Infectious disease dynamic monitoring method and device, electronic equipment and storage medium

Abstract

The application relates to the technical field of infectious disease monitoring, and particularly discloses an infectious disease dynamic monitoring method and device, an electronic device and a storage medium. The method comprises the following steps: acquiring infectious disease report information, wherein the report information comprises a disease name, an onset date and a patient work unit field; analyzing the text of the patient work unit field to obtain a school name segment and a class name segment; performing semantic matching on the school name segment and the school names in a school name data set to determine the standard school identifier corresponding to the school name segment; converting the class name segment into a standard format class name; and constructing multi-dimensional monitoring information comprising a school, a class, a disease name and an onset date based on the standard school identifier and the class name, to support infectious disease early warning analysis at the school level and the class level. In this way, the problems that small-scale outbreaks / cluster infectious diseases cannot be identified by existing automatic systems, school name association analysis is difficult and manual efficiency is low, and the like are solved.

Description

Technical Field

[0001] This application relates to the field of infectious disease surveillance technology, specifically to a method, device, electronic equipment, and storage medium for dynamic monitoring of infectious diseases. Background Technology

[0002] Schools and childcare facilities, due to their dense populations and frequent contact, are high-risk locations for the clustering or outbreak of infectious diseases. Therefore, infectious disease surveillance in these institutions is a crucial link in public health prevention and control, and is essential for achieving early intervention and precise prevention and control.

[0003] Currently, automated surveillance tools widely used in public health practice, such as the China Infectious Disease Automatic Alert and Response System (CIDARS), typically use county (district) level administrative divisions as the smallest spatial unit for anomaly detection. This design results in an excessively broad warning range, making it difficult to effectively identify and warn of infectious disease clusters or outbreaks occurring in small-scale scenarios such as a single school or even a single class. Furthermore, the "Patient's Work Unit" field in my country's infectious disease reporting system is a free-text field, lacking a unified standardized coding system. In practice, numerous alternative names and abbreviations for schools exist (e.g., "City A, District XX No. 1 Middle School" is often written as "XX No. 1 Middle School" or "District No. 1 Middle School"). This non-standardized naming makes it difficult for the existing system to perform effective school-level association and aggregation analysis of cases. To compensate for these shortcomings of CIDARS, grassroots disease control institutions often use manual processing methods in practice, namely, sorting the "Patient's Work Unit" field in the infectious disease report card using tools such as Excel, and then visually identifying potential clusters of infectious diseases. However, this method has two inherent drawbacks: First, in terms of processing efficiency, a medium-sized district or county has a large number of schools and childcare institutions, and with their alternative names and abbreviations, there are numerous valid name entries to process, making manual sorting and visual inspection extremely inefficient and consuming a lot of human resources; Second, in terms of accuracy, manual judgment is affected by subjective factors and it is difficult to accurately distinguish between different schools with similar names (e.g., "XX No. 1 Middle School" may correspond to multiple schools), which can easily lead to the omission of small-scale clusters of infectious diseases due to oversight, and cannot meet the growing demand for informatization, precision and timeliness in infectious disease prevention and control.

[0004] Therefore, there is an urgent need for a technical solution that can accurately and automatically monitor and warn schools and childcare institutions to address the shortcomings of existing technologies, such as insufficient coverage of small-scale scenarios, difficulty in analyzing non-standard texts, and low efficiency of manual methods. Summary of the Invention

[0005] This application discloses a method, device, electronic device, and storage medium for dynamic monitoring of infectious diseases, aiming to overcome the problems of excessively large warning range, difficulties in correlation analysis caused by non-standardized text information, and low efficiency and poor accuracy of manual monitoring in the prior art.

[0006] On one hand, embodiments of this application provide a method for dynamic monitoring of infectious diseases, including: acquiring infectious disease reporting information, the reporting information including disease name, onset date and patient's workplace field; parsing the text of the patient's workplace field to obtain school name fragments and class name fragments; semantically matching the school name fragments with school names in the school name dataset to determine the standard school identifier corresponding to the school name fragments; converting the class name fragments into standard format class names; and constructing multi-dimensional monitoring information including school, class, disease name and onset date based on the standard school identifier and class name to support infectious disease early warning analysis at the school and class levels.

[0007] On one hand, this application provides an infectious disease dynamic monitoring and early warning device, comprising: an information acquisition module for acquiring infectious disease report information, the report information including disease name, onset date and patient's work unit field; a text parsing module for parsing the patient's work unit field text in the report information to obtain school name fragments and class name fragments; a school matching module for semantically matching the school name fragments with school names in the school name dataset to determine the standard school identifier corresponding to the school name fragments; a class encoding module for converting the class name fragments into standard format class names; and a multi-dimensional monitoring module for constructing multi-dimensional monitoring information including school, class, disease name and onset date based on the standard school identifier and class name, to support infectious disease early warning analysis at the school level and class level.

