A data processing method and system of an old comorbidity health information cloud platform

Abstract

The application relates to the technical field of health information management, and discloses an old-age comorbidity health information cloud platform data processing method and system. Through receiving data flow from a medical device and performing preliminary analysis, the old-age comorbidity health information cloud platform data processing method and system can identify byte sequences that do not conform to preset rules in the data flow, and obtain the positions, lengths and associated device identification information of the byte sequences. After separating the non-standard byte sequences, core measurement data can continue to be transmitted, so that the high availability of the platform is ensured.

Description

Technical Field

[0001] This invention relates to the field of health information management technology, and more specifically, to a data processing method and system for an elderly comorbidity health information cloud platform. Background Technology

[0002] For health information management of elderly patients with comorbidities, existing cloud platforms face challenges such as incomplete data storage and low retrieval efficiency when processing non-standard health data generated by medical devices. In particular, when non-standard metadata is quietly introduced by medical device firmware updates, the platform's data access layer cannot accurately identify this information, hindering subsequent data processing and complex query analysis.

[0003] To ensure high platform availability, the data access layer was designed with a priority on ensuring the complete inflow of core physiological indicator data. This means that even if some metadata formats are problematic, the system will not reject the transmission of the entire data packet. When the protocol parser encounters unrecognized metadata fields, its built-in fault tolerance mechanism is triggered. This mechanism marks and temporarily stores the unrecognizable metadata portions, while continuing to pass core physiological measurement data (such as heart rate, blood pressure, and blood glucose levels) to the downstream data cleaning stage. However, faced with the aforementioned persistent non-standard metadata issues, this fault tolerance mechanism has resulted in a continuous influx of health data lacking critical device background information into the data processing flow.

[0004] Due to this fault-tolerant design of the data access layer, the raw data area gradually stored a large number of health records with incomplete metadata. While these data retained core physiological measurements, the "incompleteness" of their context caused significant problems. In subsequent layered processing of the data lake, especially when migrating data to the normalized and analytical data areas, the system was unable to accurately classify, deduplicate, and correlate data based on the missing metadata. This lack of data integrity severely impacted the usability of the data at higher levels.

[0005] Ultimately, due to the complexity and non-standardization of the metadata management repository's index structure, the query engine's execution plan generator struggles to efficiently optimize query paths when faced with complex query requests involving multiple coexisting disease patterns. This results in inefficient execution plans, and even with physical optimizations such as hierarchical storage, latency for complex queries cannot be significantly reduced. Conversely, the need for frequent backtracking and correlation of incomplete metadata can further exacerbate retrieval latency, while significantly increasing the complexity of data governance. This makes cross-patient, cross-time-dimension coexistence analysis exceptionally slow and resource-intensive, severely impacting the platform's ability to provide timely and accurate support for clinical decision-making.

[0006] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0007] This invention provides a data processing method and system for a cloud platform for health information on comorbidities in the elderly. It aims to address the problems faced by existing cloud platforms when processing non-standard health data generated by medical devices, including incomplete data storage, low retrieval efficiency, and the inability of the platform's data access layer to accurately identify information due to non-standard metadata introduced by medical device firmware updates, thus affecting subsequent data processing and complex query analysis. The technical solution of this application is as follows: Firstly, this application discloses a data processing method for an elderly comorbidity health information cloud platform, including: Receive data streams from medical devices and perform preliminary parsing of the data streams according to preset rules; In the initial analysis, byte sequences that do not conform to preset rules in the data stream are identified, and the position, length and associated device identification information of the byte sequences are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. The system collects and groups the position, length, and associated device identification information of byte sequences. It then analyzes the grouped information to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, it generates rules for parsing byte sequences. Start an isolated test environment, use rules to parse and test the data stream, and evaluate the effectiveness of the parsing rules; Once the parsing rules pass the validity evaluation, the parsing rules for the data stream are updated without interrupting the data stream transmission.

[0008] Through this technical solution, this application can effectively identify and process non-standard metadata in medical device data streams, and dynamically generate and update parsing rules without interrupting core data transmission. This solves the problems of non-standard data being unidentifiable, data integrity being compromised, and query efficiency being low in existing technologies, and significantly improves the data processing capabilities and data availability of the elderly comorbidity health information cloud platform.

[0009] Furthermore, based on the above method, the position, length, and associated device identification information of byte sequences are collected and grouped. The grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural pattern reaches a preset standard, rules for parsing the byte sequences are generated, including: Collect byte sequences that do not conform to preset rules from data packets from medical devices, and extract device operating status parameters associated with the byte sequences from the data packets; Based on the device operating status parameters, a semantic interpretation branch is selected, which corresponds to a possible semantic and decoding logic of non-standard metadata. Based on the semantic interpretation branch, specific decoding algorithms and conversion logic are applied to convert the byte sequence into structured numerical values ​​with practical meaning; Based on structured numerical values, the grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, rules for parsing byte sequences are generated.

[0010] Through this technical solution, this application can intelligently select semantic interpretation branches based on the device operating status parameters, and apply specific decoding algorithms to convert non-standard byte sequences into structured values, thereby more accurately discovering the structural patterns of non-standard data and improving the accuracy and efficiency of rule generation.

[0011] More specifically, in some implementation schemes, the grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural pattern reaches a preset standard, rules for parsing the byte sequence are generated, including: Upon receiving data packets from medical devices, continuously parse the known device operating status parameters and physiological indicators from the data packets; Based on the preset event triggering rule set, determine whether the known equipment operating status parameters or physiological indicators meet the triggering event conditions; When the conditions for triggering an event are met, an event notification is generated, which includes the event type and the device identifier in the current data packet; Based on the event type and device identifier in the event notification, the corresponding set of candidate search windows is searched from the preset search window mapping table; Within the candidate search window set, the current data packet is scanned bit-level and compared with the expected bit structure diagram corresponding to the device identifier; Bit sequences that do not conform to the expected bit structure diagram are marked as potential non-standard segments, and the starting bit offset, bit length, and relative position to the nearest known field are recorded. Potential non-standard fragments are initially semantically decomposed and associated with the triggering event information and device source information in the event notification to form a non-standard metadata record with context information; Non-standard metadata records are stored in a sparse metadata accumulation area, and sparse non-standard metadata samples from different triggering events are continuously accumulated. The accumulated samples are evaluated for validity and confidence. When the accumulated number of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

[0012] Through this technical solution, this application can identify potential non-standard segments in a refined manner by bit-level scanning and comparison with the expected bit structure diagram, and form non-standard metadata records by combining contextual information. In turn, by accumulating and evaluating samples, it provides a more reliable data foundation for rule generation, which significantly improves the accuracy of non-standard data identification and the effectiveness of rule generation.

[0013] Based on this, the accumulated samples are evaluated for validity and confidence. When the accumulated number of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation, including: Perform consistency checks on the accumulated samples to identify conflicting patterns within the samples; Adjust the weighting coefficients of the accumulated samples based on the characteristics of the conflicting patterns; Aggregate patterns with similar features but minor differences to generate a comprehensive pattern; According to the comprehensive model, the accumulated samples are evaluated for validity and confidence. When the accumulated number of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

[0014] Through this technical solution, this application can effectively handle inconsistencies and differences in samples by using consistency checks, conflict pattern adjustment weighting coefficients, and similar pattern aggregation, thereby generating more representative comprehensive patterns and improving the accuracy of sample evaluation and the reliability of rule generation.

[0015] Preferably, a consistency check is performed on the accumulated samples to identify conflicting patterns in the samples, including: Extract associated equipment operating status parameters and patient physiological condition information from each accumulated sample; The device's operating status parameters and the patient's physiological condition information are matched with preset context rules to identify the specific context in which the accumulated samples are located. Based on each identified specific context, activate a consistency check criterion for that specific context. This consistency check criterion defines the expected structure, value range, and logical association with other known data of a non-standard metadata schema within that specific context. Based on the consistency check criteria, non-standard metadata patterns in the accumulated samples are compared. Mark sample patterns that meet the expected criteria in a specific context as valid context-dependent variants; Sample patterns that deviate significantly from or logically contradict the expected standards in a specific context are identified as conflict patterns.

[0016] This technical solution combines device operating status parameters and patient physiological condition information to identify specific contextual environments and activate corresponding consistency check standards. This enables more accurate identification of conflict patterns that deviate significantly from or contradict the expected standards, thereby improving the accuracy and reliability of non-standard metadata pattern recognition.

