Comprehensive Evaluation System and Methods for Competitive Sports Ability

The comprehensive evaluation system for competitive sports ability has solved the problems of data silos and low evaluation efficiency in competitive sports ability evaluation. It has enabled personalized dynamic early warning and real-time feedback, improving the accuracy of evaluation results and the efficiency of training adjustments.

CN123117548APending Publication Date: 2026-11-17SHANGHAI SPORTS SCI INST (SHANGHAI ANTI-DOPING CENT)
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611364468.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-04
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

In existing technologies, the assessment of competitive sports ability suffers from data silos, low assessment efficiency, lack of dynamic early warning and AI-assisted decision-making. As a result, data cannot be used for longitudinal data tracking of the same athlete across cycles or for horizontal group comparison of multiple athletes. The assessment report generation efficiency is low and the degree of standardization is insufficient. There is a lack of dynamic early warning and real-time benchmarking, and the objectivity and accuracy of the assessment results are greatly affected by the experience of experts.

Method used

A comprehensive evaluation system for competitive sports ability was designed, including a data layer module, an analysis layer module, an early warning layer module, and a feedback layer module. It realizes automatic collection and integration of multi-source data, standardized processing and secure storage, constructs individual benchmark intervals through the weighted moving average method, generates personalized evaluation reports by combining triplet corpus matching algorithm, and provides real-time feedback through mobile terminals.

Benefits of technology

It has achieved standardized, intelligent, and dynamic assessment of competitive athletic abilities, improved the consistency of data integration and the accuracy of assessment and analysis, shortened the report generation cycle, provided personalized dynamic early warnings and real-time feedback, and improved the efficiency and quality of training adjustments.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123117548A_ABST
    Figure CN123117548A_ABST
Patent Text Reader

Abstract

This application discloses a comprehensive evaluation system and method for competitive sports ability, comprising: multi-source data integration and standardization, collecting multi-dimensional competitive sports-related data from athletes and completing unified encoding, standardized conversion, and then encrypted hierarchical storage; multi-dimensional data comprehensive analysis, performing multi-dimensional comparative analysis on standardized data, generating visual professional charts, and enabling cross-module data interoperability; multi-level dynamic early warning and performance benchmarking, establishing individual benchmark intervals and three-level early warning thresholds based on athletes' historical data, monitoring and triggering visual early warnings in real time, and benchmarking competitive performance against standard records in real time; intelligent report generation and real-time feedback, calling standardized evaluation corpus based on analysis and early warning information, automatically generating a comprehensive evaluation report and pushing it to mobile terminals, forming a closed loop of research-training feedback. This application can realize multi-dimensional, accurate, and dynamic comprehensive evaluation of competitive sports ability, shortening the feedback cycle of research evaluation and training application.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the interdisciplinary field of sports science and information technology, and in particular to the digital and intelligent management technology for the comprehensive assessment of competitive athletes' abilities, specifically to a comprehensive assessment system and method for competitive sports abilities. Background Technology

[0002] In the field of competitive sports research, the assessment of athletes' competitive abilities is a core link in ensuring the scientific nature of training and improving athletic performance. Its assessment dimensions cover multiple interdisciplinary directions such as physiological function, psychological state, nutritional level, sports injury, technical movements and the effect of specific training. The accuracy and timeliness of the assessment results directly affect the formulation and adjustment of competitive sports training plans.

[0003] Currently, the assessment of competitive sports abilities adopts a traditional decentralized data management and manual analysis architecture. Various test data, such as physiological and biochemical test results, psychological scale scores, and body composition analysis data, are stored in separate databases or paper records in different laboratories, without a unified standardized data interface or integrated management platform. The typical assessment process is as follows: researchers obtain single-dimensional data from various specialized laboratories, manually organize and summarize it into Excel spreadsheets and other tools, combine it with their own professional experience to complete cross-dimensional analysis, and finally manually write the assessment report and provide it to the coaching team.

[0004] Existing technologies have many shortcomings and are no longer able to meet the refined and dynamic training support needs of competitive sports. Specifically, these shortcomings include: First, the data silos are prominent, and there is a lack of standardized integration mechanisms for multi-dimensional data, making it impossible to track longitudinal data of the same athlete across cycles and to compare multiple athletes in a horizontal group, making it difficult to uncover potential relationships between data. Second, the evaluation report generation efficiency is low and the standardization is insufficient. The report relies entirely on manual writing by researchers and lacks a unified professional evaluation lexicon. This easily leads to problems such as content duplication and non-standard expression. Furthermore, the scoring logic of psychological scales is complex and the norm data is updated late, resulting in poor comparability of test results from different batches and further extending the report production cycle. Third, there is a lack of dynamic early warning and real-time benchmarking mechanisms. The existing system only realizes static data storage and does not have the function of setting personalized warning values ​​based on the individual athlete's historical trajectory. It cannot actively warn of abnormal indicators, and the competitive performance indicators cannot be synchronized with the real-time updated world records and competition qualification standards, which is out of touch with the needs of dynamic adjustment of training. Fourth, the data interpretation is not deep enough and relies heavily on expert experience. There are no AI-assisted decision-making tools to support in-depth mining and quantitative analysis of multi-dimensional data. The objectivity and accuracy of the evaluation results are greatly affected by the professional background of researchers, making it difficult to provide refined guidance for training and adjustment.