[0008] In one possible embodiment, the text parsing module is used to: establish matching rules including a set of preferred keywords and a set of alternative keywords; based on the matching rules, match the text of the patient's work unit field with the set of preferred keywords and determine whether the match is successful; if so, segment the text into school name fragments and class name fragments according to the matching position; if not, match the text with the set of alternative keywords, and if the match is successful, segment the text into school name fragments and class name fragments according to the matching position.

[0009] In one possible embodiment, the semantic matching matching mode includes at least one of the following modes: Full name string matching mode: Based on a string similarity algorithm, the school name fragment is matched with the school full name field in the school name dataset; Abbreviation string matching mode: Based on a string similarity algorithm, the school name fragment is matched with the school abbreviation or alternative name field in the school name dataset, and the corresponding school full name is returned as the standard school identifier; Hybrid matching mode: The abbreviation string matching mode is executed first, and for school name fragments that fail to match, the semantic similarity model is used to complete the matching; Semantic model matching mode: The semantic similarity model is used to match the school name fragment with the school full name field in the school name dataset.

[0010] In one possible embodiment, the semantic similarity model is a fine-tuned Sentence-BERT model. The fine-tuning process includes: extracting the standard names and their aliases from the standard school dataset, constructing triplet samples with the standard names as anchors, aliases as positive examples, and other irrelevant names as negative examples; inputting the triplet samples into the Sentence-BERT model, and training the model with a preset triplet loss function until the number of training rounds reaches a preset number or the loss value of the triplet loss function is less than the preset loss value.

[0011] In one possible embodiment, the class encoding module is used to: extract grade information and class information from the class name fragment by matching regular expression rules; call the Chinese character number conversion function to convert the Chinese characters in the grade information and class information into corresponding Arabic numerals, so as to obtain the grade number and class number respectively; determine the school segment corresponding to the grade number; and concatenate the school segment, grade number and class number according to a preset format to generate a standard format class name.

[0012] In one possible embodiment, the device further includes: a case statistics module, used to perform statistics on multi-dimensional monitoring information according to a preset sliding time window to obtain the number of cases at the school level and / or class level within the sliding time window; and an early warning triggering module, used to trigger an early warning event corresponding to the number of cases when the number of cases meets a preset threshold condition and the reporting time of the corresponding cases meets a preset timeliness constraint condition, the early warning event including school-level outbreak early warning, school-level cluster early warning or class-level cluster early warning.

[0013] In one possible embodiment, the early warning triggering module is used to: compare the number of cases with the corresponding level early warning thresholds, including class-level cluster thresholds, school-level cluster thresholds, and school-level outbreak thresholds; if the number of cases reaches the corresponding level early warning threshold, then verify whether the time difference between the last reporting time and the current analysis time in the multi-dimensional monitoring information meets the preset timeliness constraint; if so, then trigger the early warning event corresponding to the level and threshold.

[0014] On one hand, embodiments of this application provide an electronic device, which includes a processor and a memory, wherein the memory stores program code, and when the program code is executed by the processor, the processor executes any of the above-mentioned methods for dynamic monitoring of infectious diseases.

[0015] On the one hand, this application provides a computer-readable storage medium including program code, which, when the storage medium is run on an electronic device, causes the electronic device to perform any of the above-mentioned methods for dynamic monitoring of infectious diseases.

[0016] The technical solution provided in this application achieves the following beneficial effects: (1) Breakthrough in monitoring granularity from "county / district level" to "class level": A two-level monitoring model of "school-class" was established by using a complete location identifier consisting of "school identifier + class code". This makes the smallest spatial unit of monitoring and early warning more precise from "county (district)" in the existing CIDARS system to "class", solving the problem that small-scale clustered infectious diseases cannot be identified by the existing automated system and achieving full coverage of monitoring blind spots.

[0017] (2) The problem of accurately collecting school names in free text has been solved: By integrating a multi-mode strategy of "full name / abbreviation string matching" and "deep matching based on fine-tuning semantic model", the ever-changing school nicknames and abbreviations in the report can be mapped to unique standard school identifiers with high accuracy. This process replaces the "visual estimation method" that relies on human memory and judgment, and technically eliminates the data association obstacles caused by non-standard names, providing a standardized data foundation for subsequent accurate analysis.

[0018] (3) Achieved a double leap in monitoring efficiency and accuracy: The fully automated and localized data processing and early warning analysis process replaced the inefficient and error-prone manual "visual inspection method", improving the efficiency of early warning analysis. At the same time, the algorithm logic reduced subjective oversights and improved the objectivity and reliability of monitoring.

[0019] (4) Enables truly precise prevention and control: Since the early warning can be located down to a specific class, the scope of resource allocation for prevention and control measures can be precisely narrowed from the entire school or district to a single or several classes, avoiding excessive "one-size-fits-all" prevention and control, and significantly improving the efficiency of public health resource allocation and the cost-effectiveness of prevention and control measures. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the description of the embodiments are briefly described below. It should be noted that the drawings are merely exemplary embodiments of this application, and those skilled in the art can understand the technical solutions based on the drawings, but they should not be regarded as limitations on the scope of protection.