[0017] Building upon the above, after identifying sample patterns that significantly deviate from or logically contradict the expected standards in a specific context as conflict patterns, the following further steps are taken: After identifying the conflict pattern, record the timestamp of the conflict pattern occurrence and associate the timestamp with the device identifier and event type; Analyze the frequency and duration of conflict patterns; When a conflict pattern occurs consecutively a number of times within a preset time window reaches a preset threshold, or when its duration exceeds a preset duration, the conflict pattern is marked as a persistent conflict pattern. Samples marked as persistent conflict patterns are isolated from the accumulated samples, triggering further exception handling procedures.

[0018] Through this technical solution, after identifying conflict patterns, this application further analyzes their frequency and duration, which can effectively identify and isolate persistent conflict patterns, thereby avoiding interference from abnormal data in the rule generation process and improving the robustness and accuracy of data processing.

[0019] Furthermore, the frequency and duration of conflict patterns are analyzed, including: Obtain device type information, patient group information, and patient disease status information associated with the conflict pattern; Based on the equipment type information, extract the data acquisition rate, data accuracy, and data transmission stability parameters of the equipment type information from the preset set of equipment characteristic parameters; Based on patient group information and patient disease status information, the normal fluctuation range, abnormal threshold, and typical duration of physiological events of the patient group under disease status information are extracted from the preset physiological baseline parameter set. Based on the device characteristic parameters, physiological baseline parameters, and the type of conflict mode, determine the counting window size and counting threshold for assessing the frequency of conflict mode occurrence, as well as the duration threshold for assessing the duration of conflict mode. Within a defined counting window, count the number of occurrences of the conflict pattern and determine whether the number of occurrences has reached the counting threshold. Measure the duration of the conflict pattern and determine whether the duration exceeds the duration threshold; Based on the results of the number of occurrences and the duration, a comprehensive assessment is made of the frequency and duration of the conflict patterns.

[0020] This technical solution allows for the dynamic determination of evaluation parameters by comprehensively considering equipment characteristic parameters, physiological baseline parameters, and conflict mode types. It also enables a comprehensive evaluation of the frequency and duration of occurrence, thus providing a more comprehensive and accurate analysis of the characteristics of conflict modes and offering a more refined basis for subsequent anomaly handling.

[0021] Based on the above, the accumulated samples are evaluated for validity and confidence. When the accumulated number of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation, including: Obtain the device source identifier and collection timestamp for each accumulated sample; Based on the equipment source identifier and data collection timestamp, retrieve the equipment's routine operating parameters and known physiological index ranges within similar time periods from historical data; For each accumulated sample, a preliminary anomaly assessment is performed, comparing the normal operating parameters and physiological indicator ranges to identify samples that deviate from the normal range. The samples initially identified as abnormal were classified, and based on the degree of deviation and the source of the equipment, the abnormal samples were categorized as potential equipment failure abnormalities or potential rare physiological event abnormalities. Based on potential equipment failure anomaly samples, a dedicated equipment failure mode recognition mechanism is activated, which includes structured checks and numerical logic verification of specific equipment failure characteristics.

[0022] Through this technical solution, this application can identify anomalies by comparing with historical data and activating a specialized equipment failure mode recognition mechanism, thereby effectively identifying potential equipment failure anomalies or rare physiological event anomalies, thus improving the accuracy and relevance of sample evaluation.

[0023] As a technological improvement, the method also includes: Anomalies identified through the equipment failure mode recognition mechanism are assigned evaluation weights. For anomalous samples identified through the equipment failure mode identification mechanism, evaluation weights are assigned, including: Obtain the type information of the abnormal sample; Obtain patient disease combination information associated with abnormal samples; Obtain patient treatment plan information associated with abnormal samples; Based on the type information of the abnormal sample, the corresponding initial weight value is found from the preset weight configuration table; Based on the patient's disease combination information and treatment plan information, extract the risk impact factors corresponding to the disease combination information and treatment plan information from the preset risk impact factor table; The initial weight values ​​are adjusted with the risk impact factors to obtain the dynamically adjusted evaluation weights; The dynamically adjusted evaluation weights are assigned to the outlier samples.

[0024] By combining the type of abnormal sample, patient disease combination and treatment plan information, this application can dynamically adjust the evaluation weights to more comprehensively and accurately assess the risk and importance of abnormal samples, providing a more refined basis for subsequent decision-making.

[0025] Secondly, this application also discloses a cloud platform data processing system for health information on comorbidities in the elderly, comprising: The input end is used to receive data streams from medical devices and perform preliminary parsing of the data streams according to preset rules. In the preliminary parsing, byte sequences in the data stream that do not conform to the preset rules are identified, and the position, length and associated device identification information of the byte sequences are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. The generation end is used to collect and group the position, length, and associated device identification information of byte sequences. It analyzes the grouped information to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, it generates rules for parsing byte sequences. The update endpoint is used to start an isolated test environment, use rules to parse and test the data stream, and evaluate the effectiveness of the parsing rules. Once the parsing rules pass the effectiveness evaluation, the parsing rules of the data stream are updated without interrupting the data stream transmission.

[0026] Through this technical solution, the system provided in this application can automatically identify, generate rules and dynamically update non-standard metadata in medical device data streams, ensuring continuous transmission of core data, effectively solving the problem of non-standard data processing, and improving data integrity and platform operating efficiency.

[0027] Beneficial effects This application discloses a data processing method and system for a cloud platform for health information on comorbidities in the elderly. By receiving data streams from medical devices and performing preliminary parsing, the system can identify byte sequences in the data stream that do not conform to preset rules and obtain their location, length, and associated device identification information. After separating these non-standard byte sequences, core measurement data can continue to be transmitted, ensuring the high availability of the platform. This application further collects, groups, and analyzes information on these non-standard byte sequences to discover recurring structural patterns. When the reliability reaches a preset standard, rules for parsing these byte sequences are automatically generated. Subsequently, the system starts an isolated test environment to evaluate the effectiveness of the newly generated rules. After the rules pass the evaluation, the data stream parsing rules are dynamically updated without interrupting data stream transmission.

[0028] Through the above technical solution, this application effectively solves the problems in the prior art where the introduction of non-standard metadata in medical device firmware updates leads to inaccurate information identification by the data access layer, incomplete data storage, and low retrieval efficiency. Specifically, this application avoids incomplete data context caused by missing metadata by dynamically identifying and parsing non-standard metadata, thereby improving the usability of data in subsequent cleaning, transformation, and analysis stages. Simultaneously, rule updates are performed without interrupting core data transmission, ensuring the continuity and stability of platform services. Furthermore, by automatically generating and evaluating parsing rules, the cost and complexity of manual intervention are reduced, significantly improving the data processing capabilities and efficiency of the elderly comorbidity health information cloud platform, and providing more timely and accurate support for clinical decision-making. Attached Figure Description

[0029] Figure 1 This is a flowchart of a method for data processing in a cloud platform for health information on comorbidities in the elderly, provided by an embodiment of the present invention. Figure 2 This is a flowchart of a method for generating rules for parsing byte sequences provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of a cloud platform data processing system for health information on comorbidities in the elderly, provided in an embodiment of the present invention. Detailed Implementation

[0030] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0031] Reference Figure 1 , Figure 1This is a flowchart of a method for data processing in a cloud platform for health information on comorbidities in the elderly, provided by an embodiment of the present invention, including: S11, Receive data stream from medical device and perform preliminary parsing of the data stream according to preset rules; S12, in the preliminary analysis, identify byte sequences in the data stream that do not conform to the preset rules, obtain the position, length and associated device identification information of the byte sequences, and after the byte sequences are separated, allow the core measurement data in the data stream to continue to be transmitted; S13, collect and group the position, length and associated device identification information of the byte sequence, analyze the grouped information to discover the recurring structural pattern, and when the reliability of the structural pattern reaches a preset standard, generate rules for parsing the byte sequence; S14, Start an isolated test environment, use the rules to perform parsing tests on the data stream, and evaluate the effectiveness of the parsing rules; S15, after the parsing rule passes the validity evaluation, the parsing rule of the data stream is updated without interrupting the data stream transmission.

[0032] The method proposed in this application is primarily applied to an elderly comorbidity health information cloud platform, aiming to process health data streams from various medical devices. Here, "data stream" refers to a continuous sequence of real-time or near-real-time data generated and transmitted by medical devices, such as electrocardiogram (ECG) data, blood pressure monitoring data, and blood glucose measurement data. "Preset rules" refer to a set of data parsing standards predefined by the platform, used to identify and structure standard-format medical data. "Byte sequence" refers to consecutive byte fragments in the data stream that do not conform to the preset rules; these fragments typically contain non-standard metadata. "Device identification information" is used to uniquely identify the medical device generating the data stream, such as the device's serial number and MAC address. "Core measurement data" refers to the patient's physiological indicators, such as heart rate, blood pressure, and blood glucose levels; these data form the basis for the platform's health assessment and disease management. "Isolated testing environment" is a sandbox environment independent of the main data processing flow, used to safely test newly generated parsing rules and avoid interfering with existing data streams.

