Method and system for analyzing adverse event records based on field quality control report

Abstract

The invention discloses an adverse event recording analysis method and system based on a field quality control report, and relates to the field of computers, and the method comprises the following steps: S1, data collection, S2, data integration, S3, adverse event identification and marking, S4, data copying and synchronization, S5, data analysis and recording, and S6, early warning model establishment. According to the method, adverse event data in various databases related to on-site quality control can be comprehensively integrated by using a computer, adverse events are accurately identified and marked after preliminary cleaning and arrangement, data copy synchronization is performed according to the distribution condition, records are displayed in a chart report form, and the efficiency is high. Finally, possible adverse events are early warned in advance by establishing an early warning model, a more powerful basis is provided for timely finding and solving quality problems, the field quality management level is improved, and the problems that the occurrence state of the adverse events cannot be monitored in real time, and early warning cannot be given out in time to prevent further expansion of the adverse events are effectively solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of computer, in particular to a method for analyzing adverse event records based on field quality control reports and a system thereof. BACKGROUND

[0002] The field quality control report is a record and summary of the field quality control situation. It contains the monitoring data, analysis results and improvement measures of various quality indicators in the field, provides basis for timely finding and solving quality problems, and has important significance for improving the field quality management level.

[0003] In the field quality control report, the existing technology usually only collects, analyzes and records data, cannot monitor the occurrence dynamics of adverse events in real time, cannot issue early warning in time to prevent further expansion of adverse events, and only through traditional data processing means, it is difficult to conduct in-depth analysis and mining for complex adverse event causes, resulting in lack of pertinence and effectiveness when formulating improvement measures, which is relatively inconvenient. SUMMARY

[0004] The purpose of the present application is to provide a method for analyzing adverse event records based on field quality control reports and a system thereof to solve the problems raised in the background.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a method for analyzing adverse event records based on field quality control reports, comprising the following steps:

[0006] S1: data collection, connecting various databases related to field quality control, and extracting adverse event related data from these databases;

[0007] S2: data integration, preliminarily cleaning and arranging the extracted data, removing duplicate data, correcting error data, and integrating the arranged data into a database;

[0008] S3: adverse event identification and marking, based on the pre-set adverse event definition and characteristics, using data query and analysis technology in the integrated database to identify possible adverse event records, and marking the identified adverse event, assigning a unique identifier to each adverse event;

[0009] S4: data replication and synchronization, according to the distribution and requirements of adverse event records, replicating and synchronizing related data between different databases or database nodes;

[0010] S5: Data analysis and record, in-depth analysis and mining of adverse event data after replication and synchronization, using statistical analysis, data mining algorithm to find the occurrence law, trend and related factors of adverse events, and displaying the analysis results in the form of charts and reports, and recording and archiving;

[0011] S6: Establishing an early warning model, an adverse event prediction model is established through data analysis to early warn possible adverse events.

[0012] Preferably, the data collection in step S1 comprises the following steps:

[0013] S11: Connecting to various databases using a connection driver;

[0014] S12: Evaluating and classifying the severity and frequency of various adverse events related to on-site quality control, developing evaluation standards and grading system;

[0015] S13: Collecting historical adverse event data for each data source and establishing a feature database for the data source;

[0016] S14: According to the evaluation standard and the feature database of the data source, calculate the priority weight of each data source, and sort and grade the calculated priority weight to determine the priority order of different levels;

[0017] S15: In the data collection process, different data sources are collected in order according to the priority order, and the data sources with high priority are collected first;

[0018] S16: Establishing a real-time monitoring mechanism to monitor the status and priority changes of the data source at any time, and dynamically adjusting the collection order and priority according to the actual situation.

[0019] Preferably, the data integration in step S2 comprises the following steps:

[0020] S21: Data preliminary cleaning, comparing the collected data with the data records, identifying and deleting duplicate data items, verifying the data, identifying format errors and logical inconsistencies, and making appropriate corrections;

[0021] S22: Data standardization, unifying the data formats from different sources, and mapping the fields with the same and similar meanings in different databases;

[0022] S23: Data association and merging, associating and merging information from different data sources to generate a comprehensive data set containing all related information, and storing the sorted comprehensive data set in a designated database.

