:

[0029] To extract disease keywords from the medical record information;

[0030] The disease keyword is set as a disease label.

[0031] Furthermore, the annotation analysis module processes the symptom description information, wherein:

[0032] If the semantic specificity category is a specific semantic category, determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembly set to match the sample disassembly sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching result prediction, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end;

[0033] If the semantic specificity category is a non-specific semantic category, the symptom description information is sent to the receiving end.

[0034] Furthermore, the annotation analysis module determines the symptom description information data pool required for initial correction, including:

[0035] Determine the disease type corresponding to the disease type label;

[0036] The symptom description information data pool for the disease type constructed by the data pool module is determined as the symptom description information data pool required for initial correction.

[0037] Furthermore, the annotation analysis module constructs a temporary disassembly set to match the sample disassembly set corresponding to each keyword in the symptom description information data pool, including:

[0038] Using a single word in the suspected non-standard keyword as an element in a temporary disassembly set;

[0039] To determine the probability of occurrence of each element in the sample disassembly set;

[0040] To determine the sample disassembly set corresponding to the maximum probability of occurrence;

[0041] The keywords corresponding to the sample disassembly set are determined as predicted keywords.

[0042] Furthermore, the tag analysis module replaces the suspected non-standard keywords based on the matching result prediction, including:

[0043] To determine the semantic matching degree of the symptom description information after replacing the suspected non-standard keywords with the predicted keywords;

[0044] If the semantic collocation degree is greater than or equal to a predetermined semantic collocation degree threshold of the description information, retaining the symptom description information;

[0045] If the semantic collocation degree is less than a predetermined semantic collocation degree threshold of the description information, reselecting the predicted keyword to replace the corresponding suspected non-standard keyword according to the occurrence probability ranking;

[0046] Among them, the keywords corresponding to the sample disassembly set with high probability of occurrence are preferentially selected as the predicted keywords. Furthermore, the annotation analysis module reorganizes and verifies the obtained symptom description information, including:

[0047] To determine the semantic collocation of several symptom description information after the reorganization operation;

[0048] The symptom description information corresponding to the maximum semantic collocation degree is selected as the corrected symptom description information;

[0049] The reorganization operation includes word segmentation arrangement and keyword replacement.

[0050] Compared with the prior art, the present invention provides a preprocessing module, a data pool module, a label extraction module and an annotation analysis module. The preprocessing module determines the semantic specificity characterization parameters corresponding to each disease type by extracting the semantic difference between the symptom description information and the symptom description information in the standard symptom description information library, and divides the semantic specificity categories of the disease type. The data pool module determines whether it is necessary to construct a symptom description information data pool for the disease type based on the specificity category of the disease type. The label extraction module sets the disease type label for the user end. The annotation analysis module determines the semantic specificity category based on the disease type label corresponding to the user end, and processes the symptom description information, including sending the corrected symptom description information and / or symptom description information to the receiving end. The present invention improves the diagnostic efficiency and accuracy by dividing the semantic specificity categories of the disease type and correcting the symptom description information.

[0051] In particular, the present invention determines the semantic specificity characterization parameters corresponding to each disease type, which are used as the data basis for dividing semantic specificity categories. When describing symptoms, some diseases are more common, and patients usually have a higher degree of understanding of the disease. The symptom description information on the user side is often accurate, and the receiving end can determine the corresponding symptoms without much analysis. However, for some diseases, some diseases are relatively niche, and the symptom description is highly professional. Due to the patient's lack of understanding of the disease, the symptom description information sent by the user side may contain missing words, wrong words, or ambiguity. In addition, it is easy to induce unclear semantics, which affects the subsequent understanding of the symptom description and reduces the diagnostic efficiency and accuracy. Based on this, the present invention determines the semantic specificity characterization parameters corresponding to each disease type according to the semantic differences between the symptom description information corresponding to each disease type and the symptom description information in the standard symptom description information library, divides the semantic specificity categories, and constructs a symptom description information data pool for the specific semantic categories, providing a data basis for subsequent processing of the symptom description information, facilitating the optimization of the symptom description information sent by the user side, and improving the subsequent diagnostic efficiency and accuracy.