[0033] Specifically, the data processing method of this application includes the following steps: First, the system receives data streams from medical devices and performs preliminary parsing according to preset rules. For example, the data stream might be binary data packets transmitted via TCP / IP from a smart blood pressure monitor, blood glucose meter, or wearable ECG monitoring device. Upon receiving these packets, a protocol parsing module in the data access layer performs preliminary structuring processing on the packets according to preset parsing rules (e.g., based on HL7, DICOM, or a custom JSON / XML schema). These preset rules define the expected position, length, and data type of each field in the data packet.

[0034] Secondly, in the initial parsing, byte sequences in the data stream that do not conform to preset rules are identified, and their positions, lengths, and associated device identification information are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. For example, when the protocol parsing module is processing a data packet, if it finds that the byte sequence in a certain area does not conform to any field pattern defined in the preset rules—for example, a field expected to be an integer contains a string of unrecognizable characters, or there is a segment of undefined bytes between known fields—then the byte sequence will be identified as non-standard data. At this time, the system will record the starting position of the non-standard byte sequence in the original data stream, its length, and the unique identification information of the medical device to which the data packet belongs. To ensure the real-time nature of core health data, these non-standard byte sequences will be logically or physically separated from the main data stream, while the core measurement data (such as heart rate and blood pressure readings) contained in the data stream will continue to be transmitted to the subsequent data cleaning and storage modules according to the established process.

[0035] Next, the system collects and groups the position, length, and associated device identification information of byte sequences. It then analyzes the grouped information to discover recurring structural patterns. When the reliability of these patterns reaches a preset standard, rules for parsing the byte sequences are generated. For example, the system collects and groups non-standard byte sequence information from the same device or batch of devices. Suppose that in a data stream received from a specific model of blood glucose meter, a 4-byte unknown sequence is repeatedly detected at specific offsets in the data packets. Statistical analysis of these recurring sequences reveals, for example, that they always appear after a known field and that their internal bit patterns exhibit a certain regularity (e.g., always starting with "0xAA" or having a value range within a specific interval). When the frequency and consistency of this regularity reach a preset statistical confidence level, the system infers that these 4 bytes may represent specific metadata (e.g., battery level or device calibration status). Based on this inference, the system automatically generates a new parsing rule that defines how to extract these 4 bytes from that position and convert them into meaningful structured data.

[0036] Next, an isolated test environment is launched to parse the data stream using rules and evaluate the effectiveness of the parsing rules. For example, after generating new parsing rules, the system does not immediately apply them to the real-time data stream in the production environment. Instead, a separate sandbox environment is launched, which receives a copy of the real-time data stream or uses historical data samples containing non-standard byte sequences. The newly generated parsing rules are applied in this isolated environment to attempt to parse the non-standard byte sequences. The parsing results are compared with preset validation criteria, such as checking whether the parsed data conforms to the expected data type, value range, and whether there are logical conflicts with other known data. If the accuracy, completeness, and consistency of the parsing results meet the preset validity evaluation criteria, the rule is considered valid.

[0037] Finally, once the parsing rules pass the validity evaluation, the parsing rules for the data stream are updated without interrupting data transmission. For example, once the new parsing rules are verified as valid and reliable in an isolated test environment, the system will deploy them to the protocol parsing module in the production environment via hot update. This update process is seamless and will not cause interruption of ongoing data transmission or service degradation. The updated parsing module will be able to identify and correctly parse non-standard byte sequences that were previously unprocessable, thereby incorporating this metadata into the platform's data processing flow and ensuring data integrity and availability.

[0038] The data processing method for the elderly comorbidity health information cloud platform proposed in this application effectively solves the problems of unrecognizable non-standard metadata, incomplete data storage, and low efficiency of complex queries caused by firmware updates of medical devices in existing technologies. This is achieved by introducing a dynamic non-standard data parsing rule generation and update mechanism. Traditional methods often rely on preset fixed parsing rules, and when encountering non-standard data, they can only mark it as "unknown" or discard it directly, resulting in the loss of key contextual information.

[0039] The innovation of this application lies in its ability to proactively identify byte sequences in the data stream that do not conform to preset rules and separate them from the core measurement data, ensuring that core data transmission is not affected. More importantly, by collecting, grouping, and analyzing the patterns of these non-standard byte sequences, this application can automatically discover recurring structural patterns and generate new parsing rules accordingly. This process not only improves the automation of data parsing but also adapts to the ever-changing firmware and data formats of medical devices.

[0040] Furthermore, this application introduces an isolated testing environment to evaluate the effectiveness of newly generated parsing rules, ensuring the accuracy and reliability of the rules and preventing the introduction of erroneous rules into the production environment. Once a rule passes the evaluation, it can be hot-updated without interrupting data flow, guaranteeing the platform's high availability and the continuity of data processing.

[0041] Compared to existing technologies, the method in this application can transform previously unrecognizable non-standard metadata into meaningful structured information, significantly improving the data integrity of the elderly comorbidity health information cloud platform. By capturing this key device background information and operational status parameters, the platform can perform more refined data cleaning, classification, and correlation, thereby optimizing the data lake's storage structure and significantly improving the efficiency of complex queries. For example, when analyzing multiple coexisting disease patterns, complete metadata can help the query engine more accurately locate and filter relevant data, reducing unnecessary scanning and correlation operations, thus significantly reducing query latency and providing timely and accurate support for clinical decision-making. Therefore, this application demonstrates significant advancements in the intelligence, adaptability, and efficiency of data processing.

[0042] In some embodiments described above, after collecting and grouping the position, length, and associated device identification information of byte sequences, the grouped information is analyzed to discover recurring structural patterns and generate rules for parsing the byte sequences. However, in practical applications, analysis based solely on the position, length, and device identification information of byte sequences may not fully capture the complex semantics and contextual information contained in non-standard data. This may affect the reliability and accuracy of the generated parsing rules when faced with changes in device operating status or data encoding diversity, thereby limiting the effective parsing and utilization of non-standard data.

[0043] In this regard, refer to Figure 2 , Figure 2 This is a flowchart of a method for generating rules for parsing byte sequences according to an embodiment of the present invention. S13 includes: S131, collect byte sequences that do not conform to preset rules from data packets from medical devices, and extract device operating status parameters associated with byte sequences from data packets; S132, Select a semantic interpretation branch based on the device operating status parameters. The semantic interpretation branch corresponds to a possible semantic and decoding logic of non-standard metadata. S133, based on the semantic interpretation branch, applies a specific decoding algorithm and conversion logic to convert the byte sequence into a structured numerical value with practical meaning; S134: Based on the structured values, analyze the grouped information to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, generate rules for parsing byte sequences.

[0044] Specifically, the device operating status parameters can be understood as parameters reflecting the current working mode, configuration settings, sensor status, battery level, error codes, and other information of the medical device. For example, a blood glucose meter's output of non-standard byte sequences may have different meanings and structures in measurement mode, calibration mode, or standby mode. The purpose of extracting these operating status parameters is to provide important contextual information for subsequent semantic interpretation. The semantic interpretation branches can be understood as a predefined set of one or more parsing paths, each path corresponding to a possible meaning and decoding method of non-standard metadata in a specific device operating state. For example, when the device operating status parameters indicate that the device is in "error reporting mode," the system will select a semantic interpretation branch specifically for parsing error codes and diagnostic information; when the device is in "advanced measurement mode," it may select another semantic interpretation branch for parsing complex physiological indicators. Each branch contains a specific decoding algorithm and conversion logic.

[0045] This application's solution addresses the problem of insufficient reliability of parsing rules caused by the lack of contextual information when processing non-standard byte sequences in traditional methods by introducing device operating status parameters and semantic interpretation branches. Through this technical solution, the application significantly improves the parsing capability and rule generation accuracy of non-standard byte sequences in medical device data streams. By combining device operating status parameters for semantic interpretation and decoding, the generated parsing rules can not only identify the structural patterns of byte sequences but also understand their underlying meaning, effectively avoiding misjudgments or parsing failures due to a lack of contextual information. This method enables cloud platforms to process heterogeneous data from various medical devices more comprehensively and accurately, providing a more reliable data foundation for the health management of elderly patients with comorbidities, thereby improving the intelligence and robustness of data processing.