[0021] Figure 1 This is a schematic diagram of an application scenario in the embodiments of this application; Figure 2 This is a flowchart illustrating the implementation of a dynamic monitoring method for infectious diseases in this application. Figure 3 This is a schematic diagram of the structure of an infectious disease dynamic monitoring device according to an embodiment of this application; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. Unless otherwise specified, the embodiments and features in the embodiments of this application can be arbitrarily combined with each other. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than that shown here.

[0023] The terms "first," "second," etc., that may appear in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in sequences other than those illustrated or described herein.

[0024] The design concept of the embodiments of this application is briefly introduced below: Schools and childcare facilities, due to their high population density and frequent contact, are high-risk locations for infectious disease clusters or outbreaks. Currently, automated surveillance tools widely used in public health practice, such as CIDARS, typically use county (district) level administrative divisions as the smallest spatial unit for anomaly detection. This design results in an overly broad warning range, making it difficult to effectively identify and warn of infectious disease clusters or outbreaks occurring in small-scale settings such as a single school or even a single class. Furthermore, the "patient's workplace" field in my country's infectious disease reporting system is a free-text field, lacking a unified standardized coding system, which makes it difficult for the existing system to perform effective school-level correlation and aggregation analysis of cases. To compensate for these shortcomings of CIDARS, grassroots disease control institutions often use manual processing methods in practice, that is, sorting the "patient's workplace" field in the infectious disease report card using tools such as Excel, and then identifying potential clusters of infectious diseases by visual inspection. However, this method has two inherent drawbacks: First, a medium-sized district or county has a large number of schools and childcare institutions, and with their alternative names and abbreviations, there are numerous valid name entries to process, making manual sorting and visual inspection extremely inefficient and consuming a large amount of human resources. Second, manual judgment is affected by subjective factors, making it difficult to accurately distinguish between different schools with similar names, which can easily lead to the missed detection of small-scale clusters of infectious diseases. Therefore, there is an urgent need for a technical solution that can accurately and automatically monitor and provide early warning for schools and childcare institutions to address the shortcomings of existing technologies, such as insufficient coverage of small-scale scenarios, difficulty in analyzing non-standard texts, and low manual efficiency.

[0025] In view of this, embodiments of this application provide a method, device, electronic device, and storage medium for dynamic monitoring of infectious diseases. The method includes: acquiring infectious disease reporting information, which includes disease name, onset date, and patient's workplace field; parsing the text of the patient's workplace field to obtain school name fragments and class name fragments; semantically matching the school name fragments with school names in a school name dataset to determine the standard school identifier corresponding to the school name fragment; converting the class name fragment into a standard format class name; and constructing multi-dimensional monitoring information including school, class, disease name, and onset date based on the standard school identifier and class name to support infectious disease early warning analysis at the school and class levels. This precisely refines the smallest spatial unit of monitoring and early warning to the "class," solving the problem that small-scale clustered infectious diseases cannot be identified by existing automated systems, and achieving full coverage of monitoring blind spots. In addition, by integrating a multi-modal strategy of "full name / abbreviation string matching" and "deep matching based on fine-tuning semantic model", the ever-changing school nicknames and abbreviations in the report can be mapped to unique standard school identifiers with high accuracy. This technically eliminates the data association obstacles caused by non-standard names and provides a standardized data foundation for subsequent accurate analysis.

[0026] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the embodiments of this application will be described below in conjunction with the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0027] Figure 1 This is a schematic diagram illustrating an application scenario provided in one embodiment of this application. The scenario focuses on localized data processing by grassroots disease prevention and control institutions (hereinafter referred to as "CDCs"). The application scenario diagram includes a terminal device 101 and a server 102. The terminal device 101 and the server 102 communicate via a communication network.

[0028] Terminal device 101 is an electronic device used by disease control institutions. This electronic device can be a personal computer, mobile phone, tablet, laptop, e-book reader, vehicle-mounted terminal, etc. Furthermore, terminal device 101 can have a client application related to infectious disease dynamic monitoring installed. This client application can be software (e.g., an app, browser), a webpage, a mini-program, etc. The target audience (user) can use the aforementioned client application related to infectious disease dynamic monitoring through terminal device 101 to perform operations related to infectious disease dynamic monitoring.

[0029] Server 102 can be a standalone physical server, an edge device 102 in the field of cloud computing, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud storage, cloud functions, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0030] There is no limit to the number of the aforementioned terminal devices 101 and / or servers 102.