[0005] To address the shortcomings of existing technologies, there is an urgent need to develop a comprehensive evaluation system for competitive sports abilities that integrates multi-dimensional data standardization management, cross-module comprehensive analysis, multi-level dynamic early warning, and AI intelligent interpretation. This system would achieve data integration consistency, accurate evaluation and analysis, real-time early warning response, and standardized report output, thereby improving the efficiency and quality of scientific research support for competitive sports. Summary of the Invention

[0006] The purpose of this application is to provide a comprehensive evaluation system and method for competitive sports ability that integrates multi-dimensional data standardization management, cross-module comprehensive analysis, multi-level dynamic early warning and AI intelligent evaluation, so as to solve the problems of data silos, low evaluation efficiency, lack of dynamic early warning and AI-assisted decision-making in the existing technology.

[0007] According to one aspect of this application, a comprehensive evaluation system for competitive sports ability is provided, comprising a data layer module, an analysis layer module, an early warning layer module, and a feedback layer module that form a closed-loop process and are linked sequentially, wherein: The data layer module is used to realize the automatic collection and integration of multi-source competitive sports-related data, standardized processing and secure storage, and to form standardized data; The analysis layer module performs multidimensional comprehensive analysis on standardized data, generates visualization results, and enables data interoperability across testing modules; Based on individual athletes' historical data, an individual benchmark interval is constructed, a personalized warning value system is established, multi-level dynamic early warning of indicators is realized, and real-time benchmarking of competitive performance with standard records is supported. The feedback layer module calls standardized evaluation corpus based on analysis and early warning information to generate personalized intelligent evaluation reports, and completes the closed-loop storage of real-time information push and feedback data through the mobile terminal interface.

[0008] Optionally, the system has the following linked algorithm mechanism: (1) The weighted moving average mechanism based on the time decay coefficient λ (λ ranges from 0.85 to 0.95) is calculated as WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), which gives higher weight to recent data and dynamically constructs individual benchmark intervals; (2) Based on the corpus precise matching mechanism of the intersection operation of indicator coding, threshold range, and warning level triplet labels, the evaluation corpus that best matches the athlete's current indicator status is retrieved from the standardized corpus; (3) Early warning reverse trigger supplementary collection mechanism: When an indicator triggers an orange or red warning, the system automatically pushes a designated psychological scale supplementary test task to the corresponding athlete, and the supplementary test results are fed back to the analysis layer module in real time to complete the re-analysis. (4) Feedback data self-iterative update mechanism: After the push is completed, the evaluation report, early warning record and training adjustment record will be back-synchronized to the storage module as the historical data input for the weighted moving average calculation of the next evaluation cycle, so as to realize the self-iterative update of the individual benchmark interval accumulated with the evaluation cycle.

[0009] Optionally, the data layer module includes a data access unit, a data standardization processing unit, and a data security storage unit, wherein: The data access unit connects to laboratory equipment, imports standardized templates, and utilizes a questionnaire platform to automatically collect multi-source data. The questionnaire platform is used for automatic scoring of the POMS and PSQI psychological scales, and during the scoring process, an iceberg profile deviation identification algorithm is used to provide priority-based warnings. The data access unit also includes a built-in corpus feedback collection interface, recording the adoption status and modifications made by coaches or researchers to each corpus segment in the evaluation report. This adoption rate data is used as a supervisory signal input to the corpus weight self-learning module of the feedback layer module. The corpus weight self-learning module dynamically adjusts the priority ranking of corpus items under similar indicator states based on historical adoption rates, achieving self-iterative optimization of the corpus retrieval strategy over the evaluation cycle. For example, when an athlete's vitality score is lower than their fatigue score, a specific marker is automatically triggered and the warning priority is upgraded. The data access unit also has a built-in corpus feedback collection interface to record the operation status of each corpus segment in each evaluation report by coaches or researchers, including three types of status markers: adoption (used directly without modification), modification (used after editing the corpus content), and deletion (removed from the report). The operation status of each corpus entry is associated with and stored with the corresponding indicator code, threshold range, and warning level triplet label. The above adoption rate statistics serve as a supervision signal and are periodically input into the corpus weight self-learning module in the feedback layer module. The corpus weight self-learning module dynamically adjusts the priority ranking weight of the corresponding corpus entry in triplet matching retrieval based on the historical adoption rate of each corpus entry under the same triplet label conditions, realizing continuous iterative optimization of the corpus retrieval strategy as the evaluation cycle accumulates.

[0010] The data standardization processing unit has a built-in dictionary of core indicators for competitive sports and an automatically updated norm database that is cross-categorized by gender, age group, sport, and training cycle stage. This database is used to achieve unified data encoding and standardized conversion, forming standardized data. The norm database is set with automatic update trigger conditions. When the amount of new data in any stratified subgroup reaches a preset threshold (usually 30 records by default), the mean and standard deviation of that stratum are automatically recalculated to ensure the timeliness and comparability of the norm data. The data security storage unit adopts a distributed database architecture, encrypts sensitive data using the AES encryption algorithm, and sets up a three-level role-based access control system for researchers, coaches, and athletes.

[0011] Optionally, the analysis layer module includes a comparative analysis unit, a visualization chart generation unit, and a cross-module data retrieval unit, wherein: The comparative analysis unit supports multi-dimensional comparisons of the same athlete across cycles, athletes from the same team / event across different categories, and cross-analysis of different categories of indicators, and provides customizable filtering conditions; The visualization chart generation unit automatically generates radar charts, trend charts, and scatter plots, drawing reference lines for key physiological indicators based on the mean ± 2 standard deviations of group data of the same sex, age group, and sport in the norm database. This visually presents the deviation of individual athlete indicators from the group level. When the amount of data in the target stratified subgroup is lower than the statistical reliability threshold (default 50 records), the system automatically merges it upwards to a broader stratum (in the following order: same sport group → same sex and age group → same sex group), and marks the stratification range on which the current reference line is based with a confidence level indicator in the chart, ensuring the statistical reliability of the reference line while reminding the user of the data limitations. The cross-module data call unit uses an indicator association mapping library to enable data interoperability between different test modules. The indicator association mapping library is configured with mapping rules for each monitoring indicator. For example, when the creatine kinase indicator triggers an abnormal warning, it automatically calls the body composition data of the physiological module and links it with the protein intake data of the dietary analysis module to form a triangular verification chain of "blood biochemistry - physiological function - nutrient intake".