[0046] This application further proposes analyzing the grouped information to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, rules for parsing the byte sequence are generated, including: Upon receiving a data packet from the medical device, continuously parse the known device operating status parameters and physiological indicators from the data packet; Based on the preset event triggering rule set, determine whether the known equipment operating status parameters or physiological indicators meet the triggering event conditions; When the triggering event conditions are met, an event notification is generated, which includes the event type and the device identifier in the current data packet; Based on the event type and device identifier in the event notification, the corresponding set of candidate search windows is searched from a preset search window mapping table; Within the candidate search window set, the current data packet is scanned bit-level and compared with the expected bit structure diagram corresponding to the device identifier; Bit sequences that do not conform to the expected bit structure diagram are marked as potential non-standard segments, and the starting bit offset, bit length, and relative position to the nearest known field are recorded. The potential non-standard fragments are initially semantically decomposed and associated with the triggering event information and device source information in the event notification to form a non-standard metadata record with context information; The non-standard metadata records are stored in the sparse metadata accumulation area, and sparse non-standard metadata samples from different triggering events are continuously accumulated. The accumulated samples are evaluated for validity and confidence. When the accumulated amount of the samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

[0047] Specifically, upon receiving data packets from medical devices, the system is configured to continuously parse known device operating status parameters and physiological indicators from the packets. This step aims to establish a baseline so that abnormal or non-standard data that does not conform to the normal data pattern can be subsequently identified. Device operating status parameters can be understood as the internal state information of the medical device at a specific point in time, such as battery level, calibration status, and operating mode; physiological indicators refer to the patient's vital signs data, such as heart rate, blood pressure, and blood glucose levels.

[0048] Furthermore, based on a preset event triggering rule set, it is determined whether known device operating status parameters or physiological indicators meet the triggering event conditions. This event triggering rule set defines a series of conditions that may indicate the occurrence of non-standard data, such as abnormal device status, physiological indicators exceeding the normal range, or the occurrence of specific operational events. When a triggering event condition is detected, an event notification is generated. This event notification includes the event type and the device identifier in the current data packet. The event notification serves as a signal to instruct the system to focus its attention on a specific data area that may contain non-standard metadata.

[0049] Based on this, according to the event type and device identifier in the event notification, the corresponding candidate search window set is retrieved from a pre-defined search window mapping table. The search window mapping table stores the association between different event types and device identifiers and specific areas (i.e., search windows) in the data packet where non-standard data may appear. This effectively narrows down the search scope for non-standard data, improving search efficiency. Within the candidate search window set, the current data packet is scanned bit-level and compared with the expected bit structure diagram corresponding to the device identifier. Bit-level scanning is a refined data analysis method that can directly identify data patterns at the binary level, while the expected bit structure diagram provides a precise structural template for known standard data in that region.

[0050] When a bit sequence that does not conform to the expected bit structure diagram is found, the sequence is marked as a potential non-standard fragment, and its starting bit offset, bit length, and relative position to the nearest known field are recorded. This information is crucial for subsequent non-standard fragment analysis and rule generation. Subsequently, the potential non-standard fragments undergo preliminary semantic decomposition and are associated with triggering event information and device origin information in event notifications, thereby forming non-standard metadata records with contextual information. By associating contextual information, isolated non-standard fragments can be given preliminary business meaning, such as indicating specific device warnings, error codes, or special measurement modes.

[0051] Finally, these non-standard metadata records with contextual information are stored in a sparse metadata accumulation area, continuously accumulating sparse non-standard metadata samples from different triggering events. The sparse metadata accumulation area serves as a temporary storage area to collect a sufficient number of non-standard data samples to overcome the randomness and uncertainty of individual samples. The accumulated samples undergo validity and confidence assessments. When the accumulated sample count reaches a preset threshold and the assessment result reaches a preset confidence level, the sample is marked as usable for rule generation. The validity assessment aims to check the internal consistency of the sample and its logical compatibility with known data, while the confidence assessment measures the representativeness and reliability of the sample, ensuring that the generated parsing rules have high accuracy and stability.

[0052] This application's solution, by introducing an event-triggered mechanism and contextual association, can more accurately locate and identify potential non-standard segments in the data stream. Through the above technical solution, this application can significantly improve the accuracy and robustness of identifying non-standard metadata in the medical device data stream of the elderly comorbidity health information cloud platform. By introducing an event-driven identification mechanism and contextual association, it can more effectively focus on key areas in the data stream that may contain non-standard information, avoiding inefficient scanning of the entire data stream. Bit-level scanning and comparison with the expected structure diagram ensure the accurate location of non-standard segments. Furthermore, by accumulating sparse metadata samples with contextual information and conducting rigorous validity and confidence assessments, the quality and reliability of the samples used to generate parsing rules are greatly enhanced. This allows the generated parsing rules to more accurately and stably handle diverse non-standard data, effectively reducing the false positive and false negative rates, and providing a more reliable data foundation for the health management of elderly patients with comorbidities.

[0053] This application further proposes a more refined evaluation mechanism, which aims to improve the accuracy and reliability of validity and confidence assessment of accumulated samples through steps such as consistency checks, weighting coefficient adjustments, and pattern aggregation.

[0054] Specifically, the aforementioned process involves evaluating the validity and confidence of the accumulated samples. When the accumulated number of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation, including: A consistency check is performed on the accumulated samples to identify conflicting patterns in the samples; The weighting coefficients of the accumulated samples are adjusted according to the characteristics of the conflicting patterns. Aggregate patterns with similar features but minor differences to generate a comprehensive pattern; According to the comprehensive mode, the accumulated samples are evaluated for validity and confidence. When the accumulated amount of the samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

[0055] The consistency check on the accumulated samples refers to cross-referencing and logically verifying the accumulated non-standard metadata samples to identify contradictory or inconsistent patterns. For example, if different samples from the same device type operating in similar conditions provide contradictory semantic interpretations of the same byte sequence, these sample patterns will be identified as conflicting patterns. The purpose is to discover potential errors, noise, or ambiguity in the data.

[0056] Furthermore, adjusting the weighting coefficients of the accumulated samples based on the characteristics of the conflicting patterns means that after identifying a conflicting pattern, the system dynamically adjusts the weights of relevant samples in subsequent evaluations based on characteristics such as the severity of the conflict, its frequency of occurrence, or the type of equipment involved. For example, samples associated with high-frequency conflicting patterns may be assigned lower weights, while samples associated with stable patterns may maintain higher weights. The aim is to reduce the negative impact of conflicting samples on the overall evaluation results and improve the robustness of the evaluation.

[0057] Furthermore, aggregating patterns with similar characteristics but minor differences to generate a comprehensive pattern can be understood as using techniques such as clustering analysis or fuzzy matching to merge non-standard metadata patterns that are highly similar in structure, length, or contextual information but have subtle differences, forming a more representative and generalizable comprehensive pattern. For example, if a non-standard fragment appears as "0xAB CDEF" and "0xAB CDFF" in different samples, the system can aggregate it into a comprehensive pattern, such as "0xAB CDXX", where "XX" represents the variable part. The purpose is to eliminate the influence of data noise and random variations, and extract more stable underlying patterns.

[0058] Therefore, based on the comprehensive model, the accumulated samples are evaluated for validity and confidence. When the accumulated amount of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation. This means that after preprocessing steps such as consistency checks, weighted adjustments, and model aggregation, the final validity and confidence evaluations will be based on these more refined and reliable comprehensive models. The evaluation process comprehensively considers the accumulated amount of samples, the stability of the comprehensive model, and the degree of matching with known data or preset rules to determine whether the comprehensive model meets the criteria for generating parsed rules.

[0059] This application's solution proactively identifies and locates conflict patterns in the data by performing consistency checks on accumulated samples, thus providing a clear basis for subsequent processing. By adjusting the weighting coefficients of samples according to the characteristics of conflict patterns, the system can intelligently reduce the impact of unreliable or contradictory data, ensuring the anti-interference capability of the evaluation process. Simultaneously, aggregating patterns with similar characteristics but minor differences allows the system to extract more stable and representative comprehensive patterns from noise and variation, avoiding misjudgments due to subtle differences. Finally, validity and confidence assessments are performed based on these refined comprehensive patterns, enabling the generated analytical rules to more accurately reflect the actual data patterns and effectively addressing the challenges posed by noise, inconsistencies, and subtle differences in the original accumulated samples to the accuracy of rule generation.