[0023] Preferably, the adverse event identification and marking in step S3 comprises the following steps:

[0024] S31: defining adverse events, writing and reviewing standard documents about adverse events, and determining the criteria for considering adverse events;

[0025] S32: establishing predefined rules, designing specific IF-THEN rules according to the definition and characteristics of adverse events;

[0026] S33: applying identification, applying the predefined IF-THEN rules to each record extracted one by one, identifying adverse events that meet the conditions, marking records that meet the adverse event conditions, and assigning a unique identifier to each adverse event;

[0027] S34: automatically reviewing the automatically marked adverse events, and adjusting and optimizing the IF-THEN rules according to the review results.

[0028] Preferably, the IF-THEN rule in step S32 is specifically:

[0029] R K : IF (f1∈S k1 ) ∧ (f2∈S k2 ) ∧... ∧ (f m ∈S km ) THEN (y=c k )

[0030] wherein f i is a feature in each data in a data set D, R k is a rule for each data, S Ki is the value range of the feature f i , y is the classification result, and c k is the label of the kth adverse event.

[0031] Preferably, the unique identifier assigned in step S33 adopts a hash function, specifically:

[0032] ID=Hash(f1,f2,...,f m )

[0033] wherein f i is the feature value of the record, and Hash is a hash function.

[0034] Preferably, the data replication and synchronization monitoring in step S4 comprises the following steps:

[0035] S41: determining the adverse event definition and characteristics of data replication pre-set, analyzing data distribution and formulating replication strategies, and determining the data range to be replicated, target database, and replication frequency.

[0036] S42: By the master-slave replication function of the database built-in MySQL, screening and searching are performed according to the set adverse event definition and characteristics, and records possibly meeting the conditions are found out.

[0037] Preferably, in the step S5, the analysis result is displayed in a chart form by using the Matplotlib data visualization tool, and the analyzed result is interpreted to determine the occurrence rule, possible inducement and trend of the adverse event.

[0038] The application also provides a system for analyzing adverse event records based on on-site quality control reports, comprising:

[0039] A data acquisition module is configured to connect to various databases related to on-site quality control and extract adverse event related data from the databases;

[0040] A data integration module is configured to preliminarily clean and arrange the extracted data, remove duplicate data, correct error data, and integrate the arranged data into a database;

[0041] An adverse event identification and marking module is configured to identify possible adverse event records and mark the identified adverse events based on pre-set adverse event definition and characteristics by using data query and analysis techniques in the integrated database, and assign a unique identifier to each adverse event;

[0042] A data replication and synchronization module is configured to replicate and synchronize related data between different databases or database nodes according to the distribution of adverse event records and requirements;

[0043] A data analysis and record module is configured to deeply analyze and mine the adverse event data after replication and synchronization, discover the occurrence rule, trend and associated factors of adverse events by using statistical analysis and data mining algorithms, display the analysis result in a chart or report form, and record and archive the result;

[0044] An early warning module is configured to establish an adverse event prediction model by data analysis, and early warn possible adverse events.

[0045] Preferably, the early warning module comprises:

[0046] A data input and processing submodule is configured to receive and process real-time data, clean and standardize the data;

[0047] A feature engineering submodule is configured to select and construct features that have an impact on prediction;

[0048] A prediction model construction submodule is configured to select a machine learning model and train and verify the model;

[0049] A pre-warning threshold setting sub-module is configured to set a probability threshold for triggering a pre-warning and adjust sensitivity;

[0050] A pre-warning triggering and response sub-module is configured to automatically trigger a pre-warning and define a response measure;

[0051] A result recording and feedback sub-module is configured to record a pre-warning event and a processing result and perform feedback analysis;

[0052] A visualization and reporting sub-module is configured to visualize a pre-warning result and generate an analysis report.