[0031] It should be noted that the infectious disease dynamic monitoring method in this embodiment can be executed by either terminal device 101 or server 102 alone, or by both terminal device 101 and server 102. For example, when executed by terminal device 101 alone, terminal device 101 obtains infectious disease report information, which includes the disease name, onset date, and patient's workplace field; it parses the text of the patient's workplace field to obtain school name fragments and class name fragments; it performs semantic matching between the school name fragments and school names in the school name dataset to determine the standard school identifier corresponding to the school name fragments; it converts the class name fragments into standard format class names; and based on the standard school identifiers and class names, it constructs multi-dimensional monitoring information including school, class, disease name, and onset date to support infectious disease early warning analysis at the school and class levels.

[0032] It is important to emphasize that all data obtained in this application embodiment was acquired legally and compliantly, such as report data exported from the internal information system of the disease control agency through legitimate procedures. Furthermore, when all methods and processes are completed independently on the terminal device 101, no continuous network connection or remote server intervention is required. This helps meet the data privacy and security requirements in disease control operations and adapts to grassroots working environments with limited network conditions.

[0033] The following is in conjunction with the appendix Figure 2 This application provides a detailed and complete description of the technical solutions in its embodiments. Obviously, the described embodiments are only a part of, and not all, of these embodiments. Based on this application, those skilled in the art can obtain other embodiments without creative effort.

[0034] like Figure 2 The illustration shows a method for dynamic monitoring of infectious diseases provided in this application, which mainly includes the following steps: S201: Obtain infectious disease reporting information, which includes the disease name, date of onset, and patient's workplace field.

[0035] In this embodiment, raw data is obtained from the infectious disease reporting system (China Disease Prevention and Control Information System). This raw data includes fields such as disease name, onset date, patient's workplace, and population classification. After obtaining this raw data, further preprocessing is performed, including: Cases belonging to a pre-defined target infectious disease list are selected from the raw data. This list covers Class A, B, and C infectious diseases as stipulated in the "Law of the People's Republic of China on the Prevention and Control of Infectious Diseases," as well as key local monitoring diseases such as chickenpox and fever with thrombocytopenia syndrome. Cases belonging to school-related populations (such as students, kindergarten children, and teachers) are also selected from the raw data. Based on this, the filtered report cards can be obtained. Furthermore, duplicate report records are eliminated based on the unique identifier field, i.e., the report card ID.

[0036] After preprocessing, the aforementioned infectious disease reporting information is obtained, which is a standardized data stream used for subsequent analysis.

[0037] S202: Parse the text of the patient's workplace field to obtain fragments of the school name and class name.

[0038] In this embodiment, after obtaining infectious disease report information, geographical prefixes such as "XX City" and "XX District," punctuation marks, and redundant spaces are further removed from the text to form a clean text to be parsed. Then, a matching rule containing a priority keyword set and a candidate keyword set is established to segment and extract school name fragments and class name fragments from the text to be parsed. The specific logic includes: First, define a priority keyword set, which includes general educational entity words such as primary school, middle school, kindergarten, and university; define a candidate keyword set, which includes regular expression patterns adapted to colloquial expressions, such as "Xiaoxiao School" or "Xiaozhong Middle School". Then, perform hierarchical matching extraction: match the text to be parsed with the priority keyword set and determine whether the match is successful; if so, the matched keyword and the text before it are used together as the school name fragment, and the remaining text after the keyword is used as the class name fragment; if not, then match the text to be parsed with the candidate keyword set. If the match is successful, the same logic is used to extract the school name fragment and the class name fragment.

[0039] If neither of the above two types of keywords is matched successfully, the entire cleaned text will be used as the school name fragment, and the class name fragment will be recorded as a missing value.

[0040] Using the methods described above, you can obtain fragments of the school name and class name corresponding to infectious disease reporting information.

[0041] S203: Perform semantic matching between the school name fragment and the school names in the school name dataset to determine the standard school identifier corresponding to the school name fragment.

[0042] In this embodiment of the application, after obtaining the school name fragment and class name fragment corresponding to the infectious disease reporting information, the school name fragment is further semantically matched with the school names in the school name dataset to determine the standard school identifier corresponding to the school name fragment. The semantic matching strategy can be selected according to the field completeness of the standard school name dataset and the requirement for matching accuracy. The specific matching mode includes at least one of the following modes: Full Name String Matching: Based on string similarity algorithms, such as the Jaccard distance algorithm, this method matches school name fragments with the standardized full name field of the school in the school name dataset. This mode is suitable for work scenarios where the school name dataset only contains the standardized full name field of the school and where high matching efficiency is required.

[0043] Abbreviation string matching: Based on string similarity algorithms, such as the Jaccard distance algorithm, this method matches school name fragments with the school abbreviation or alias fields in the school name dataset and returns the corresponding full school name as the standard school identifier. This mode is suitable for working scenarios where the school name dataset contains both the standard full school name and the school abbreviation or alias fields, and the abbreviation / alias field information is complete.

[0044] Hybrid matching: First, perform abbreviation string matching. For school name fragments that fail to match, further use a finely tuned semantic similarity model to complete the matching. This mode is suitable for work scenarios where the school name dataset contains both the standard school name field and the school nickname or abbreviation field, but some school nickname or abbreviation field information is missing, and it is necessary to take into account both matching coverage and accuracy.