[0060] Through the above technical solutions, this application can significantly improve the accuracy and reliability of validity and confidence assessment of non-standard metadata samples. By introducing consistency checks, the system can proactively identify and handle contradictions in the data, avoiding the inclusion of erroneous information in rule generation. Dynamic adjustment of weighting coefficients enables the system to flexibly respond to differences in data quality, reducing the impact of low-quality data and thus enhancing the robustness of the assessment. The pattern aggregation mechanism effectively filters out noise and random variations in the data, enabling the system to identify more stable and generalizable underlying data structures. These improvements collectively ensure that the final generated parsing rules have higher accuracy, stability, and adaptability, providing a more reliable data processing foundation for the elderly comorbidity health information cloud platform.

[0061] In some embodiments of this application described above, when evaluating the validity and confidence of accumulated samples, a consistency check needs to be performed on the accumulated samples to identify conflicting patterns within the samples. Specifically, performing a consistency check on the accumulated samples to identify conflicting patterns within the samples may include the following steps: Extract associated device operating status parameters and patient physiological condition information from each of the accumulated samples; The device operating status parameters and the patient's physiological condition information are matched with preset context rules to identify the specific context environment in which the accumulated samples are located. Based on each identified specific context, activate a consistency check criterion for that specific context, which defines the expected structure, value range, and logical association with other known data of a non-standard metadata schema in that specific context. According to the consistency check criteria, the non-standard metadata patterns in the accumulated samples are compared. Sample patterns that meet the expected criteria in the specific context are marked as valid context-dependent variants; Sample patterns that deviate significantly from or logically contradict the expected standards in the specific context are identified as conflict patterns.

[0062] Specifically, when performing consistency checks on accumulated samples, it is first necessary to extract associated device operating status parameters and patient physiological condition information from each accumulated sample. Device operating status parameters may include the device's power status, sensor calibration status, transmission mode, etc., while patient physiological condition information may include the patient's age, gender, diagnosis, current medications, underlying diseases, etc. The extraction of this information aims to provide foundational data for subsequent contextual recognition.

[0063] Furthermore, the extracted device operating status parameters and patient physiological condition information are matched with preset context rules to identify the specific context in which the accumulated samples exist. The preset context rules can be a series of logical conditions defining different combinations of device states and patient physiological conditions. For example, when a device is "charging" and the patient is "sleep," it may constitute a specific context. Through this matching, the specific situation in which each non-standard metadata sample is generated can be accurately defined.

[0064] Based on this, a consistency check criterion is activated for each identified specific context. This consistency check criterion is predefined and used to regulate the expected structure, value range, and logical relationships with other known data of non-standard metadata schemas within a specific context. For example, in the context of "patient sleep, device charging," the expected value range of a non-standard metadata field might be limited to a narrow range, or its structure should conform to a specific encoding format.

[0065] Subsequently, non-standard metadata patterns in the accumulated samples are compared according to the activated consistency check criteria. This comparison process aims to check whether the structure and values ​​of the samples conform to expectations in that specific context.

[0066] Therefore, sample patterns that conform to the expected criteria in a specific context are marked as valid context-dependent variants. These variants are considered reliable non-standard metadata patterns that can be used for subsequent rule generation.

[0067] Conversely, sample patterns that deviate significantly from or logically contradict expected standards within a specific context are identified as conflicting patterns. Significant deviations may include index values ​​exceeding the expected range, structural inconsistencies, or logical inconsistencies with other known data. These conflicting patterns may indicate data acquisition errors, equipment malfunctions, unusual physiological events, or unknown non-standard data formats, requiring further review and processing.

[0068] This application's solution, by introducing a context-dependent consistency check mechanism, enables a more refined assessment of the reliability of non-standard metadata samples. Through this technical solution, the application significantly improves the accuracy and robustness of non-standard metadata pattern recognition. By associating samples with specific contextual environments and applying context-dependent consistency check criteria, truly meaningful conflict patterns can be identified more effectively, such as anomalous data caused by equipment failure, data transmission errors, or rare physiological events. This helps reduce false positives and false negatives, ensuring higher quality and reliability of the non-standard metadata samples used to generate parsing rules. Furthermore, by distinguishing between valid context-dependent variants and conflict patterns, this application can more flexibly adapt to the diversity and complexity of medical device data formats, providing a more accurate and reliable data processing foundation for elderly comorbidity health information cloud platforms.

[0069] In some of the embodiments described above in this application, although sample patterns that significantly deviate from or logically contradict the expected standards in a specific context—i.e., conflicting patterns—can be identified, if these conflicting patterns persist or occur frequently, they may indicate deeper data collection anomalies, equipment malfunctions, or rare physiological events. Simply identifying these patterns without further differentiation and processing may lead to inaccuracies in subsequent rule generation or lag in the system's response to anomalies, thereby affecting the data reliability of the elderly comorbidity health information cloud platform.

[0070] In response, this application further proposes that, after identifying sample patterns that significantly deviate from or logically contradict the expected standards in a specific context as conflict patterns, it also includes: After identifying the conflict pattern, the timestamp of the occurrence of the conflict pattern is recorded, and the timestamp is associated with the device identifier and the event type; The frequency and duration of the aforementioned conflict patterns were analyzed. When the number of times the conflict pattern occurs consecutively within a preset time window reaches a preset threshold, or when its duration exceeds a preset duration, the conflict pattern is marked as a persistent conflict pattern. The samples marked as persistent conflict patterns are isolated from the accumulated samples, and further anomaly handling procedures are triggered.

[0071] Specifically, after identifying a conflict pattern, the system immediately records the timestamp of its occurrence. This timestamp, along with the associated device identifier and event type, is stored for subsequent tracing and analysis. The device identifier uniquely identifies the medical device from which the data originated, while the event type specifies the particular event category that caused the conflict pattern, such as incorrect data format or values ​​outside the acceptable range. This association provides rich contextual information for subsequent analysis.

[0072] Furthermore, the frequency and duration of conflict patterns are analyzed. Frequency refers to the number of times the same conflict pattern occurs within a specific time window; duration refers to the length of time a conflict pattern lasts from its first occurrence to its disappearance. These analyses aim to assess the severity and stability of conflict patterns.

[0073] When a conflict pattern occurs consecutively a number of times within a preset time window reaches a preset threshold, or its duration exceeds a preset duration, the conflict pattern will be marked as a persistent conflict pattern. The preset time window, preset threshold, and preset duration can be configured according to specific application scenarios, device types, and data sensitivity. For example, for high-frequency data acquisition devices, the preset threshold may be higher, while for key physiological indicators, the preset duration may be shorter.

[0074] Once marked as a persistent conflict pattern, these samples will be isolated from the accumulated samples. This means they will no longer participate in subsequent rule generation, preventing the generation of inaccurate parsing rules based on erroneous or anomalous data. Simultaneously, the system will trigger further anomaly handling procedures, such as sending alerts to the system administrator, initiating device diagnostics, or suspending data reception on the device, to ensure timely intervention for potential problems.

[0075] This application's solution effectively addresses the limitations of basic solutions that merely identify conflict patterns without differentiating and addressing their severity by introducing analysis of the frequency and duration of conflict patterns and isolating and handling persistent conflict patterns based on the analysis results. Through this technical solution, the robustness and accuracy of data processing in the elderly comorbidity health information cloud platform are significantly improved. This solution not only identifies non-standard metadata conflict patterns in the data stream but also differentiates and effectively handles persistent and systemic conflict patterns. This avoids generating erroneous parsing rules due to persistent abnormal data, ensuring the accuracy and reliability of data parsing rules. Furthermore, timely isolation and anomaly handling of persistent conflict patterns enable early warning and intervention for potential equipment failures or rare physiological events, thereby reducing the risk of data error propagation, improving the data quality and operational stability of the entire health information cloud platform, and providing more reliable data support for the health management of elderly patients with comorbidities.

[0076] Specifically, the steps described above for analyzing the frequency and duration of conflict patterns can be understood as conducting in-depth quantitative assessments of the identified conflict patterns through multi-dimensional data correlation and parameter extraction, in order to more accurately determine their nature and potential impact.

[0077] The analysis process includes the following steps: Obtain device type information, patient group information, and patient disease status information associated with the conflict mode; Based on the device type information, extract the data acquisition rate, data accuracy, and data transmission stability parameters of the device type information from a preset set of device characteristic parameters; Based on the patient group information and the patient's disease status information, extract the normal fluctuation range, abnormal threshold, and typical duration of physiological events of the patient group under the disease status information from the preset physiological baseline parameter set; Based on the device characteristic parameters, the physiological baseline parameters, and the type of conflict mode, determine the counting window size and counting threshold for assessing the frequency of occurrence of the conflict mode, and the duration threshold for assessing the duration of the conflict mode; Within the defined counting window, the number of occurrences of the conflict mode is counted, and it is determined whether the number of occurrences reaches the counting threshold. Measure the duration of the conflict mode and determine whether the duration exceeds the duration threshold; Based on the results of the determination of the number of occurrences and the determination of the duration, a comprehensive evaluation is performed on the occurrence frequency and duration of the conflict pattern.