[0012] Optionally, the early warning layer module includes a warning value setting unit, a performance benchmarking unit, and an early warning trigger display unit, wherein: The warning value setting unit calculates the historical average of individual athlete indicators using a weighted moving average method with a time decay coefficient λ, which ranges from 0.85 to 0.95. This weights the historical data, and the calculation formula is: WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), where n is the total number of data periods, i is the data period number, and Xi is the indicator value in the i-th period. Recent data receives higher weight. Based on this, an individual baseline interval is constructed using the weighted moving standard deviation, with the weighted average value ± 1.5 times the weighted moving standard deviation. Three warning thresholds (red, orange, and yellow) are set for different types of indicators, and manual adjustment is supported. The different types of indicators generally include recovery-sensitive indicators and slow-change indicators. For recovery-sensitive indicators (such as creatine kinase, blood lactate, etc.), a weighted moving standard deviation of 1.5 times is used as the yellow warning coefficient, and for slow-change indicators (such as hemoglobin, body fat percentage, etc.), a weighted moving standard deviation of 2.0 times is used as the yellow warning coefficient. This achieves adaptive matching of the warning threshold with the physiological characteristics of the indicator, avoiding the problem of warning hypersensitivity or sluggishness caused by fixed coefficients. The performance benchmarking unit compares athletes' competitive results with the latest competition records, world records, and event qualification records in real time. The early warning triggering display unit triggers visual early warnings through multi-colored icons on the front-end interactive interface, and generates a summary list of early warnings containing information such as abnormal indicators and early warning levels on the system homepage. At the same time, the early warning layer module has the ability to send reverse supplementary collection instructions. When any indicator triggers an orange or red early warning, the system automatically identifies the corresponding athlete's identifier and the type of indicator that triggered the early warning. Through the questionnaire platform interface in the data access unit, it pushes the designated psychological scale retest task to the corresponding athlete according to the preset "early warning level - scale type" mapping rule. After the retest is completed, the retest results, along with the indicator identifier that triggered the early warning, are fed back to the analysis layer module in real time. The comparative analysis unit then correlates and re-analyzes the results with the original analysis results to update the current early warning status, realizing an active closed loop of "early warning triggering → supplementary collection → dynamic reassessment".

[0013] Optionally, the feedback layer module includes a professional evaluation lexicon management unit, an intelligent report generation unit, and a real-time feedback push unit, wherein: The professional evaluation corpus management unit stores standardized evaluation expressions according to the dimensions of physiology, psychology, nutrition, and sports injury, and supports manual addition and automatic updating of corpus. The intelligent report generation unit uses a triplet corpus matching algorithm based on indicator threshold classification to call standardized corpora of corresponding dimensions in the professional evaluation corpus. It generates an evaluation report containing data charts, anomaly analysis, and improvement suggestions, and supports export and editing. Simultaneously, the intelligent report generation unit receives the output results from the performance benchmarking unit. When the gap between an athlete's performance and the benchmark record exceeds a preset threshold, the system dynamically increases the weight of improvement suggestion corpora in the report; when the performance is within the benchmark range, the weight of positive evaluation corpora is increased, achieving dynamic and personalized adjustments to the report content. Each standardized corpus in the professional evaluation corpus is labeled with a triplet tag consisting of an indicator code, a threshold range, and a warning level. The intelligent report generation unit determines the optimal matching corpus set through triplet intersection operations; that is, corpora that simultaneously meet the three conditions of indicator code matching, current indicator value falling within the corresponding threshold range, and consistent warning level have the highest priority. It should be noted that the technical implementation path of the intelligent report generation unit is: a combination of rule matching retrieval based on triplet labels (rule layer, deterministic output) and the historical adoption rate feedback mechanism of the corpus weight self-learning module (learning layer, iterative optimization based on supervision signals), rather than a generative model based on deep learning, to ensure the controllability and consistency of the report output.

[0014] Specifically, the intelligent report generation unit generates an evaluation report based on a triplet corpus matching algorithm through the following steps: Step 1, Triple Tag Retrieval: Each standardized corpus in the professional evaluation corpus is labeled with a triple tag consisting of an indicator code, a threshold range, and a warning level; the intelligent report generation unit uses the current athlete's indicator code, the range in which the current indicator value falls, and the current warning level as query keys, and determines the optimal matching corpus set that simultaneously meets the three conditions through triple intersection operation; Step 2, Dynamic Adjustment of Corpus Weights: Receive the output results from the performance benchmarking unit. When the gap between an athlete's performance and the qualifying record exceeds a preset threshold, dynamically increase the weight of improvement suggestions in the report; when the performance is within the qualifying range, increase the weight of positive evaluations; simultaneously, based on the historical adoption rate ranking output by the corpus weight self-learning module, further adjust the priority of similar corpus items. Step 3, Report Integration and Output: The optimal matching corpus is structurally integrated with the data charts, anomaly analysis, and targeted improvement suggestions generated by the analysis layer module to generate a personalized comprehensive evaluation report with four dimensions: physiological, psychological, nutritional, and athletic performance. The report supports format export and manual editing and optimization.