[0045] Optionally, semantic model matching directly uses a specially tuned semantic understanding model, such as the Sentence-BERT model. Based on this model, the semantic similarity between the school name fragment and the standardized full name field of the school in the dataset is calculated, and matching is performed according to a preset similarity threshold. This mode is suitable for work scenarios where the dataset only contains the standardized name field, without alternative or abbreviation fields, and where high matching accuracy is required. The fine-tuning process of the semantic understanding model includes: Training data construction: Extract the standard names and their aliases from the standard school dataset, and construct triplet samples with the standard name as the anchor, its alias as positive examples, and other irrelevant names as negative examples.

[0046] Efficient parameter fine-tuning: Using LoRA technology, only the Key and Value projection matrices in the attention mechanism of the pre-trained model are fine-tuned (typical parameters: rank r=8, scaling factor lora_alpha=32).

[0047] Model training: Use the AdamW optimizer (learning rate 2e-5, weight decay 0.01) to train the model for a preset number of rounds with a preset triplet loss function, such as 2 rounds, 3 rounds, or 4 rounds. There is no specific limit here, or train until the loss value of the triplet loss function is less than the preset loss value, so that the model can better distinguish between the canonical name and its aliases, as well as irrelevant names.

[0048] When performing matching applications through the fine-tuned semantic model, generate semantic embedding vectors for the parsed school name fragments and calculate the cosine similarity with the pre-computed standard school canonical name embedding vectors. Take the result with the highest similarity. If its value exceeds the preset threshold, such as 0.7, 0.8, or 0.9, etc., which is specifically adjusted according to the matching accuracy and matching efficiency, then use this canonical name as the matching standard school identifier.

[0049] S204: Convert the class name fragment into a class name in standard format.

[0050] In the embodiments of this application, converting the class name fragment into a class name in standard format includes: extracting grade information and class information from the class name fragment through regular rule matching; calling a Chinese numeral conversion function to convert the Chinese numerals in the grade information and class information into corresponding Arabic numerals, respectively obtaining the grade numeral and the class numeral; determining the school stage corresponding to the grade numeral; and splicing the school stage, grade numeral, and class numeral in a preset format to generate a class name in standard format.

[0051] Specifically, convert the extracted heterogeneous class name fragments, such as "Class 5, Grade 9" and "Grade 9, Class 5", into class names in standard format. This process includes: Regular matching extraction: By designing a regular expression rule group that covers various common expression formats, for example, for formats such as those with parentheses, all Chinese numeral formats, and abbreviation formats, identify and extract the school stage information, grade information (which may be Arabic numerals or Chinese numerals), and class number information from the class name fragment.

[0052] Chinese numeral conversion: Establish a mapping table from Chinese numerals such as "one" to "ninety-nine" to Arabic numerals. When the extracted grade or class information is a Chinese numeral, convert it to an Arabic numeral according to this mapping table.

[0053] School stage mapping and splicing: Automatically determine the school stage it belongs to according to the converted grade value. For example, determine the school stage of grades 1-6 as "primary school", and the school stage of grades 7-9 as "junior high school", etc. Finally, generate a class name in standard format according to the preset format of "school stage + grade numeral + grade + class number + class", such as "Class 5, Grade 3, Junior High School".

[0054] The above method can standardize heterogeneous class names, thereby facilitating subsequent data processing and improving processing efficiency and accuracy.

[0055] S205: Based on standard school identifiers and class names, construct multi-dimensional monitoring information including school, class, disease name, and onset date to support infectious disease early warning analysis at the school and class levels.

[0056] In this embodiment, the standard school identifier obtained in S204 and the standard format class name obtained in S203 constitute the complete location identifier for each case. Based on this, multi-dimensional monitoring information including school, class, disease, and time is constructed. Based on this multi-dimensional monitoring information, further infectious disease early warning analysis is performed, specifically including the following: According to a preset sliding time window, the multi-dimensional monitoring information is statistically analyzed to obtain the number of cases at the school level and / or class level within the sliding time window. Specifically, based on the date of onset, the number of cases with the combination of "school + disease" and "school + class + disease" is counted separately according to a preset sliding time window (e.g., 7 days).

[0057] When the number of cases meets a preset threshold and the reporting time of the corresponding cases meets a preset timeliness constraint, a warning event corresponding to the number of cases is triggered. Warning events include school-level outbreak warnings, school-level cluster warnings, or class-level cluster warnings. Specifically, different warning thresholds are first set, for example: Class-level cluster warning: Within the same class of the same school, the number of cases of a specific disease within a sliding window is ≥2; School-level cluster warning: Within the same school (across all classes), the number of cases of a specific disease within a sliding window is ≥5; School-level outbreak warning: Within the same school (across all classes), the number of cases of a specific disease within a sliding window is ≥10. Then, the number of cases is compared with the corresponding level warning threshold. If the number of cases reaches the corresponding level warning threshold, the time difference between the last reporting time and the current analysis time in the multi-dimensional monitoring information is checked to see if it meets the preset time constraint conditions. For example, it is determined whether the last reporting time is within 3 days before the statistical day. If so, the corresponding level warning event is triggered and warning information is generated. The warning information includes at least the warning type, risk level, location identifier (school and class), number of cases, and time range.