[0052] In particular, the present invention determines the semantic specificity category based on the disease label corresponding to the user end, and predicts and replaces the symptom description information that needs to be corrected based on the data in the symptom description information data pool. In actual situations, the symptom description information of specific diseases is relatively professional, and the symptom description information sent by the user end may have problems such as non-standard and non-standard keywords, resulting in unclear semantics. Based on this, in order to enable the receiving end to quickly and accurately receive the symptom description information of the user end, the present invention considers determining a symptom description information data pool for the symptom description information classified as a specific semantic category, and splitting and matching suspected non-standard keywords to predict and replace the suspected non-standard keywords based on the matching results, thereby improving the reliability of optimizing the user end symptom description information and improving the subsequent diagnostic efficiency and accuracy.

[0053] In particular, the present invention reorganizes and verifies the symptom description information after the predictive replacement process. Preferably, in order to ensure the accuracy of the symptom description information received by the receiving end, the symptom description information after the predictive replacement process needs to be verified. Based on this, the present invention performs word segmentation and keyword replacement on the symptom description information after the predictive replacement process, and calculates the semantic collocation degree of several symptom description information after the reorganization operation to determine the corrected symptom description information, thereby improving the subsequent diagnostic efficiency and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 This is a schematic structural diagram of a cloud platform-based hypertension management system according to an embodiment of the invention;

[0055] Figure 2 A logical block diagram for classifying the semantically specific categories of diseases according to an embodiment of the invention;

[0056] Figure 3 A logic block diagram for determining whether it is necessary to construct a symptom description information data pool for the disease according to an embodiment of the invention;

[0057] Figure 4 This is a logic block diagram of replacing the suspected non-standard keywords based on the matching result prediction according to an embodiment of the invention. DETAILED DESCRIPTION

[0058] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0059] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the scope of protection of the present invention.

[0060] It should be noted that, in the description of the present invention, unless otherwise expressly specified or limited, the term "connection" should be understood in a broad sense. For example, it can mean a fixed connection, a detachable connection, or an integral connection; it can mean a mechanical connection or an electrical connection; it can mean a direct connection or an indirect connection through an intermediate medium; it can mean internal communication between two components. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0061] See also Figure 1 , Figure 1 This is a schematic diagram of the structure of a cloud-based hypertension management system according to an embodiment of the invention. The cloud-based hypertension management system of the present invention includes:

[0062] A preprocessing module is used to extract symptom description information for various diseases that cause hypertension, determine the semantic specificity representation parameters corresponding to each disease based on the semantic differences between the symptom description information corresponding to each disease and the symptom description information in the standard symptom description information database, and classify the semantic specificity categories of the diseases;

[0063] A data pool module, connected to the pre-processing module, is used to determine whether it is necessary to build a symptom description information data pool for the disease type according to the specific category of the disease type;

[0064] A tag extraction module is used to extract the medical record information of the user terminal and set the disease type tag for the user terminal based on the keywords in the medical record information;

[0065] The labeling analysis module is connected to the data pool module and the label extraction module to receive the symptom description information sent by the user terminal, determine the semantic specificity category according to the disease label corresponding to the user terminal, and process the symptom description information, including:

[0066] Determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembled set to match the sample disassembled sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching results, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end;

[0067] Or, send symptom description information to the receiving end;

[0068] The suspected non-standard keywords are decomposed into a set of single elements.

[0069] Specifically, the symptom description information can be a sentence or a paragraph. There is no limitation on the source of the symptom description information. For example, some symptom description information sent by the user when consulting on the cloud platform used by the hospital can be obtained in advance with authorization. Of course, other methods can also be used, which will not be repeated here.

[0070] Specifically, there is no limitation on the source of the standard symptom description information database. For example, a medical literature database can be obtained in advance to obtain some symptom description information as corpus, and a standard symptom description information database can be constructed after data cleaning. Alternatively, an existing open source related corpus database can be used as the standard symptom description information database. This will not be elaborated here.

[0071] Specifically, the symptom description information data pool of a single disease can be obtained by pre-obtaining medical literature for the corresponding disease, screening out the symptom description information in the obtained medical literature, and constructing the symptom description information data pool of the corresponding disease after data cleaning.