[0078] The acquisition of information related to conflict patterns, including device type, patient population, and patient disease status, aims to provide necessary context for subsequent analysis. Device type information may include the specific model and manufacturer of the medical device; patient population information may refer to demographic characteristics such as age group, gender, and region; and patient disease status information may cover their diagnosis, disease stage, and complications. This information forms the basis for refined analysis.

[0079] Furthermore, based on the device type information, data acquisition rate, data accuracy, and data transmission stability parameters are extracted from a preset set of device characteristic parameters. Data acquisition rate refers to the number of data points the device can generate per second or per minute; data accuracy refers to the accuracy of the device's measurement results; and data transmission stability parameters reflect the possibility of data loss or corruption during transmission. These parameters help to understand the device's own data generation and transmission characteristics, thus enabling the differentiation between data fluctuations caused by inherent device characteristics and genuine anomalies when assessing conflict patterns.

[0080] Simultaneously, based on patient group information and patient disease status information, the normal fluctuation range, abnormal threshold, and typical duration of physiological events are extracted from a pre-defined physiological baseline parameter set. The physiological baseline parameter set is a database containing a wealth of medical knowledge and clinical experience. It defines the normal range of variation for various physiological indicators (such as heart rate, blood pressure, and blood glucose) under specific patient groups and disease states, the criteria for judging abnormalities exceeding these ranges, and the typical duration of specific physiological events (such as arrhythmias and hypoglycemic episodes). By incorporating this physiological background knowledge, it is possible to more accurately determine whether conflict patterns are related to the patient's actual physiological condition and the degree of abnormality.

[0081] Based on this, and according to the extracted device characteristic parameters, physiological baseline parameters, and conflict mode types, the counting window size and counting threshold for assessing the frequency of conflict mode occurrences, and the duration threshold for assessing the duration of conflict modes are determined. The counting window size defines the time period within which conflict modes are counted, while the counting threshold sets the number of occurrences within that window to be considered high-frequency. The duration threshold is used to determine whether the duration of a single conflict mode is excessively long. These thresholds are set dynamically and intelligently, adjusting according to the specific characteristics of the device, patient, and conflict mode to adapt to different clinical scenarios and data characteristics.

[0082] Within a defined counting window, the system automatically counts the number of times a conflict mode occurs and determines whether it has reached a preset counting threshold. Simultaneously, it measures the duration of each conflict mode and determines whether it exceeds a preset duration threshold.

[0083] Finally, based on the results of judging the frequency and duration of occurrence, a comprehensive assessment is made of the frequency and duration of conflict patterns. This comprehensive assessment provides a more holistic and accurate perspective on understanding the severity and potential impact of conflict patterns, avoiding misjudgments that may arise from a single indicator.

[0084] This application's solution introduces multi-dimensional contextual information, including device characteristics, patient physiological baselines, and the type of conflict pattern itself. This transforms the analysis of the frequency and duration of conflict patterns from simple counting and timing into an intelligent assessment based on professional knowledge and clinical experience. This technical solution significantly improves the accuracy of identifying and assessing conflict patterns in non-standard metadata. By comprehensively considering device characteristics, patient physiological baselines, and conflict pattern types, it avoids misinterpreting normal device fluctuations or physiological changes as conflict patterns, thereby reducing false alarms. Simultaneously, genuine, clinically significant conflict patterns can be identified and labeled more promptly and accurately, providing a reliable basis for subsequent anomaly handling and rule generation. This refined analytical approach enables the elderly comorbidity health information cloud platform to handle complex and variable medical data more intelligently, improving the robustness and reliability of data processing.

[0085] In some of the embodiments described above in this application, the validity and confidence assessment of accumulated non-standard metadata samples mainly focuses on consistency checks and pattern aggregation within the samples. However, in practical applications, the occurrence of non-standard data fragments may not only be due to unknown data structures, but may also stem from potential malfunctions of medical devices or rare physiological events experienced by patients. Failure to effectively distinguish these potential root causes may result in insufficient robustness in the generation of parsing rules, and may even lead to the misclassification of anomalous data as parsable non-standard patterns, thereby affecting the accuracy and security of data processing. To address this, this application further proposes a more refined evaluation method, aiming to more accurately identify and classify the potential sources of non-standard samples by combining historical data and anomaly detection mechanisms, thereby improving the reliability of rule generation.

[0086] The above-mentioned validity and confidence assessment of accumulated samples, when the accumulated amount of the samples reaches a preset threshold and the assessment result reaches a preset confidence level, the samples are marked as usable for rule generation, including: Obtain the device source identifier and collection timestamp for each of the accumulated samples; Based on the device source identifier and the collection timestamp, retrieve the device's routine operating parameters and known physiological index ranges within a similar time period from historical data; For each of the accumulated samples, a preliminary anomaly assessment is performed, and the normal operating parameters and the range of physiological indicators are compared to identify samples that deviate from the normal range. The samples initially identified as abnormal are classified, and based on the degree of deviation and the source of the equipment, the abnormal samples are classified as potential equipment failure abnormalities or potential rare physiological event abnormalities. Based on the potential equipment failure anomaly samples, a dedicated equipment failure mode recognition mechanism is activated, which includes structured checks and numerical logic verification of specific equipment failure characteristics.

[0087] Specifically, obtaining the device source identifier and collection timestamp for each accumulated sample aims to provide necessary contextual information for subsequent data retrieval and anomaly detection. The device source identifier is used to uniquely identify the medical device that generated the non-standard sample, while the collection timestamp provides the time point when the data was generated. The combination of the two allows for precise tracing of the data source and the timing of the event.

[0088] Furthermore, based on the acquired device source identifier and data collection timestamp, the system retrieves the device's routine operating parameters and known physiological index ranges for a similar time period from historical data storage. Routine operating parameters may include the device's battery level, sensor calibration status, and internal error logs, while the physiological index range refers to the fluctuation range of physiological parameters typically monitored by the device (such as heart rate, blood pressure, and blood glucose) under normal conditions. This historical data serves as a baseline, providing a reference for determining the anomaly of the current sample.

[0089] Based on this, an initial anomaly assessment is performed on each accumulated sample. This assessment process compares the contextual information of the current sample (e.g., known data associated with non-standard byte sequences) with retrieved ranges of normal operating parameters and physiological indicators. If certain characteristics of the current sample deviate significantly from historical normal ranges, it is identified as a preliminary anomaly sample. For example, if a non-standard byte sequence appears simultaneously with an abnormal decrease in device battery power, it may indicate a problem with the device.

[0090] Subsequently, samples initially identified as abnormal are categorized. This categorization process, based on factors such as the degree of deviation and the origin of the equipment, classifies abnormal samples into potential equipment malfunctions or potential rare physiological events. For example, if the deviation is significant and highly correlated with the equipment's own operating parameters, it is more likely to be classified as an equipment malfunction; if the deviation is significant but associated with a patient's specific physiological condition or rare disease, it may be classified as a rare physiological event. This categorization helps in adopting different treatment strategies for different types of abnormalities.

[0091] Finally, based on the anomaly samples categorized as potential equipment failures, the system activates a dedicated equipment failure mode recognition mechanism. This mechanism is designed to perform structured checks and numerical logic verification on specific equipment failure characteristics. For example, it may include identification logic for predefined fault code patterns, abnormal sensor reading combinations, or specific hardware error signals. By activating this mechanism, samples suspected of equipment failures can be analyzed in depth to confirm whether there are actual equipment problems, thereby avoiding misclassification of data anomalies caused by equipment failures as non-standard data requiring the generation of parsing rules.

[0092] The proposed solution effectively addresses the limitations of existing evaluation methods in distinguishing non-standard data sources by introducing historical data comparison and anomaly classification mechanisms.

[0093] In some of the embodiments described above in this application, although potential equipment malfunctions or rare physiological events can be identified and a specialized equipment failure pattern recognition mechanism can be activated for inspection, in practical applications, different types of abnormal samples, as well as their associated patient disease combinations and treatment plans, may have drastically different risk levels and urgency. If all abnormal samples are processed without differentiation, it may lead to improper resource allocation and an inability to respond promptly to high-risk events, thereby affecting the efficiency and safety of health management for elderly patients with comorbidities.