[0058] Based on the above methods, rapid early warning and analysis of infectious diseases at the school and class levels can be achieved.

[0059] It should be noted that each step of the above method can be implemented using various programming languages ​​and toolchains, such as R or Python. As a specific embodiment of the present invention, the inventors have developed and verified a complete implementation program, which specifically includes a semantic model fine-tuning module for school name matching tasks. This module constructs triplet training data (using the standardized school name as the anchor, its aliases as positive examples, and other irrelevant names as negative examples) and employs a triplet loss function and efficient parameter fine-tuning techniques (such as LoRA) to specifically optimize the pre-trained semantic model (such as Sentence-BERT), thereby significantly improving the semantic matching accuracy for school name variants (such as abbreviations and aliases). Those skilled in the art can reproduce or adaptively develop various programs to implement the present invention based on the disclosed fine-tuning method, data processing logic, and other technical features provided in this application. The provider has verified that by processing actual data using the program of this specific embodiment, the matching accuracy of school names can reach a practical level, fully verifying the technical effect and feasibility of the present invention.

[0060] To facilitate understanding, an example is given: The original report states that the patient's workplace is "Class 5, Grade 12, No. 1 Middle School of XX District". After parsing in S202, the school name fragment "Class 5, Grade 12" and the class name fragment "Class 5, Grade 12" are obtained. After matching in S203 (e.g., using string similarity matching), "Class 5, Grade 12" is mapped to the standard school identifier "Class 5, Grade 12, No. 1 Middle School of XX City". After standardization in S204, "Class 5, Grade 12" is converted into the standard format class name "Class 5, Grade 12". In S205, if statistics show that there are 3 cases of chickenpox in Class 5, Grade 12, No. 1 Middle School of XX City in the last 7 days, a "Class-level Cluster Warning" is triggered.

[0061] Based on the same inventive concept, embodiments of this application also provide a device for dynamic monitoring of infectious diseases. For example... Figure 3 The diagram shown is a structural schematic of an infectious disease dynamic monitoring device, which may include: The information acquisition module 301 is used to acquire infectious disease reporting information, which includes the disease name, onset date and patient's workplace field. Text parsing module 302 is used to parse the patient's work unit field text in the report information to obtain school name fragments and class name fragments; The school matching module 303 is used to semantically match school name fragments with school names in the school name dataset in order to determine the standard school identifier corresponding to the school name fragments; The class coding module 304 is used to convert class name fragments into standard format class names; The multidimensional monitoring module 305 is used to construct multidimensional monitoring information, including school, class, disease name, and onset date, based on standard school identifiers and class names, to support infectious disease early warning analysis at the school and class levels.

[0062] In one possible embodiment, the text parsing module 302 is used to: establish a matching rule containing a set of preferred keywords and a set of alternative keywords; based on the matching rule, match the text of the patient's work unit field with the set of preferred keywords and determine whether the match is successful; if so, segment the text into school name fragments and class name fragments according to the matching position; if not, match the text with the set of alternative keywords, and if the match is successful, segment the text into school name fragments and class name fragments according to the matching position.

[0063] In one possible embodiment, the semantic matching matching mode includes at least one of the following modes: Full name string matching mode: Based on a string similarity algorithm, the school name fragment is matched with the school full name field in the school name dataset; Abbreviation string matching mode: Based on a string similarity algorithm, the school name fragment is matched with the school abbreviation or alternative name field in the school name dataset, and the corresponding school full name is returned as the standard school identifier; Hybrid matching mode: The abbreviation string matching mode is executed first, and for school name fragments that fail to match, the semantic similarity model is used to complete the matching; Semantic model matching mode: The semantic similarity model is used to match the school name fragment with the school full name field in the school name dataset.

[0064] In one possible embodiment, the semantic similarity model is a fine-tuned Sentence-BERT model. The fine-tuning process includes: extracting the standard names and their aliases from the standard school dataset, constructing triplet samples with the standard names as anchors, aliases as positive examples, and other irrelevant names as negative examples; inputting the triplet samples into the Sentence-BERT model, and training the model with a preset triplet loss function until the number of training rounds reaches a preset number or the loss value of the triplet loss function is less than the preset loss value.

[0065] In one possible embodiment, the class encoding module 304 is used to: extract grade information and class information from the class name fragment by matching regular expression rules; call the Chinese character number conversion function to convert the Chinese character numbers in the grade information and class information into corresponding Arabic numerals, so as to obtain the grade number and class number respectively; determine the school segment corresponding to the grade number; and concatenate the school segment, grade number and class number according to a preset format to generate a standard format class name.