[0072] Specifically, the preprocessing module determines the semantic specific representation parameters corresponding to each disease type based on the semantic difference between the symptom description information corresponding to each disease type and the symptom description information in the standard symptom description information library, including:

[0073] Used to determine the occurrence probability of each keyword in the symptom description information corresponding to the disease in the symptom description information database. It can be understood that the occurrence probability here is the average of the corresponding occurrence probabilities of each keyword, and the occurrence probability of a single keyword is the probability of the corresponding keyword appearing in each sentence in the symptom description information database;

[0074] To determine the semantic collocation degree in the symptom description information corresponding to the disease type, where the semantic collocation degree is the average of the semantic collocation degrees of the symptom description information corresponding to the disease type;

[0075] The ratio of the occurrence probability to a preset occurrence probability threshold is used to obtain a first semantic specificity representation factor;

[0076] The ratio of the semantic collocation degree to a preset semantic collocation degree threshold is used to obtain a second semantic specificity representation factor;

[0077] The first semantic-specific representation factor and the second semantic-specific representation factor are weighted and summed to determine a semantic-specific representation parameter.

[0078] Specifically, the preset occurrence probability threshold is calculated in advance, and historical symptom description information corresponding to several diseases is obtained in advance to determine the historical occurrence probability of each keyword, and 0.89 times the average value of the historical occurrence probability is set as the preset occurrence probability threshold.

[0079] Specifically, the preset semantic collocation threshold is calculated in advance, and historical symptom description information corresponding to several diseases is obtained in advance to determine the historical semantic collocation, and 0.92 times the average value of the historical semantic collocation is set as the preset semantic collocation threshold.

[0080] Specifically, the sum of the weight coefficients of the first semantic-specific representation factor and the second semantic-specific representation factor is 1, the weight coefficient of the first semantic-specific representation factor is 0.56, and the weight coefficient of the second semantic-specific representation factor is 0.44.

[0081] Specifically, there is no limitation on the method of extracting keywords. For example, the TF-IDF method can be used. The specific steps are as follows:

[0082] Determine the number of occurrences of a word in the corresponding symptom description information and the probability of the number of occurrences relative to the total number of words in the corresponding symptom description information;

[0083] Determine the logarithm of the total number of corresponding symptom descriptions and the number of corresponding symptom descriptions containing the word;

[0084] Determine the product of the probability of occurrence and the logarithm;

[0085] Determine the word corresponding to the maximum product as the symptom keyword corresponding to the symptom description information;

[0086] Of course, the TextRank algorithm, N-gram analysis method, etc. can also be used, or the above methods can be used in combination. It is only necessary to ensure that keywords can be extracted. Those skilled in the art can make a choice based on actual conditions, and will not go into details here.

[0087] It is understandable that there may be more than one keyword in the corresponding symptom description information. In this case, the products can be arranged by size and the first few digits of the products can be determined to be keywords to make the corrected symptom description information more accurate. For example, the selected keywords can be half of the total number of words in the symptom description information.

[0088] Specifically, the present invention determines the semantic specificity characterization parameters corresponding to each disease type, which are used as the data basis for dividing semantic specificity categories. When describing symptoms, some diseases are more common, and patients usually have a higher degree of understanding of the disease. The symptom description information on the user side is often accurate, and the receiving end can determine the corresponding symptoms without much analysis. However, for some diseases, some diseases are relatively niche, and the symptom description is highly professional. Due to the patient's lack of understanding of the disease, the symptom description information sent by the user side may contain missing words, wrong words, or ambiguity. In addition, it is easy to induce unclear semantics, which affects the subsequent understanding of the symptom description and reduces the diagnostic efficiency and accuracy. Based on this, the present invention determines the semantic specificity characterization parameters corresponding to each disease type according to the semantic differences between the symptom description information corresponding to each disease type and the symptom description information in the standard symptom description information library, divides the semantic specificity categories, and constructs a symptom description information data pool for the specific semantic categories, providing a data basis for subsequent processing of the symptom description information, facilitating the optimization of the symptom description information sent by the user side, and improving the subsequent diagnostic efficiency and accuracy.