[0053] Technical effects and advantages of the present application:

[0054] The present application can comprehensively integrate adverse event data in various databases related to on-site quality control by using a computer, accurately identify and mark adverse events after preliminary cleaning and arrangement, replicate and synchronize data according to distribution, deeply analyze and mine adverse event rules, and display and record in the form of charts and reports, so as to provide a more powerful basis for timely discovery and solution of quality problems by establishing a pre-warning model to pre-warn possible adverse events, improve on-site quality management level, and effectively solve the problems of inability to monitor adverse event dynamics in real time, inability to timely issue a pre-warning to prevent adverse events from further expanding, and inability to deeply analyze and mine complex adverse event causes by using traditional data processing means, resulting in lack of pertinence and effectiveness of improvement measures. BRIEF DESCRIPTION OF DRAWINGS

[0055] Figure 1 A method flowchart for analyzing adverse event records based on on-site quality control reports is provided.

[0056] Figure 2 A data collection flowchart is provided.

[0057] Figure 3 A data integration flowchart is provided.

[0058] Figure 4 An adverse event identification and marking flowchart is provided.

[0059] Figure 5 A data replication and synchronization flowchart is provided.

[0060] Figure 6 A system block diagram for analyzing adverse event records based on on-site quality control reports is provided.

[0061] Figure 7 A pre-warning module block diagram is provided.

[0062] In the figure: 1, data acquisition module; 2, data integration module; 3, adverse event identification and marking module; 4, data replication and synchronization module; 5, data analysis and recording module; 6, early warning module; 601, data input and processing submodule; 602, feature engineering submodule; 603, prediction model construction submodule; 604, early warning threshold setting submodule; 605, early warning trigger and response submodule; 606, result recording and feedback submodule; 607, visualization and reporting submodule. DETAILED DESCRIPTION

[0063] The technical solutions in the embodiments of the present application will be described clearly and completely below in combination with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0064] The present application provides an adverse event record analysis method based on on-site quality control report analysis as shown in the figure, comprising the following steps: Figures 1-7

[0065] S1: data acquisition, connecting various databases related to on-site quality control, and extracting adverse event related data from these databases;

[0066] S2: data integration, preliminarily cleaning and arranging the extracted data, removing duplicate data, correcting error data, and integrating the arranged data into a database;

[0067] S3: adverse event identification and marking, based on pre-set adverse event definition and characteristics, using data query and analysis technology in the integrated database to identify possible adverse event records, and marking the identified adverse events, assigning a unique identifier to each adverse event;

[0068] S4: data replication and synchronization, according to the distribution of adverse event records and requirements, replicating and synchronizing related data between different databases or database nodes;

[0069] S5: data analysis and recording, in-depth analysis and mining of adverse event data after replication and synchronization, using statistical analysis and data mining algorithms to find the occurrence regularity, trend and associated factors of adverse events, and displaying the analysis results in the form of charts and reports, and recording and archiving;

[0070] S6: establishing an early warning model, establishing an adverse event prediction model through data analysis to early warn possible adverse events.

[0071] ​Through the above steps, the adverse event data in various databases related to on-site quality control can be comprehensively integrated, the adverse events are accurately identified and marked after preliminary cleaning and arrangement, the data is copied and synchronized according to the distribution, the adverse event rules are analyzed and mined in depth, and the records are displayed in the form of charts and reports, finally the possible adverse events are early warned through the establishment of a warning model, which provides a stronger basis for timely discovery and solution of quality problems, improves the on-site quality management level, effectively solves the problems that the adverse event occurrence dynamics cannot be monitored in real time, the warning cannot be sent in time to prevent the adverse events from further expanding, and the causes of complex adverse events are difficult to be deeply analyzed and mined through traditional data processing means, resulting in lack of pertinence and effectiveness of the improvement measures

[0072] Further, the data collection in step S1 includes the following steps:

[0073] S11: connecting with various databases by using a connection driver;

[0074] S12: evaluating and classifying the severity and frequency of various adverse events related to on-site quality control, and formulating evaluation standards and grading system;

[0075] S13: collecting historical adverse event data of each data source and establishing a feature database of the data source;

[0076] S14: calculating the priority weight of each data source according to the evaluation standard and the feature database of the data source, sorting and grading the calculated priority weight, and determining the priority order of different levels;

[0077] S15: in the data collection process, different data sources are collected in sequence according to the priority order, and the data source with high priority is collected first;

[0078] S16: establishing a real-time monitoring mechanism to monitor the state and priority change of the data source at any time, and dynamically adjusting the collection order and priority according to the actual situation.