[0066] In one possible embodiment, the device further includes: a case statistics module, used to perform statistics on multi-dimensional monitoring information according to a preset sliding time window to obtain the number of cases at the school level and / or class level within the sliding time window; and an early warning triggering module, used to trigger an early warning event corresponding to the number of cases when the number of cases meets a preset threshold condition and the reporting time of the corresponding cases meets a preset timeliness constraint condition, the early warning event including school-level outbreak early warning, school-level cluster early warning or class-level cluster early warning.

[0067] In one possible embodiment, the early warning triggering module is used to: compare the number of cases with the corresponding level early warning thresholds, including class-level cluster thresholds, school-level cluster thresholds, and school-level outbreak thresholds; if the number of cases reaches the corresponding level early warning threshold, then verify whether the time difference between the last reporting time and the current analysis time in the multi-dimensional monitoring information meets the preset timeliness constraint; if so, then trigger the early warning event corresponding to the level and threshold.

[0068] The technical effects of the aforementioned infectious disease dynamic monitoring device can be referenced from the infectious disease dynamic monitoring method, and will not be elaborated here.

[0069] In some possible implementations, the infectious disease dynamic monitoring device according to this application may include at least a processor and a memory. The memory stores program code that, when executed by the processor, causes the processor to perform the steps of the infectious disease dynamic monitoring method according to various exemplary embodiments of this application described in this specification. For example, the processor may perform actions such as... Figure 2 The steps are shown in the figure.

[0070] Based on the same inventive concept, this application also provides an electronic device that can realize the functions of the aforementioned method and device for dynamic monitoring of infectious diseases. (Refer to...) Figure 4 Electronic devices include: At least one processor 401 and a memory 402 connected to at least one processor 401. In this embodiment, the specific connection medium between the processor 401 and the memory 402 is not limited. Figure 4 The example shown is the connection between processor 401 and memory 402 via bus 400. Bus 400 is... Figure 4 The connections between other components are shown in bold lines only and are not intended to be limiting. The 400 bus can be divided into address bus, data bus, control bus, etc., for ease of representation. Figure 4 The term is represented by a single thick line, but this does not imply that there is only one bus or one type of bus. Alternatively, processor 401 can also be called a controller; there is no restriction on the name.

[0071] In this embodiment, memory 402 stores instructions executable by at least one processor 401. By executing the instructions stored in memory 402, at least one processor 401 can perform the aforementioned method for dynamic monitoring of infectious diseases. Processor 401 can implement... Figure 3 The functions of each module in the device shown.

[0072] The processor 401 is the control center of the device. It can connect to various parts of the control device through various interfaces and lines. By running or executing instructions stored in memory 402 and calling data stored in memory 402, the processor can perform various functions and process data, thereby monitoring the device as a whole.

[0073] In one possible design, processor 401 may include one or more processing units. Processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly handles the operating system, user interface, and applications, and the modem processor mainly handles wireless communication. It is understood that the modem processor may also not be integrated into processor 401. In some embodiments, processor 401 and memory 402 may be implemented on the same chip; in some embodiments, they may also be implemented on separate chips.

[0074] Processor 401 can be a general-purpose processor, such as a central processing unit (CPU), digital signal processor, application-specific integrated circuit, field-programmable gate array or other programmable logic device, discrete gate or transistor logic device, or discrete hardware component, capable of implementing or executing the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the infectious disease dynamic monitoring method disclosed in the embodiments of this application can be directly manifested as being executed by a hardware processor, or executed by a combination of hardware and software modules within the processor.

[0075] Memory 402, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. Memory 402 may include at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory, random access memory (RAM), static random access memory (SRAM), programmable read-only memory (PROM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), magnetic storage, magnetic disk, optical disk, etc. Memory 402 can be any other medium capable of carrying or storing desired program code in the form of instructions or data structures that can be accessed by a computer, but is not limited thereto. In the embodiments of this application, memory 402 can also be a circuit or any other device capable of implementing storage functions for storing program instructions and / or data.

[0076] By designing and programming the processor 401, the code corresponding to the infectious disease dynamic monitoring method described in the foregoing embodiments can be embedded into the chip, thereby enabling the chip to execute the code during operation. Figure 2 The steps of the infectious disease dynamic monitoring method of the illustrated embodiment are as follows. How to design and program the processor 401 is a technique well-known to those skilled in the art and will not be described further here.

[0077] Based on the same inventive concept, embodiments of this application also provide a storage medium storing computer instructions that, when executed on a computer, cause the computer to perform the aforementioned method for dynamic monitoring of infectious diseases.

[0078] In some possible implementations, various aspects of the infectious disease dynamic monitoring method provided in this application can also be implemented in the form of a program product, which includes program code that, when the program product is run on a device, causes the control device to perform the steps in the infectious disease dynamic monitoring method according to the various exemplary embodiments of this application described above.