[0089] See also Figure 2 , Figure 2 This is a logic block diagram of the embodiment of the invention for dividing the semantic specificity categories of the disease. Specifically, the pre-processing module divides the semantic specificity categories of the disease, wherein:

[0090] If the semantic specificity representation parameter is greater than or equal to the benchmark threshold, the semantic specificity category is classified as a nonspecific semantic category;

[0091] If the semantic specificity representation parameter is less than the benchmark threshold, the semantic specificity category is classified as a specific semantic category.

[0092] Specifically, the baseline threshold is selected in the interval [0.63, 0.82].

[0093] See also Figure 3 , Figure 3 This is a logic block diagram of an embodiment of the invention for determining whether it is necessary to build a symptom description information data pool for the disease. Specifically, the data pool module determines whether it is necessary to build a symptom description information data pool for the disease based on the specific category of the disease, wherein:

[0094] If the semantic specificity category is a non-specific semantic category, it is determined that there is no need to construct a symptom description information data pool;

[0095] If the semantic specificity category is a specific semantic category, it is determined that a symptom description information data pool needs to be constructed.

[0096] Specifically, the specific semantic categories correspond one-to-one to the symptom description information data pool, which will not be elaborated here.

[0097] Specifically, the label extraction module sets disease labels for the user side, including:

[0098] To extract disease keywords from the medical record information;

[0099] The disease keyword is set as a disease label.

[0100] Specifically, the disease labels correspond one-to-one to the symptom description information data pool, which will not be elaborated here.

[0101] Specifically, the annotation analysis module processes the symptom description information, where:

[0102] If the semantic specificity category is a specific semantic category, determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembly set to match the sample disassembly sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching result prediction, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end;

[0103] If the semantic specificity category is a non-specific semantic category, the symptom description information is sent to the receiving end.

[0104] Specifically, the present invention determines the semantic specificity category based on the disease label corresponding to the user end, and predicts and replaces the symptom description information that needs to be corrected based on the data in the symptom description information data pool. In actual situations, the symptom description information of specific diseases is relatively professional, and the symptom description information sent by the user end may have problems such as non-standard and non-standard keywords, resulting in unclear semantics. Based on this, in order to enable the receiving end to quickly and accurately receive the symptom description information of the user end, the present invention considers determining a symptom description information data pool for the symptom description information classified as a specific semantic category, and splitting and matching suspected non-standard keywords to predict and replace the suspected non-standard keywords based on the matching results, thereby improving the reliability of optimizing the user end symptom description information and improving the subsequent diagnostic efficiency and accuracy.

[0105] Specifically, the annotation analysis module determines the symptom description information data pool required for initial correction, including:

[0106] Determine the disease type corresponding to the disease type label;

[0107] The symptom description information data pool for the disease type constructed by the data pool module is determined as the symptom description information data pool required for initial correction.

[0108] Specifically, the annotation analysis module constructs a temporary disassembly set to match the sample disassembly set corresponding to each keyword in the symptom description information data pool, including:

[0109] Using a single word in the suspected non-standard keyword as an element in a temporary disassembly set;

[0110] To determine the probability of occurrence of each element in the sample disassembly set;

[0111] To determine the sample disassembly set corresponding to the maximum probability of occurrence;

[0112] The keywords corresponding to the sample disassembly set are determined as predicted keywords.

[0113] Specifically, suspected non-standard keywords are non-standard keywords that may lead to unclear expression of symptom information. For example, "I found that my blood pressure was high during routine examination today, so I took elsabetan urgently." The suspected non-standard keyword in this sentence is "elsabetan", which will not be repeated here.

[0114] Specifically, the elements can be individual Chinese characters or individual English letters, as long as they can exist independently and can be combined, which will not be elaborated here.

[0115] Specifically, in implementation, the elements in the temporary disassembly set include "E, Sha, Bei, Tan", and the elements in the sample disassembly set are "E, Bei, Sha, Tan", so when determining the probability of occurrence, it should be ensured that the elements in the temporary disassembly set exist or similar elements exist in the sample disassembly set to ensure the accuracy of the predicted keywords.