[0079] Step S11 can ensure the comprehensiveness and universality of data collection, and obtain a large amount of adverse event data related to on-site quality control. Secondly, step S12 evaluates and classifies the severity and frequency of adverse events, and formulates evaluation criteria and grading system, which helps to focus on key adverse events more targetedly, provides clear direction for subsequent analysis and processing, step S13 establishes the feature database of data source, which can provide basis for subsequent priority weight calculation and other operations of data source, step S14 calculates priority weight and sorts and grades, which can make data collection more planned and focused, and collect data sources with high priority first, improve the efficiency and quality of data collection, step S15 collects in turn according to the priority order, which avoids disorganized data collection, so that key data can be collected first, and step S16 establishes real-time monitoring mechanism, which can dynamically adjust the collection order and priority according to the actual situation, so as to ensure that data collection can always adapt to the changing needs of on-site quality control, and provide solid data basis and strong guarantee for subsequent comprehensive and accurate analysis of adverse events and improvement of on-site quality management level

[0080] Further, the data integration in step S2 includes the following steps:

[0081] S21: preliminary data cleaning, comparing the collected data with the data records, identifying and deleting duplicate data items, checking the data, identifying format errors and logical inconsistencies, and making corresponding corrections. By comparing and deleting duplicate data items, checking and correcting format errors and logical inconsistencies, the quality and accuracy of the data can be effectively improved, and redundant and incorrect information can be removed, laying a solid foundation for subsequent data processing and analysis;

[0082] S22: data standardization, unifying the data format from different sources, mapping the fields with the same or similar meanings in different databases, so that the data from different sources can be unified in format, and the fields with the same or similar meanings are mapped, which facilitates unified processing and analysis in subsequent operations, avoids confusion and errors caused by data format differences, and improves the compatibility and operability of the data;

[0083] S23: data association and merging, associating and merging information from different data sources to generate a comprehensive data set containing all related information, and then storing the sorted comprehensive data set in a specified database. Integrating information from different data sources into a comprehensive data set realizes data integration and sharing, can more comprehensively reflect the relevant situation of on-site quality control, provides more rich and accurate data sources for deeper data analysis and decision-making, and helps to more comprehensively grasp the overall situation of on-site quality control, which is of great significance to improving on-site quality management level.

[0084] Further, the adverse event identification and marking in step S3 includes the following steps:

[0085] S31: defining adverse events, writing and reviewing standard documents about adverse events, defining the standards for considering adverse events, defining the standards for considering adverse events, providing clear definition and basis for subsequent identification work, avoiding ambiguity and ambiguity in the definition of adverse events, and making the identification work have rules to follow;

[0086] S32: establishing predefined rules, designing specific IF-THEN rules according to the definition and characteristics of adverse events, designing specific IF-THEN rules according to the definition and characteristics of adverse events, making the identification process more systematic and standardized, and accurately capturing adverse events that meet the conditions, improving the accuracy and efficiency of identification;

[0087] S33: application identification, applying the predefined IF-THEN rules to each record extracted, identifying adverse events that meet the conditions, and marking records that meet the adverse event conditions, and assigning a unique identifier to each adverse event, which can carefully find adverse events that meet the conditions and mark and assign a unique identifier, ensuring accurate identification and tracking of adverse events, and providing clear objects for subsequent analysis and processing;

[0088] S34: automatically reviewing the automatically marked adverse events, adjusting and optimizing the IF-THEN rules according to the review results, so that the rules can continuously adapt to changes in actual conditions, continuously improve the accuracy and adaptability of adverse event identification, and better serve the on-site quality management, and provide more reliable support for timely discovery and solution of quality problems.