[0079] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0080] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, as well as combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0081] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0082] These computer program instructions can also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0083] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for dynamic monitoring of infectious diseases, characterized in that, include: Obtain infectious disease reporting information, which includes fields for disease name, date of onset, and patient's workplace; Parse the text of the patient's workplace field to obtain fragments of the school name and class name; The school name fragment is semantically matched with the school names in the school name dataset to determine the standard school identifier corresponding to the school name fragment; Convert the class name fragment into a standard format class name; Based on the standard school identifier and the class name, multi-dimensional monitoring information including school, class, disease name, and onset date is constructed to support infectious disease early warning analysis at the school and class levels.

2. The method according to claim 1, characterized in that, The text of the patient's workplace field is parsed to obtain school name fragments and class name fragments, including: Establish matching rules that include a priority keyword set and a candidate keyword set; Based on the matching rules, the text of the patient's work unit field is matched with the set of priority keywords, and it is determined whether the match is successful. If so, the text is segmented into the school name fragment and the class name fragment based on the matching position; If not, the text is matched with the set of candidate keywords, and if a match is successful, the text is segmented into the school name fragment and the class name fragment according to the matching position.

3. The method according to claim 1, characterized in that, The semantic matching pattern includes at least one of the following patterns: Full Name String Matching Mode: Based on a string similarity algorithm, the school name fragment is matched with the full name field of the school in the school name dataset; Abbreviation string matching mode: Based on the string similarity algorithm, the school name fragment is matched with the school abbreviation or alias field in the school name dataset, and the corresponding full school name is returned as the standard school identifier; Hybrid matching mode: First, the abbreviation string matching mode is executed, and for school name fragments that fail to match, a semantic similarity model is used to complete the matching; Semantic model matching mode: Using a semantic similarity model, the school name fragment is matched with the full school name field in the school name dataset.

4. The method according to claim 3, characterized in that, The semantic similarity model is a fine-tuned Sentence-BERT model, and the fine-tuning process includes: From the standard school dataset, extract the standard name and its alias, and construct triplet samples with the standard name as the anchor, the alias as positive examples, and other irrelevant names as negative examples; The triplet samples are input into the Sentence-BERT model, and the model is trained using a preset triplet loss function until the number of training rounds reaches a preset number or the loss value of the triplet loss function is less than the preset loss value.

5. The method according to claim 1, characterized in that, The step of converting the class name fragment into a standard format class name includes: The grade and class information are extracted from the class name fragment by matching regular expression rules. Call the Chinese character to numeral conversion function to convert the Chinese characters and numerals in the grade information and class information into corresponding Arabic numerals, and obtain the grade number and class number respectively; Determine the grade level corresponding to the grade number; The school segment, grade number, and class number are concatenated according to a preset format to generate the class name in the standard format.

6. The method according to claim 1, characterized in that, After constructing the multi-dimensional monitoring information including school, class, disease name, and onset date, it also includes: The multi-dimensional monitoring information is statistically analyzed according to a preset sliding time window to obtain the number of cases at the school level and / or class level within the sliding time window; When the number of cases meets a preset threshold condition and the reporting time of the corresponding cases meets a preset timeliness constraint condition, an early warning event corresponding to the number of cases is triggered. The early warning event includes school-level outbreak early warning, school-level cluster early warning, or class-level cluster early warning.

7. The method according to claim 6, characterized in that, When the number of cases meets a preset threshold condition and the reporting time of the corresponding cases meets a preset timeliness constraint condition, an early warning event corresponding to the number of cases is triggered, including: The number of cases is compared with the corresponding tiered warning thresholds, which include class-level cluster thresholds, school-level cluster thresholds, and school-level outbreak thresholds. If the number of cases reaches the corresponding level of early warning threshold, then verify whether the time difference between the last reporting time and the current analysis time in the multi-dimensional monitoring information meets the preset timeliness constraint. If so, then trigger the warning event corresponding to that level and threshold.

8. A dynamic monitoring and early warning device for infectious diseases, characterized in that, include: The information acquisition module is used to acquire infectious disease report information, which includes disease name, onset date and patient's work unit fields; The text parsing module is used to parse the patient's work unit field text in the report information to obtain school name fragments and class name fragments; The school matching module is used to semantically match the school name fragment with the school names in the school name dataset to determine the standard school identifier corresponding to the school name fragment; The class coding module is used to convert the class name fragment into a standard format class name; The multidimensional monitoring module is used to construct multidimensional monitoring information, including school, class, disease name, and onset date, based on the standard school identifier and class name, to support infectious disease early warning analysis at the school and class levels.

9. An electronic device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein when the computer program is executed by the processor, the processor performs the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The device contains a computer program that, when run on an electronic device, causes the electronic device to perform the method as described in any one of claims 1 to 7.

Images