[0015] The real-time feedback push unit pushes reports and warning information to coaches and athletes through the WeChat official account and APP mobile terminal interface. After the real-time feedback push unit completes the push, it will reverse synchronize the current assessment report, warning record and coach training adjustment record to the data security storage unit. The feedback data serves as the historical data input for the calculation of the individual benchmark interval moving average in the next assessment cycle, realizing the self-iterative update of the benchmark interval as it accumulates with the assessment cycle.

[0016] Optionally, the intelligent report generation unit also works with a corpus weight self-learning module to achieve continuous iterative optimization of the corpus retrieval strategy. Specifically, the data access unit has a built-in corpus feedback collection interface that records the three types of operation states of coaches or researchers in the evaluation report: adoption, modification, and deletion. The above operation states are associated with the corresponding corpus entries' indicator codes, threshold ranges, and warning level triplet labels and stored. The above adoption rate statistics are periodically input into the corpus weight self-learning module as a supervision signal. The corpus weight self-learning module dynamically adjusts the priority ranking weight of the corresponding corpus entries in triplet matching retrieval based on the historical adoption rate of each corpus entry under the same triplet label conditions, and sends the updated weight value back to the intelligent report generation unit, forming a self-learning closed loop of "report generation → coach feedback → weight update → retrieval optimization", realizing continuous iterative optimization of the corpus retrieval strategy as the evaluation cycle accumulates.

[0017] In particular, the present invention also provides a method for comprehensive evaluation of competitive athletic ability, which is applied to the aforementioned comprehensive evaluation system for competitive athletic ability. The evaluation method includes the following steps: S1. Multi-source data integration and standardization: Collect multi-dimensional competitive sports-related data from athletes, complete unified encoding and standardized conversion, and then perform encrypted and hierarchical storage. S2. Multi-dimensional data comprehensive analysis: Performs multi-dimensional comparative analysis on standardized data, generates visual professional charts, and enables cross-module data interoperability. S3, Multi-level dynamic early warning and performance benchmarking: Based on the athlete's historical data, establish individual benchmark intervals and three-level early warning thresholds, monitor and trigger visual early warnings in real time, and benchmark the competitive performance against the standard records in real time; S4, intelligent report generation and real-time feedback: Based on analysis and early warning information, it calls standardized evaluation corpus, automatically generates a comprehensive evaluation report and pushes it to the designated mobile terminal, forming a closed loop of scientific research-training feedback.

[0018] Optionally, in step S1, the process of multi-source data integration and standardization includes multi-source data acquisition, standardization transformation, and encrypted hierarchical storage, wherein: In multi-source data acquisition, physiological and biochemical data and body composition data are automatically synchronized by connecting to laboratory equipment via the HL7 protocol, structured data are imported using standardized Excel / CSV templates, and psychological scale data are collected and automatically scored using a built-in questionnaire platform. Standardization conversion is achieved through a dictionary of core indicators for competitive sports and a classification norm database; The encrypted hierarchical storage adopts a distributed database architecture, uses the AES encryption algorithm to encrypt and store sensitive data, and allocates data access scope based on three-level role permissions.

[0019] Optionally, in step S2, the process of multi-dimensional data comprehensive analysis includes multi-dimensional comparative analysis, generation of professional visualization charts, and cross-module data interoperability, wherein: Multi-dimensional comparative analysis includes longitudinal analysis of the same athlete across cycles, horizontal analysis of athletes from the same team / same project, and cross-analysis of different categories of indicators. It also allows users to customize filtering conditions such as "only the most recent N tests" and "specific training phase". The visualization features professional charts including radar charts, trend charts, and scatter plots, and draws reference lines for the mean ± 2 standard deviations for key physiological indicators; Cross-module data interoperability includes the rehabilitation assessment module calling FMS / Y balance test results and the dietary analysis module linking body composition data.

[0020] Optionally, in step S3, the process of multi-level dynamic early warning and performance benchmarking includes, The weighted moving average method, using a time decay coefficient λ (ranging from 0.85 to 0.95), is used to weight historical data. The calculation formula is: WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), where n is the total number of data periods, i is the data period number, and Xi is the indicator value of the i-th period. Recent data receives higher weight. Based on this, an individual benchmark interval is constructed using the weighted moving standard deviation. Three levels of warning thresholds (red, orange, and yellow) are set for different types of indicators, and manual adjustment is supported. When an indicator exceeds the threshold, a warning is triggered via a multi-colored icon on the front-end UI, and a summary list of warnings is generated on the system homepage. Real-time benchmarking of athletic performance compares athletes' performance against the latest competition records, world records, and event qualification records. When an indicator triggers an orange or red alert, a reverse trigger supplementary data collection sub-step is executed: based on the "alert level - scale type" mapping rule, the system pushes a designated psychological scale supplementary test task to the corresponding athlete through the questionnaire platform. After the supplementary test is completed, the results, along with the indicator that triggered the alert, are fed back in real time to the multi-dimensional comparative analysis process in step S2 to complete the correlation reanalysis and update the current alert status, thus realizing a closed loop of "alert triggering → supplementary data collection → dynamic reassessment".