[0116] See also Figure 4 , Figure 4 This is a logic block diagram of replacing the suspected non-standard keywords based on the matching result prediction of an embodiment of the invention. Specifically, the annotation analysis module replaces the suspected non-standard keywords based on the matching result prediction, including:

[0117] To determine the semantic matching degree of the symptom description information after replacing the suspected non-standard keywords with the predicted keywords;

[0118] If the semantic collocation degree is greater than or equal to a predetermined semantic collocation degree threshold of the description information, retaining the symptom description information;

[0119] If the semantic collocation degree is less than a predetermined semantic collocation degree threshold of the description information, reselecting the predicted keyword to replace the corresponding suspected non-standard keyword according to the occurrence probability ranking;

[0120] Among them, the keywords corresponding to the sample disassembly set with a high probability of occurrence are preferentially selected as the predicted keywords.

[0121] Specifically, there is no limitation on the method of determining the semantic collocation. For example, the cosine similarity method can be used to vectorize the keywords in the symptom description information, calculate the cosine similarity between the keywords, and determine the mean cosine similarity as the semantic collocation of the symptom description information. Of course, other methods can also be used, which will not be repeated here.

[0122] Specifically, the closer the cosine similarity value is to 1, the more similar the two vectors are, that is, the higher the semantic compatibility between them. The closer the value is to 0, the less similar the two vectors are, that is, the lower the semantic compatibility between them. Therefore, the closer the average cosine similarity is to 1, the higher the semantic compatibility of the symptom description information is. We will not elaborate on this here.

[0123] Specifically, the purpose of setting the semantic collocation threshold of the descriptive information is to characterize whether the semantic collocation of the symptom description information after replacing the suspected non-standard keywords can meet the standards of daily language. Therefore, the semantic collocation threshold of the descriptive information can be determined based on the preset semantic collocation threshold, and is usually set to between 0.9 and 1.1 times the semantic collocation threshold.

[0124] Specifically, the annotation analysis module reorganizes and verifies the obtained symptom description information, including:

[0125] To determine the semantic collocation of several symptom description information after the reorganization operation;

[0126] The symptom description information corresponding to the maximum semantic collocation degree is selected as the corrected symptom description information;

[0127] The reorganization operation includes word segmentation arrangement and keyword replacement.

[0128] Specifically, there is no limitation on the method of performing keyword replacement. For example, the following method can be used:

[0129] Determine a keyword that needs to be replaced and several similar keywords corresponding to the keyword;

[0130] Replace the corresponding keywords with similar keywords respectively to obtain several replaced symptom description information;

[0131] The word segmentation arrangement includes performing word segmentation processing on the replaced symptom description information, and arranging and combining the obtained word segments to obtain a number of new symptom description information.

[0132] Specifically, there is no limitation on the method of determining similar keywords. For example, similar keywords may be homophones of keywords or synonyms of keywords, which will not be elaborated here.

[0133] Specifically, the present invention reorganizes and verifies the symptom description information after the predictive replacement process. Preferably, in order to ensure the accuracy of the symptom description information received by the receiving end, the symptom description information after the predictive replacement process needs to be verified. Based on this, the present invention performs word segmentation arrangement and keyword replacement on the symptom description information after the predictive replacement process, and calculates the semantic collocation degree of several symptom description information after the reorganization operation to determine the corrected symptom description information, thereby improving the subsequent diagnostic efficiency and accuracy.

[0134] Thus far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it will be readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.

Claims

1. A hypertension management system based on a cloud platform, characterized in that: include: A preprocessing module is used to extract symptom description information for various diseases that cause hypertension, determine the semantic specificity representation parameters corresponding to each disease based on the semantic differences between the symptom description information corresponding to each disease and the symptom description information in the standard symptom description information database, and classify the semantic specificity categories of the diseases; A data pool module, connected to the pre-processing module, is used to determine whether it is necessary to build a symptom description information data pool for the disease type according to the specific category of the disease type; A tag extraction module is used to extract the medical record information of the user terminal and set the disease type tag for the user terminal based on the keywords in the medical record information; The labeling analysis module is connected to the data pool module and the label extraction module to receive the symptom description information sent by the user terminal, determine the semantic specificity category according to the disease label corresponding to the user terminal, and process the symptom description information, including: Determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembled set to match the sample disassembled sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching results, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end; Or, send symptom description information to the receiving end; wherein, the suspected non-standard keywords are decomposed into a single element set; The preprocessing module determines the semantic specific characterization parameters corresponding to each disease type based on the semantic difference between the symptom description information corresponding to each disease type and the symptom description information in the standard symptom description information library, including: To determine the probability of occurrence of each keyword in the symptom description information corresponding to the disease in the symptom description information database; To determine the semantic collocation degree in the symptom description information corresponding to the disease; The ratio of the occurrence probability to a preset occurrence probability threshold is used to obtain a first semantic specificity representation factor; The ratio of the semantic collocation degree to a preset semantic collocation degree threshold is used to obtain a second semantic specificity representation factor; The first semantic-specific representation factor and the second semantic-specific representation factor are weighted and summed to determine a semantic-specific representation parameter.