[0094] In response, this application further proposes to assign evaluation weights to abnormal samples that have undergone the aforementioned equipment failure mode identification mechanism.

[0095] Specifically, assigning evaluation weights to abnormal samples identified by the equipment failure mode recognition mechanism includes: Obtain the type information of the abnormal sample; Obtain patient disease combination information associated with the abnormal sample; Obtain patient treatment plan information associated with the abnormal sample; Based on the type information of the abnormal sample, the corresponding initial weight value is found from the preset weight configuration table; Based on the patient's disease combination information and the patient's treatment plan information, extract the risk impact factors corresponding to the disease combination information and the treatment plan information from a preset risk impact factor table; The initial weight values ​​are adjusted with the risk impact factors to obtain the dynamically adjusted evaluation weights; The dynamically adjusted evaluation weights are assigned to the abnormal samples.

[0096] The abnormal sample type information can be understood as the classification of abnormal events, such as device sensor malfunction, data transmission error, abnormal fluctuations in specific physiological indicators, and drug side effects. This type information can be predefined and categorized according to the characteristics of the medical device and common abnormal patterns. The patient disease combination information refers to the collection of multiple diseases suffered by the patient, such as hypertension, diabetes, and coronary heart disease. This combination of diseases may significantly increase the risk of certain abnormal events. The patient treatment plan information refers to the drug treatment, physical therapy, or other interventions the patient is currently receiving. Different treatment plans may affect the patient's physiological state, thereby altering the risk assessment of abnormal events. The weight configuration table is a pre-set database or rule set that stores the initial weight values ​​corresponding to different abnormal sample types.

[0097] This application's solution effectively addresses the problem of lacking refined risk level differentiation after identifying abnormal samples by introducing a mechanism that assigns evaluation weights to them. Through this technical solution, this application enables more refined and personalized risk assessment and priority management of abnormal data in the elderly comorbidity health information cloud platform. Compared to basic solutions that only identify and classify abnormalities, this application, by introducing dynamic evaluation weights, allows the system to quantify and differentiate the risk level of abnormal samples based on various factors such as the type of abnormal event, the patient's disease combination, and treatment plan. This significantly improves the intelligence level of abnormal event handling, avoids resource waste, and ensures that medical resources are prioritized for patients in need and for the most urgent abnormal situations. Especially for elderly patients with comorbidities, whose health conditions are complex and variable, this solution can more accurately identify and respond to potential emergencies, thereby effectively improving the quality and safety of health management for elderly patients with comorbidities.

[0098] refer to Figure 3 , Figure 3 This is a schematic diagram of the structure of a cloud platform data processing system for health information on comorbidities in the elderly, provided in an embodiment of the present invention, including: The input end is used to receive data streams from medical devices and perform preliminary parsing of the data streams according to preset rules. In the preliminary parsing, byte sequences in the data stream that do not conform to the preset rules are identified, and the position, length and associated device identification information of the byte sequences are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. The generation end is used to collect and group the position, length, and associated device identification information of the byte sequence, analyze the grouped information to discover the recurring structural patterns, and generate rules for parsing the byte sequence when the reliability of the structural patterns reaches a preset standard. The update endpoint is used to start an isolated test environment, use the rules to perform parsing tests on the data stream, and evaluate the effectiveness of the parsing rules; when the parsing rules pass the effectiveness evaluation, the parsing rules of the data stream are updated without interrupting the data stream transmission.

[0099] This system aims to address the problems of incomplete data storage and low retrieval efficiency faced by traditional cloud platforms when processing non-standard health data generated by medical devices. Through a systematic component design, this application can effectively identify and process byte sequences in the data stream that do not conform to preset rules, ensuring the continuous transmission of core measurement data, while dynamically generating and updating parsing rules to adapt to changes in non-standard data. Therefore, this application can improve the data integrity and availability of the elderly comorbidity health information cloud platform and optimize the data processing workflow.

[0100] The core of the elderly comorbidity health information cloud platform data processing system proposed in this application lies in its modular design to achieve intelligent processing of non-standard medical data streams.

[0101] Specifically, the system includes an input terminal. The above embodiments have already described receiving a data stream from a medical device, performing preliminary parsing of the data stream according to preset rules, identifying byte sequences in the data stream that do not conform to the preset rules during preliminary parsing, obtaining the position, length, and associated device identification information of the byte sequences, and ensuring the continued transmission of core measurement data in the data stream after the byte sequences are separated; these details will not be repeated here. It is important to emphasize that the input terminal can be configured to include a data receiving module and a preliminary parsing module. The data receiving module is responsible for establishing a communication connection with the medical device, for example, receiving TCP / IP packets or Bluetooth data streams through a network interface card or wireless communication module. The preliminary parsing module can be a programmable protocol parser, internally storing a preset set of parsing rules for preliminary structured processing of the received data stream. When the preliminary parsing module identifies byte sequences that do not conform to the preset rules, it can mark and cache them, while ensuring that the core measurement data in the data stream is extracted and forwarded to subsequent processing stages.

[0102] Furthermore, the system includes a generation end. The above embodiments have already described the collection and grouping of byte sequences, their lengths, and associated device identification information. The grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural pattern reaches a preset standard, rules for parsing the byte sequences are generated, which will not be elaborated further here. It is important to emphasize that the generation end can be configured to include a data collection and grouping module and a rule generation module. The data collection and grouping module is responsible for receiving and accumulating relevant information of non-standard byte sequences from the input end and grouping them according to dimensions such as device identification and timestamps. The rule generation module can be an engine based on machine learning or pattern recognition algorithms. By performing statistical analysis and pattern mining on the grouped non-standard byte sequence data, it automatically identifies recurring structural patterns. For example, this module can use techniques such as sequence alignment algorithms, frequency analysis, or contextual correlation analysis. When the identified regularity reaches a preset confidence threshold, new data parsing rules are automatically constructed.

[0103] In addition, the system also includes an update endpoint. The above embodiments have already described starting an isolated test environment, using rules to parse and test the data stream, evaluating the effectiveness of the parsing rules, and updating the parsing rules of the data stream without interrupting data stream transmission after the parsing rules pass the effectiveness evaluation; these details will not be repeated here. It is important to emphasize that the update endpoint can be configured to include a test environment management module and a rule deployment module. The test environment management module is responsible for dynamically creating and managing an isolated sandbox environment that can simulate the data stream of the production environment and apply newly generated parsing rules for testing within this environment. The rule deployment module is responsible for integrating the new rules into the protocol parser of the production environment through hot deployment or seamless switching after the parsing rules pass the effectiveness evaluation, thereby achieving dynamic updates of the parsing rules without interrupting the running data processing service.

[0104] The data processing system for the elderly comorbidity health information cloud platform proposed in this application demonstrates significant progress in processing non-standard medical data compared to existing technologies. Traditional systems often rely on static, predefined parsing rules. When medical device firmware is updated or new non-standard metadata is introduced, these systems cannot effectively identify and process it, resulting in compromised data integrity and consequently affecting the efficiency of subsequent data analysis and querying.

[0105] The system described in this application achieves adaptive processing of non-standard data streams through the collaborative work of its input, generation, and update ends. The input end intelligently identifies and separates non-standard byte sequences, ensuring the continuous transmission of core physiological data and avoiding interruptions to the overall data stream due to local data anomalies. The generation end, through continuous learning and pattern recognition of these non-standard byte sequences, dynamically generates new parsing rules, greatly improving the system's adaptability to unknown or changing data formats. Furthermore, the isolation testing mechanism introduced in the update end ensures the accuracy and reliability of the new rules, and the uninterrupted update mechanism guarantees the platform's high availability and the continuity of data processing.

[0106] Through the aforementioned system design, this application can transform previously unrecognizable non-standard metadata into meaningful structured information, thereby significantly improving the data integrity of the elderly comorbidity health information cloud platform. This enables the platform to perform more refined data cleaning, classification, and association, optimize data storage structure, and greatly improve the efficiency of complex queries. For example, when analyzing multiple coexisting disease patterns, it can provide more timely and accurate data support, thus providing a more reliable basis for clinical decision-making.

[0107] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A data processing method for an elderly comorbidity health information cloud platform, characterized in that, include: Receive data streams from medical devices and perform preliminary parsing of the data streams according to preset rules; In the initial analysis, byte sequences in the data stream that do not conform to the preset rules are identified, and the position, length and associated device identification information of the byte sequences are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. The location, length, and associated device identification information of the byte sequence are collected and grouped. The grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, rules for parsing the byte sequence are generated. Start an isolated test environment, use the rules to parse the data stream, and evaluate the effectiveness of the parsing rules; Once the parsing rule passes the validity evaluation, the parsing rule for the data stream is updated without interrupting the data stream transmission.