[0021] Optionally, in step S4, the process of generating intelligent reports and providing real-time feedback includes, Based on the indicator types of the analysis and early warning information, standardized evaluation corpora are matched for physiological, psychological, nutritional, and sports injury dimensions; By using a triplet corpus matching algorithm based on indicator threshold classification, combined with data charts, anomaly analysis, and targeted improvement suggestions, personalized or team-wide comprehensive evaluation reports are generated. The reports support format export and manual editing optimization. During the report generation process, the system considers overall competitive performance. When the gap between an athlete's competitive performance and the target record exceeds a preset threshold, the system dynamically increases the weight of improvement suggestion corpus in the report; when the performance is within the target range, the weight of positive evaluation corpus is increased, realizing dynamic and personalized adjustment of the report content. The assessment report and early warning information are pushed to the bound mobile terminals of coaches and athletes through WeChat official accounts and apps. After the push is completed, the assessment report, early warning records and coach training adjustment records are back-synchronized to the data security storage unit. The feedback data serves as the historical data input for the weighted moving average calculation of step S3 in the next assessment cycle, realizing the self-iterative update of the individual benchmark interval as the assessment cycle accumulates. At the same time, the coach's adoption status and modification records of the report corpus paragraphs are synchronized to the corpus weight self-learning module, serving as the supervisory signal input for the dynamic adjustment of corpus priority in the next cycle, forming a complete methodological closed loop of 'assessment-feedback-learning-reassessment'.

[0022] Compared with existing technologies, the comprehensive evaluation system and method for competitive sports ability provided in this application have the following beneficial effects or advantages: This application achieves standardization, intelligence, and dynamism in competitive sports ability assessment through a closed-loop linkage of the data layer, analysis layer, early warning layer, and feedback layer modules. It also ensures the integration, consistency, and storage security of multi-dimensional data. By automating data collection through multiple channels in the data access unit and unifying encoding and standardized conversion in the data standardization processing unit, it eliminates the risk of errors from manual entry and cross-system data transcription. Data from three channels—automatic synchronization via the HL7 protocol, import from Excel / CSV templates, and data collected from the questionnaire platform—all include a unified timestamp encoding, making timeline alignment errors from different sources controllable. Compared to the inconsistencies in time stamps caused by researchers manually entering data across systems in traditional manual summarization methods (errors can reach several hours to several days), this application fundamentally eliminates cross-dimensional data time misalignment, ensuring the horizontal comparability of multi-dimensional indicators. The aforementioned time alignment accuracy is achieved based on a unified conversion mechanism of HL7 protocol UTC timestamps, template-imported sampling time fields, and questionnaire platform server timestamps.

[0023] Regarding report generation efficiency, this application verifies the efficiency improvement through the following comparative tests: Compared with the traditional manual writing mode, this application's system significantly shortens the assessment report generation cycle through automated corpus retrieval and report formatting, improving efficiency by hundreds of times. Furthermore, the automatic update function of the norm database ensures the cross-period comparability of psychological scale data; simultaneously, the AES encryption and three-level role-based access management of the data security storage unit achieve secure storage and hierarchical access to sensitive data, preventing data leakage. In addition, it achieves accurate identification and quantitative comparison of competitive athletic abilities. Through multi-dimensional comparative analysis and custom filtering functions in the comparative analysis unit, it meets the data analysis needs of different assessment scenarios. Combined with the professional charts and reference lines drawn from the mean ± 2 standard deviation in the visualization chart generation unit, it intuitively displays the degree to which athletes' indicators deviate from the group mean and their own historical performance, helping researchers to keenly detect potential risks such as overtraining and functional decline, improving the accuracy and intuitiveness of the analysis results.

[0024] This application establishes a proactive and personalized dynamic early warning and performance benchmarking system. The personalized benchmark interval has a significant advantage in early warning sensitivity compared to the industry-standard fixed reference range. Taking creatine kinase as an example: when using a fixed reference range (the upper limit of the male resting value is approximately 350 U / L), for athletes whose baseline value has been consistently at 150 U / L, an increase to 300 U / L (a 100% increase) will not trigger an early warning. However, the individual benchmark interval constructed based on the weighted moving average method in this application can identify abnormal states where the athlete's relative increase reaches the yellow warning threshold. The early warning sensitivity is significantly better than the fixed threshold scheme, enabling the coaching team to intervene and adjust the training plan in a timely manner before the athlete exhibits obvious injury symptoms. By setting personalized benchmark intervals and three-level warning thresholds based on the weighted moving average method in the warning value setting unit, the warning thresholds can be customized. Combined with the visual warning and centralized display of the warning trigger display unit, the system transforms from "passive query" to "proactive notification". Coaching teams can adjust training plans in time before athletes suffer injuries or functional decline, reducing the risk of injuries in competitive sports. At the same time, the real-time benchmarking function of the performance benchmarking unit intuitively displays the gap between athletes and standard records, providing a clear direction for performance improvement.

[0025] In addition, this application uses a triplet matching algorithm to lock the corpus output results under the same indicator state, ensuring that the report descriptions corresponding to the same indicator type, the same threshold range, and the same warning level are completely consistent. This achieves rigid uniformity of the evaluation standard at the algorithm level, eliminating the interference of subjective experience differences on the evaluation conclusion. The above efficiency improvement is directly due to the automatic replacement of the corpus retrieval process by the triplet matching algorithm and the elimination of cross-dimensional data integration steps by the unified coding system.

[0026] It is generally accepted in the existing technical field that psychological scale data and physiological and biochemical index data are difficult to directly correlate and visualize on the same analytical platform due to significant differences in their dimensional systems and inconsistent collection frequencies (psychological scales are collected periodically, while physiological indicators are collected continuously or at high frequencies). This application overcomes the aforementioned technical bias by converting the raw scores of psychological scales into standardized scores in the same rating system as physiological indicators through "four-dimensional cross-normative standardized conversion," and by combining it with a unified time axis coding mechanism. For the first time, it achieves direct comparison and presentation of physiological and biochemical indicators and psychological scale scores on the same radar chart and trend chart, providing a technical foundation for researchers to identify the risk of overtraining from a cross-psychological perspective.