[0089] Further, the IF-THEN rule in step S32 is specifically:

[0090] R K : IF (f1∈S k1 ) ∧ (f2∈S k2 ) ∧... ∧ (f m ∈S km ) THEN (y=c k )

[0091] In a data set D, f i is a feature in each data, R k is a rule for each data, S Ki is the value range of the feature f i , y is the classification result, and c k is the label of the kth adverse event.

[0092] The unique identification in step S33 adopts a hash function, specifically:

[0093] ID = Hash (f1, f2,..., f m )

[0094] Wherein f i is the characteristic value of the record, and Hash is a hash function.

[0095] Further, the data replication and synchronization monitoring in step S4 includes the following steps:

[0096] S41: Determine the pre-set adverse event definition and characteristics of data replication, analyze data distribution and develop replication strategies, and clearly define the data range, target database and replication frequency that need to be replicated, so that the data replication work has a clear goal and direction. Clearly defining the data range that needs to be replicated can avoid unnecessary data replication, improve replication efficiency, and at the same time ensure the timeliness and consistency of the data;

[0097] S42: Through the built-in MySQL master-slave replication function of the database, filter and search according to the set adverse event definition and characteristics, and find out records that may meet the conditions. This can quickly and accurately find out records that may meet the conditions with the help of the professional functions of the database. This greatly improves the accuracy and efficiency of data replication, reduces the workload and error rate of manual screening, makes the data replication work more efficient and reliable, and provides a more accurate and complete data basis for subsequent in-depth analysis and processing of data. It is conducive to better monitoring and management of adverse events and improves the level of on-site quality management.

[0098] In step S5, the analysis results are displayed in the form of charts using the Matplotlib data visualization tool, and the analyzed results are interpreted to determine the occurrence pattern, possible causes and trends of adverse events. Displaying the analysis results in the form of charts can make the analysis results more intuitive and visual. Compared with simple text description, charts can help people quickly and clearly grasp the characteristics and trends of the data, such as the occurrence pattern, possible causes and trends of adverse events. Through intuitive chart display and interpretation, the occurrence point and potential problems of adverse events can be accurately located, thereby providing a strong basis for subsequent quality management decisions, such as strengthening monitoring for a specific time period, taking corrective measures for a specific cause, or adjusting resource allocation according to trends, to effectively reduce the probability of adverse events and improve the level of on-site quality management.

[0099] The present application also provides a system for analyzing adverse event records based on on-site quality control reports, comprising:

[0100] Data collection module 1, for establishing connections with various databases related to on-site quality control, and extracting adverse event-related data from these databases;

[0101] Data integration module 2, for preliminary cleaning and sorting of the extracted data, removing duplicate data, correcting erroneous data, and integrating the sorted data into a database;

[0102] Adverse event identification and labeling module 3, based on pre-set adverse event definitions and characteristics, uses data query and analysis techniques in the integrated database to identify possible adverse event records, and labels the identified adverse events, assigning each adverse event a unique identifier;

[0103] Data replication and synchronization module 4, for replicating and synchronizing relevant data between different databases or database nodes according to the distribution of adverse event records and requirements;

[0104] Data analysis and record module 5, for in-depth analysis and mining of adverse event data after replication and synchronization, using statistical analysis and data mining algorithms to discover the occurrence patterns, trends and associated factors of adverse events, and presenting the analysis results in the form of charts and reports, and recording them;

[0105] Early warning module 6, through data analysis to establish an adverse event prediction model, for early warning of possible adverse events.