2. The cloud-based hypertension management system according to claim 1, characterized in that: The pre-processing module divides the disease into semantically specific categories, wherein: If the semantic specificity representation parameter is greater than or equal to the benchmark threshold, the semantic specificity category is classified as a nonspecific semantic category; If the semantic specificity representation parameter is less than the benchmark threshold, the semantic specificity category is classified as a specific semantic category.

3. The cloud-based hypertension management system according to claim 2, characterized in that: The data pool module determines whether it is necessary to build a symptom description information data pool for the disease type based on the specific category of the disease type, wherein: If the semantic specificity category is a non-specific semantic category, it is determined that there is no need to construct a symptom description information data pool; If the semantic specificity category is a specific semantic category, it is determined that a symptom description information data pool needs to be constructed.

4. The cloud-based hypertension management system according to claim 1, characterized in that: The label extraction module sets the disease type label for the user terminal, including: To extract disease keywords from the medical record information; The disease keyword is set as a disease label.

5. The cloud-based hypertension management system according to claim 2, characterized in that: The annotation analysis module processes the symptom description information, wherein: If the semantic specificity category is a specific semantic category, determine the symptom description information data pool required for initial correction, disassemble the suspected non-standard keywords, construct a temporary disassembly set to match the sample disassembly sets corresponding to each keyword in the symptom description information data pool, replace the suspected non-standard keywords based on the matching result prediction, reorganize and verify the obtained symptom description information, obtain the corrected symptom description information, and send the corrected symptom description information to the receiving end; If the semantic specificity category is a non-specific semantic category, the symptom description information is sent to the receiving end.

6. The cloud-based hypertension management system according to claim 1, characterized in that: The annotation analysis module determines the symptom description information data pool required for initial correction. include, Determine the disease type corresponding to the disease type label; The symptom description information data pool for the disease type constructed by the data pool module is determined as the symptom description information data pool required for initial correction.

7. The cloud-based hypertension management system according to claim 1, characterized in that: The annotation analysis module constructs a temporary disassembly set to match the sample disassembly set corresponding to each keyword in the symptom description information data pool, including: Using a single word in the suspected non-standard keyword as an element in a temporary disassembly set; To determine the probability of occurrence of each element in the sample disassembly set; To determine the sample disassembly set corresponding to the maximum probability of occurrence; The keywords corresponding to the sample disassembly set are determined as predicted keywords.

8. The cloud-based hypertension management system according to claim 7, characterized in that: The tag analysis module predicts and replaces the suspected non-standard keywords based on the matching results, including: To determine the semantic matching degree of the symptom description information after replacing the suspected non-standard keywords with the predicted keywords; If the semantic collocation degree is greater than or equal to a predetermined semantic collocation degree threshold of the description information, retaining the symptom description information; If the semantic collocation degree is less than a predetermined semantic collocation degree threshold of the description information, reselecting the predicted keyword to replace the corresponding suspected non-standard keyword according to the occurrence probability ranking; Among them, the keywords corresponding to the sample disassembly set with a high probability of occurrence are preferentially selected as the predicted keywords.

9. The cloud-based hypertension management system according to claim 1, characterized in that: The annotation analysis module reorganizes and verifies the obtained symptom description information, including: To determine the semantic collocation of several symptom description information after the reorganization operation; The symptom description information corresponding to the maximum semantic collocation degree is selected as the corrected symptom description information; The reorganization operation includes word segmentation arrangement and keyword replacement.

Images