[0027] In summary, this application shortens the feedback loop delay between research and training. By using the WeChat official account and mobile APP push function of the real-time feedback push unit, the evaluation report and early warning information can be directly delivered to coaches and athletes. This solves the problem of excessively long "settling" time of test data in the laboratory in the traditional model, ensuring that the research evaluation results can directly guide the training load arrangement of the day and achieve deep synergy between scientific research support and actual combat training.

[0028] The above and other objects, advantages and features of this application will become more apparent to those skilled in the art from the following detailed description of specific embodiments of this application in conjunction with the accompanying drawings. Attached Figure Description

[0029] The following sections will describe some specific embodiments of this application in detail by way of example and not limitation, with reference to the accompanying drawings. The same reference numerals in the drawings denote the same or similar parts or components. Those skilled in the art should understand that these drawings are not necessarily drawn to scale. In the drawings: Figure 1 This is a schematic diagram of the framework structure of a comprehensive evaluation system for competitive sports ability according to an embodiment of this application; Figure 2 This is a schematic diagram of the workflow of a comprehensive evaluation method for competitive sports ability according to an embodiment of this application; Figure 3 This is a schematic diagram illustrating the specific workflow of a comprehensive evaluation method for competitive sports ability according to an embodiment of this application. Detailed Implementation

[0030] The technical solution of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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] Furthermore, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other. Example 1:

[0032] The comprehensive evaluation system for competitive sports ability described in this embodiment, such as Figure 1 As shown, it includes a data layer module, an analysis layer module, an early warning layer module, and a feedback layer module that form a closed loop throughout the entire process and are linked sequentially. Detailed functional descriptions of each module and unit are provided in the invention description.

[0033] In one implementation, the data layer module includes a data access unit (automatic collection of multi-source data and scoring of psychological scales), a data standardization processing unit (unified encoding and norm-standardized conversion), and a data security storage unit (AES encryption and three-level access control).

[0034] In one implementation, the analysis layer module includes a comparative analysis unit (multi-dimensional comparative analysis), a visualization chart generation unit (radar chart / trend chart / scatter plot and reference line drawing), and a cross-module data call unit (cross-module inter-invoking driven by indicator association mapping library).

[0035] In one implementation, the early warning layer module includes an alert value setting unit (using a weighted moving average method to construct individual baseline intervals and three-level early warning thresholds), a performance benchmarking unit (comparing competitive performance with standard records in real time), and an early warning trigger display unit (using multi-color icons to visualize early warnings and summary lists).

[0036] In one implementation, the feedback layer module includes a professional evaluation corpus management unit (storage and update of standardized corpus by dimension), an intelligent report generation unit (generation of personalized reports driven by triple matching algorithm), and a real-time feedback push unit (mobile terminal push and closed-loop storage of feedback data). Example 2:

[0037] This embodiment describes a method for comprehensive evaluation of competitive athletic ability, which is applied to the aforementioned comprehensive evaluation system for competitive athletic ability, such as... Figure 2 As shown, the assessment method includes the following steps: S1. Multi-source data integration and standardization: Collect multi-dimensional competitive sports-related data from athletes, complete unified encoding and standardized conversion, and then perform encrypted and hierarchical storage. S2. Multi-dimensional data comprehensive analysis: Performs multi-dimensional comparative analysis on standardized data, generates visual professional charts, and enables cross-module data interoperability. S3, Multi-level dynamic early warning and performance benchmarking: Based on the athlete's historical data, establish individual benchmark intervals and three-level early warning thresholds, monitor and trigger visual early warnings in real time, and benchmark the competitive performance against the standard records in real time; S4, intelligent report generation and real-time feedback: Based on analysis and early warning information, it calls standardized evaluation corpus, automatically generates a comprehensive evaluation report and pushes it to the designated mobile terminal, forming a closed loop of scientific research-training feedback.

[0038] Please continue reading. Figure 3 .

[0039] In one implementation, step S1, the process of multi-source data integration and standardization includes multi-source data acquisition, standardization conversion, and encrypted hierarchical storage, wherein: In multi-source data acquisition, physiological and biochemical data and body composition data are automatically synchronized by connecting to laboratory equipment via the HL7 protocol, structured data are imported using standardized Excel / CSV templates, and psychological scale data are collected and automatically scored using a built-in questionnaire platform. Standardization conversion is achieved through a dictionary of core indicators for competitive sports and a classification norm database; The encrypted hierarchical storage adopts a distributed database architecture, uses the AES encryption algorithm to encrypt and store sensitive data, and allocates data access scope based on three-level role permissions.

[0040] In one implementation, step S2, the process of multi-dimensional data comprehensive analysis includes multi-dimensional comparative analysis, generation of professional visualization charts, and cross-module data interoperability, wherein: Multi-dimensional comparative analysis includes longitudinal analysis of the same athlete across cycles, horizontal analysis of athletes from the same team / same project, and cross-analysis of different categories of indicators. It also allows users to customize filtering conditions such as "only the most recent N tests" and "specific training phase". The visualization features professional charts including radar charts, trend charts, and scatter plots, and draws reference lines for the mean ± 2 standard deviations for key physiological indicators; Cross-module data interoperability includes the rehabilitation assessment module calling FMS / Y balance test results and the dietary analysis module linking body composition data.