[0106] Early warning module 6 includes:

[0107] Data input and processing submodule 601, for receiving and processing real-time data, performing data cleaning and standardization, ensuring the accuracy and consistency of data input to subsequent modules, providing a reliable data foundation for subsequent analysis and prediction work, and avoiding errors caused by data quality problems;

[0108] Feature engineering submodule 602, for selecting and constructing features that have an impact on prediction, which helps to highlight key information, improve the accuracy and efficiency of the prediction model, avoid interference from irrelevant or redundant features, and make the model more focused on predicting key information such as adverse events;

[0109] Prediction model construction submodule 603, for selecting machine learning models and training and validating them, which can select the most suitable model according to different data characteristics and prediction requirements, improve the performance and generalization ability of the model through training and validation, and more accurately predict the occurrence of adverse events and other situations;

[0110] The early warning threshold setting submodule 604 is used for setting a probability threshold of early warning triggering and adjusting sensitivity, and can flexibly set the early warning standard according to the actual situation, can avoid false positives and can discover potential problems in time, makes the early warning mechanism more accurate and effective, and guarantees the reliability and practicality of the system;

[0111] The early warning triggering and response submodule 605 is used for automatically triggering early warning and defining response measures, can timely issue an alarm when an adverse event occurs, and at the same time, can clearly define the relevant response measures, improves the response speed and processing efficiency of the adverse event, and reduces the loss possibly caused by the adverse event;

[0112] The result recording and feedback submodule 606 is used for recording the early warning event and processing result, and performing feedback analysis, is helpful for tracking and evaluating the operation of the early warning system, provides a basis for subsequent optimization and improvement, and at the same time, can provide reference experience for processing of subsequent similar events;

[0113] The visualization and reporting submodule 607 is used for visualizing the early warning result and generating an analysis report, the visualized early warning result makes the early warning information more intuitive and easy to understand, facilitates relevant personnel to quickly understand the situation and make decisions, and the generated report also provides detailed data for subsequent data analysis and management.

[0114] Finally, it should be noted that: the above only describes the preferred embodiments of the present application, and is not used to limit the present application, although the present application is described in detail with reference to the foregoing embodiments, for those skilled in the art, the technical solutions recorded in the foregoing embodiments can still be modified, or some technical features can be replaced, any modification, equivalent replacement, improvement, etc. within the spirit and principles of the present application, should be included in the protection scope of the present application.

Claims

1. A method for analyzing adverse event records based on on-site quality control reports, characterized by, The method comprises the following steps: S1: data collection, connecting various databases related to on-site quality control, and extracting adverse event related data from these databases; S2: data integration, preliminary cleaning and sorting of the extracted data, removing duplicate data, correcting error data, and integrating the sorted data into a database; S3: adverse event identification and labeling, based on pre-set adverse event definitions and characteristics, using data query and analysis techniques in the integrated database to identify possible adverse event records, and labeling the identified adverse event, assigning a unique identifier to each adverse event; S4: data replication and synchronization, according to the distribution of adverse event records and requirements, replicating and synchronizing related data between different databases or database nodes; S5: data analysis and record, in-depth analysis and mining of adverse event data after replication and synchronization, using statistical analysis and data mining algorithms to find the occurrence regularity, trend and associated factors of adverse events, and displaying the analysis results in the form of charts and reports, and recording them; S6: Establishing an early warning model, establishing an adverse event prediction model through data analysis to early warn possible adverse events.

2. The method for analyzing adverse event records based on on-site quality control reports according to claim 1, wherein, The data collection in step S1 comprises the following steps: S11: connecting with various databases using connection drivers; S12: evaluating and classifying the severity and frequency of various adverse events related to on-site quality control, developing evaluation criteria and grading system; S13: for each data source, collect its historical adverse event data, and establish a feature database of the data source; S14: according to the evaluation criteria and the feature database of the data source, calculate the priority weight of each data source, and sort and grade the calculated priority weight to determine the priority order of different levels; S15: during data collection, collect different data sources according to the priority order, and collect the data sources with high priority first; S16: Establish a real-time monitoring mechanism to monitor the status and priority changes of the data source at any time, and dynamically adjust the collection order and priority according to the actual situation.

3. The method for analyzing adverse event records based on on-site quality control reports according to claim 1, wherein, The data integration in step S2 comprises the following steps: S21: preliminary data cleaning, comparing the collected data with the data records, identifying and deleting duplicate data items, checking the data, identifying format errors and logical inconsistencies, and making appropriate corrections; S22: data standardization, unifying the data formats from different sources, and mapping the fields with the same or similar meanings in different databases; S23: data association and merging, associating and merging information from different data sources to generate a comprehensive data set containing all related information, and storing the sorted comprehensive data set in a designated database.