2. The data processing method for an elderly comorbidity health information cloud platform according to claim 1, characterized in that, The process involves collecting and grouping the position, length, and associated device identification information of the byte sequences, analyzing the grouped information to discover recurring structural patterns, and generating rules for parsing the byte sequences when the reliability of the structural patterns reaches a preset standard. These rules include: Collect byte sequences from data packets from the medical device that do not conform to the preset rules, and extract device operating status parameters associated with the byte sequences from the data packets; Based on the device operating status parameters, a semantic interpretation branch is selected, and the semantic interpretation branch corresponds to a possible semantic and decoding logic of non-standard metadata; Based on the semantic interpretation branch, a specific decoding algorithm and conversion logic are applied to convert the byte sequence into a structured numerical value with practical meaning; Based on the structured numerical values, the grouped information is analyzed to discover recurring structural patterns. When the reliability of the structural patterns reaches a preset standard, rules for parsing the byte sequence are generated.

3. The data processing method for an elderly comorbidity health information cloud platform according to claim 2, characterized in that, The analysis of the grouped information to discover recurring structural patterns, and the generation of rules for parsing the byte sequence when the reliability of the structural patterns reaches a preset standard, includes: Upon receiving a data packet from the medical device, continuously parse the known device operating status parameters and physiological indicators from the data packet; Based on the preset event triggering rule set, determine whether the known equipment operating status parameters or physiological indicators meet the triggering event conditions; When the triggering event conditions are met, an event notification is generated, which includes the event type and the device identifier in the current data packet; Based on the event type and device identifier in the event notification, the corresponding set of candidate search windows is searched from a preset search window mapping table; Within the candidate search window set, the current data packet is scanned bit-level and compared with the expected bit structure diagram corresponding to the device identifier; Bit sequences that do not conform to the expected bit structure diagram are marked as potential non-standard segments, and the starting bit offset, bit length, and relative position to the nearest known field are recorded. The potential non-standard fragments are initially semantically decomposed and associated with the triggering event information and device source information in the event notification to form a non-standard metadata record with context information; The non-standard metadata records are stored in the sparse metadata accumulation area, and sparse non-standard metadata samples from different triggering events are continuously accumulated. The accumulated samples are evaluated for validity and confidence. When the accumulated amount of the samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

4. The data processing method for an elderly comorbidity health information cloud platform according to claim 3, characterized in that, The process of evaluating the validity and confidence of the accumulated samples, and marking the samples as usable for rule generation when the accumulated amount of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, includes: A consistency check is performed on the accumulated samples to identify conflicting patterns in the samples; The weighting coefficients of the accumulated samples are adjusted according to the characteristics of the conflicting patterns. Aggregate patterns with similar features but minor differences to generate a comprehensive pattern; According to the comprehensive mode, the accumulated samples are evaluated for validity and confidence. When the accumulated amount of the samples reaches a preset threshold and the evaluation result reaches a preset confidence level, the samples are marked as usable for rule generation.

5. The data processing method for an elderly comorbidity health information cloud platform according to claim 4, characterized in that, The process of performing a consistency check on the accumulated samples to identify conflicting patterns within the samples includes: Extract associated device operating status parameters and patient physiological condition information from each of the accumulated samples; The device operating status parameters and the patient's physiological condition information are matched with preset context rules to identify the specific context environment in which the accumulated samples are located. Based on each identified specific context, activate a consistency check criterion for that specific context, which defines the expected structure, value range, and logical association with other known data of a non-standard metadata schema in that specific context. According to the consistency check criteria, the non-standard metadata patterns in the accumulated samples are compared. Sample patterns that meet the expected criteria in the specific context are marked as valid context-dependent variants; Sample patterns that deviate significantly from or logically contradict the expected standards in the specific context are identified as conflict patterns.

6. The data processing method for an elderly comorbidity health information cloud platform according to claim 5, characterized in that, After identifying sample patterns that significantly deviate from or logically contradict the expected standards in the specific context as conflict patterns, the method further includes: After identifying the conflict pattern, the timestamp of the occurrence of the conflict pattern is recorded, and the timestamp is associated with the device identifier and the event type; The frequency and duration of the aforementioned conflict patterns were analyzed. When the number of times the conflict pattern occurs consecutively within a preset time window reaches a preset threshold, or when its duration exceeds a preset duration, the conflict pattern is marked as a persistent conflict pattern. The samples marked as persistent conflict patterns are isolated from the accumulated samples, and further anomaly handling procedures are triggered.

7. The data processing method for an elderly comorbidity health information cloud platform according to claim 6, characterized in that, The analysis of the frequency and duration of the conflict patterns includes: Obtain device type information, patient group information, and patient disease status information associated with the conflict mode; Based on the device type information, extract the data acquisition rate, data accuracy, and data transmission stability parameters of the device type information from a preset set of device characteristic parameters; Based on the patient group information and the patient's disease status information, extract the normal fluctuation range, abnormal threshold, and typical duration of physiological events of the patient group under the disease status information from the preset physiological baseline parameter set; Based on the device characteristic parameters, the physiological baseline parameters, and the type of conflict mode, determine the counting window size and counting threshold for assessing the frequency of occurrence of the conflict mode, and the duration threshold for assessing the duration of the conflict mode; Within the defined counting window, the number of occurrences of the conflict mode is counted, and it is determined whether the number of occurrences reaches the counting threshold. Measure the duration of the conflict mode and determine whether the duration exceeds the duration threshold; Based on the results of the determination of the number of occurrences and the determination of the duration, a comprehensive evaluation is performed on the occurrence frequency and duration of the conflict pattern.

8. The data processing method for an elderly comorbidity health information cloud platform according to claim 4, characterized in that, The process of evaluating the validity and confidence of the accumulated samples, and marking the samples as usable for rule generation when the accumulated amount of samples reaches a preset threshold and the evaluation result reaches a preset confidence level, includes: Obtain the device source identifier and collection timestamp for each of the accumulated samples; Based on the device source identifier and the collection timestamp, retrieve the device's routine operating parameters and known physiological index ranges within a similar time period from historical data; For each of the accumulated samples, a preliminary anomaly assessment is performed, and the normal operating parameters and the range of physiological indicators are compared to identify samples that deviate from the normal range. The samples initially identified as abnormal are classified, and based on the degree of deviation and the source of the equipment, the abnormal samples are classified as potential equipment failure abnormalities or potential rare physiological event abnormalities. Based on the potential equipment failure anomaly samples, a dedicated equipment failure mode recognition mechanism is activated, which includes structured checks and numerical logic verification of specific equipment failure characteristics.

9. A data processing method for an elderly comorbidity health information cloud platform according to claim 8, characterized in that, The method further includes: Anomalies identified by the aforementioned equipment failure mode recognition mechanism are assigned evaluation weights. Assigning evaluation weights to abnormal samples identified by the equipment failure mode recognition mechanism includes: Obtain the type information of the abnormal sample; Obtain patient disease combination information associated with the abnormal sample; Obtain patient treatment plan information associated with the abnormal sample; Based on the type information of the abnormal sample, the corresponding initial weight value is found from the preset weight configuration table; Based on the patient's disease combination information and the patient's treatment plan information, extract the risk impact factors corresponding to the disease combination information and the treatment plan information from a preset risk impact factor table; The initial weight values ​​are adjusted with the risk impact factors to obtain the dynamically adjusted evaluation weights; The dynamically adjusted evaluation weights are assigned to the abnormal samples.

10. A data processing system for a cloud platform for health information on comorbidities in the elderly, characterized in that, include: The input end is used to receive data streams from medical devices and perform preliminary parsing of the data streams according to preset rules. In the preliminary parsing, byte sequences in the data stream that do not conform to the preset rules are identified, and the position, length and associated device identification information of the byte sequences are obtained. After the byte sequences are separated, the core measurement data in the data stream continues to be transmitted. The generation end is used to collect and group the position, length, and associated device identification information of the byte sequence, analyze the grouped information to discover the recurring structural patterns, and generate rules for parsing the byte sequence when the reliability of the structural patterns reaches a preset standard. The update endpoint is used to start an isolated test environment, use the rules to perform parsing tests on the data stream, and evaluate the effectiveness of the parsing rules; when the parsing rules pass the effectiveness evaluation, the parsing rules of the data stream are updated without interrupting the data stream transmission.

Images