[0041] In one implementation, step S3, the process of multi-level dynamic early warning and performance benchmarking includes, The system analyzes athletes' historical test data using a weighted moving average method to generate individual indicator baseline ranges, and supports manual adjustment of the baseline ranges and three-level warning thresholds (red, orange, and yellow). When an indicator exceeds a threshold, a warning is triggered via a multi-colored icon on the front-end UI, and a summary list of warnings is generated on the system homepage. Real-time benchmarking of competitive results involves comparing athletes' competitive performance with the latest competition records, world records, and event qualification records item by item.

[0042] In one implementation, step S4, the process of generating and providing real-time feedback of the intelligent report, includes: Based on the indicator types of the analysis and early warning information, standardized evaluation corpora are matched for physiological, psychological, nutritional, and sports injury dimensions; By using a triplet corpus matching algorithm based on index threshold classification, combined with data charts, anomaly analysis, and targeted improvement suggestions, personalized or team-wide comprehensive evaluation reports are generated. The reports also support format export and manual editing and optimization. The assessment report and early warning information will be pushed to the coaches' and athletes' linked mobile terminals through the official account and APP.

[0043] Taking the application of the comprehensive evaluation system and method for competitive sports ability of the present invention in the comprehensive evaluation of swimmers' body composition, psychological state and special performance as an example, the module linkage, unit function and execution steps of the method of the system are explained in detail.

[0044] System hardware and software fundamentals The comprehensive evaluation system for competitive sports ability of this invention is built on a cloud computing platform and adopts a B / S architecture. This embodiment uses the following technology stack as an example for specific illustration, but the invention is not limited to this. Other equivalent front-end frameworks, back-end frameworks, and distributed database architectures can be selected according to actual needs: the front-end interactive interface is developed based on Vue.js, the back-end adopts the SpringBoot framework, the distributed database adopts the Hadoop+HBase architecture, and has built-in HL7 protocol parsing plugins, natural language generation algorithm plugins, and weighted moving average statistical analysis plugins. At the same time, it connects to the WeChat Official Account Open Platform and the self-developed competitive sports training APP interface to realize cloud storage, automated processing, and mobile terminal push of data.

[0045] Connection relationship between system modules and units The comprehensive evaluation system for competitive sports ability of this invention includes a data layer module, an analysis layer module, an early warning layer module, and a feedback layer module linked sequentially, with each module forming a closed loop of data flow. The data layer module's data access unit, data standardization processing unit, and data security storage unit are linked serially. The output of the data access unit is connected to the input of the data standardization processing unit, and the output of the data standardization processing unit is simultaneously connected to the input of the data security storage unit and the input of the comparison analysis unit in the analysis layer module. The output of the cross-module data retrieval unit in the analysis layer module is connected to the input of the comparison analysis unit, and the output of the comparison analysis unit is connected to the input of the visualization chart generation unit. The output of the visualization chart generation unit... Simultaneously, it connects to the input end of the early warning trigger display unit of the early warning layer module and the input end of the AI ​​intelligent report generation unit of the feedback layer module; the output end of the warning value setting unit of the early warning layer module is connected to the input end of the early warning trigger display unit, and the output end of the comparative analysis unit is also connected to the input end of the performance benchmarking unit. The output ends of the performance benchmarking unit and the early warning trigger display unit are both connected to the input end of the AI ​​intelligent report generation unit; the output end of the professional evaluation linguistics management unit of the feedback layer module is connected to the input end of the AI ​​intelligent report generation unit, and the output end of the AI ​​intelligent report generation unit is connected to the input end of the real-time feedback push unit. The feedback data of the real-time feedback push unit can be reverse-synchronized to the data security storage unit to complete the closed-loop storage of the evaluation data. In addition, the feedback layer module has a built-in corpus weight self-learning module. Its input receives adoption rate statistics from the corpus feedback collection interface in the data access unit, as well as the current corpus matching record (including corpus entry identifier, corresponding triplet label, and corpus position weight in the report) from the intelligent report generation unit. The output of the corpus weight self-learning module is connected to the corpus priority sorting input of the intelligent report generation unit, and sends the updated weight values ​​of each corpus entry back to the intelligent report generation unit. In the next triplet matching retrieval, the corpus entries with higher adoption rates are prioritized, forming a learning closed loop of "report generation → coach feedback → weight update → retrieval optimization".