4. The method for analyzing adverse event records based on on-site quality control reports according to claim 1, wherein, The adverse event identification and labeling in step S3 comprises the following steps: S31: define adverse events, write and review standard documents about adverse events, and clearly define the standards for adverse events; S32: Establish pre-defined rules, design specific IF-THEN rules based on the definition and characteristics of adverse events; S33: Application identification, for each record extracted, apply the pre-defined IF-THEN rule, identify the adverse events that meet the conditions, and for the records that meet the adverse event conditions, mark and assign a unique identifier to each adverse event; S34: Automatic review of automatically marked adverse events, and according to the review results, adjust and optimize the IF-THEN rule.

5. The method for analyzing adverse event records based on on-site quality control reports according to claim 4, wherein, The IF-THEN rule in step S32 is specifically: R K : IF (f1∈S k1 ) ∧ (f2∈S k2 ) ∧... ∧ (f m ∈S km ) THEN (y = c k ) It is in a data set D, f i For each feature in the data, R k For each rule of the data, S Ki is the value range of the feature f i y is the classification result, c k is the label of the kth adverse event.

6. The method for analyzing adverse event records based on on-site quality control reports according to claim 4, wherein, The unique identifier assigned in step S33 uses a hash function, specifically: ID = Hash (f1, f2,..., f m ) where f i Recorded feature values, Hash is a hash function.

7. The method for analyzing adverse event records based on on-site quality control reports according to claim 1, wherein, The data replication and synchronization in step S4 includes the following steps: S41: Determine the pre-set adverse event definition and characteristics of data replication, analyze data distribution and develop a replication strategy, and clearly define the data range, target database and replication frequency that need to be replicated; S42: Through the master-slave replication function of MySQL built-in database, filter and search according to the set adverse event definition and characteristics, and find out the records that may meet the conditions.

8. The method for analyzing adverse event records based on on-site quality control reports according to claim 1, wherein, In step S5, the analysis results are displayed in the form of charts using Matplotlib data visualization tools, and the results are interpreted to clarify the occurrence rules, possible causes and trends of adverse events.

9. A system for analyzing adverse event records based on site quality control reports, the system comprising: It includes: A data acquisition module (1) for connecting various databases related to on-site quality control and extracting adverse event related data from these databases; A data integration module (2) for preliminary cleaning and sorting of extracted data, removing duplicate data, correcting error data, and integrating sorted data into a database; An adverse event identification and marking module (3) that uses data query and analysis techniques based on pre-set adverse event definitions and characteristics to identify possible adverse event records and mark identified adverse events, and assign a unique identifier to each adverse event; A data replication and synchronization module (4) for replicating and synchronizing related data between different databases or database nodes according to the distribution of adverse event records and requirements; A data analysis and record module (5) for in-depth analysis and mining of adverse event data after replication and synchronization, using statistical analysis and data mining algorithms to discover adverse event occurrence rules, trends and associated factors, and displaying analysis results in the form of charts and reports, and recording archives; An early warning module (6) that establishes an adverse event prediction model through data analysis to provide early warning of possible adverse events.

10. The system for analyzing adverse event records based on live quality control reports according to claim 9, wherein, The early warning module (6) includes: A data input and processing sub-module (601) for receiving and processing real-time data, performing data cleaning and standardization; A feature engineering sub-module (602) for selecting and constructing features that affect prediction; A prediction model construction sub-module (603) for selecting machine learning models and training and validating them; A warning threshold setting sub-module (604) for setting the probability threshold for triggering an early warning and adjusting the sensitivity; An early warning trigger and response sub-module (605) for automatically triggering an early warning and defining response measures; A result recording and feedback submodule (606) is configured to record the early warning events and processing results, and perform feedback analysis. A visualization and reporting submodule (607) is configured to visualize the early warning results and generate an analysis report.

Images