[0046] Specific Implementation Steps of the Comprehensive Assessment Method for Competitive Sports Ability This embodiment takes the monthly comprehensive evaluation of male 100-meter freestyle swimmers from a provincial swimming team as an example to implement the comprehensive evaluation method for competitive sports ability of the present invention. The specific steps are as follows: S1. Multi-source data integration and standardization This step utilizes the data layer module to collect, standardize, and encrypt the swimmer's multi-dimensional data. Specifically: Multi-source data acquisition: The data access unit connects to the body composition analyzer and physiological and biochemical analyzer of the sports medicine laboratory via the HL7 protocol or standard device data interface (e.g., RS232 serial port, USB or manufacturer's proprietary API), automatically synchronizing the athlete's body composition data (body fat percentage 12.5%, muscle mass 68.2kg) and blood biochemical data (hemoglobin 152g / L, creatine kinase 380U / L); importing the athlete's recent dietary survey structured data via a standardized Excel template; and pushing the POMS mood status scale and PSQI sleep quality scale to the athlete through the built-in questionnaire platform. After the athlete completes the questionnaire, the questionnaire platform automatically completes the scoring, obtaining the scores of each dimension of the POMS scale (tension-anxiety, depression-depression, anger-hostility, vitality-activity, fatigue-sluggishness, confusion-perplexity) and the total mood disorder score (TMD). The fatigue dimension score is 18 points and the anxiety dimension score is 12 points, with the TMD in the high range; the PSQI scale total score is 6 points (≥5 points indicate sleep quality problems). The specific judgment logic of the iceberg profile bias identification algorithm is as follows: The POMS scale includes six dimensions: Vitality-Active (V), Tension-Anxiety (T), Depression-Frustration (D), Anger-Hospitality (A), Fatigue-Sluggishness (F), and Confusion-Perplexity (C). Normatively healthy athletes exhibit an "iceberg-shaped" profile with "high vitality and low scores in the other five dimensions" (V dimension scores are significantly higher than the mean scores of T, D, A, F, and C). This algorithm performs the following two-level judgment: Level 1 Judgment (Vigor-Fatigue Ratio Test): Calculate the V / (F+1) ratio. When the ratio is lower than the preset threshold Rv (default Rv=1.0, i.e. fatigue dimension score ≥ vitality dimension score), trigger Level 1 specificity marker. The system upgrades the athlete's POMS warning priority from "normal" to "attention". The second level of judgment (overall negative emotional load test): Calculate the total mood disorder score (TMD) = (T + D + A + F + C) - V. When the TMD exceeds 1.5 times the standard deviation of the norm mean for the same gender, age group, and sport, a second-level specificity marker is triggered. The system further upgrades the warning priority to "abnormal" and marks it in orange in the warning summary list, triggering the warning reverse supplementary data collection mechanism. Taking this embodiment as an example, the athlete's fatigue dimension score is 18 points, and the vitality dimension score (assumed to be 10 points). V / (F+1) = 10 / 19 ≈ 0.53, which is lower than Rv = 1.0, triggering a first-level marker. The calculated TMD is in the high range, so the system sets the warning priority to the "attention" level and records it in the warning summary list.

[0047] Standardization Conversion: The data standardization processing unit assigns unique codes to indicators such as "body fat percentage", "creatine kinase", and "POMS fatigue dimension score" through the built-in competitive sports core indicator dictionary, completing the unified coding of data; at the same time, it calls the norm database classified by gender (male), age group (20-25 years old), and sport (swimming) to convert the original scores of POMS and PSQI scales into standardized scores. Among them, the standardized score of POMS fatigue dimension is in the high range, and the total score of PSQI scale is 6 points, which is higher than the cutoff value (≥5 points). The standardized rating is that there is a mild problem with sleep quality.

[0048] Encrypted hierarchical storage: The data security storage unit synchronizes the aforementioned standardized data to a distributed database and uses the AES encryption algorithm to encrypt sensitive data such as blood biochemistry data and psychological scale scores. At the same time, based on a three-level role-based access control system, researchers are granted full access to the data, coaches are granted access to standardized psychological scale rating results (such as "normal / attention / abnormal" levels) and other data analysis results, the original scores of psychological scales and the original data of each dimension are only accessible to psychological researchers, and athletes are granted access to personal body composition and specialty performance data.

[0049] S2, Multi-dimensional Data Comprehensive Analysis This step utilizes the analysis layer module to perform multi-dimensional comparative analysis, visualization, and cross-module data exchange on standardized data. Specifically: Cross-module data access: The cross-module data access unit obtains the FMS functional movement screening results (score of 18 points, higher than the injury risk cutoff value of 14 points, indicating normal movement pattern) of the rehabilitation assessment module through the indicator association mapping library. It then associates and accesses the FMS test results (score of 18 points) of the rehabilitation assessment module, the recent dietary survey data and body composition data of the dietary analysis module, and the physiological and biochemical data to provide full-dimensional data support for multi-dimensional comparative analysis.

[0050] Multi-dimensional comparative analysis: Researchers set the screening criteria in the comparative analysis unit as "test data of the past 3 months" and "male 100-meter freestyle swimmers in the same team". The unit automatically completed three analyses: First, longitudinal analysis of the athlete across cycles, comparing the changes in body fat percentage and creatine kinase over the past 3 months, found that creatine kinase increased by 20% compared to the previous month; Second, cross-sectional analysis of athletes in the same team, comparing the athlete's body fat percentage and hemoglobin with the average values ​​of athletes in the same team and event, found that the body fat percentage was 1.2% lower than the team average, and the hemoglobin was 152g / L, within the team average range, but higher than the warning value for exercise-induced anemia screening in male athletes (140g / L), indicating no risk of anemia at present; Third, cross-analysis of indicators, cross-analyzing protein intake in the dietary survey with muscle mass and body fat percentage, found that the daily protein intake was 1.8g / kg, lower than the recommended intake for swimming (2.0-2.2g / kg / day, referring to the ISSN Sports Nutrition Guidelines).

[0051] Visualization chart generation: Based on the above analysis results, the visualization chart generation unit automatically generates a trend chart of the athlete's physiological indicators over the past 3 months, a radar chart comparing the indicators of athletes in the same team, a scatter plot of protein intake and body fat percentage, and draws a reference line of "mean ± 2 standard deviations" for key physiological indicators such as creatine kinase and hemoglobin. The creatine kinase value is at the upper limit of the reference line, close to the warning threshold.

[0052] S3, Multi-level Dynamic Early Warning and Performance Benchmarking This step utilizes the early warning layer module to complete personalized early warnings, monitor abnormal indicators, and benchmark against swimming segment performance. Specifically: Personalized warning value setting: The warning value setting unit statistically analyzes the athlete's creatine kinase data over the past 12 months using a weighted moving average method. Combining the sports medicine reference range (the upper limit of CK resting reference for male athletes is approximately 350 U / L) with the athlete's historical average over the past 12 months, an individual baseline range of 150-350 U / L is generated. Three warning thresholds are set: red, orange, and yellow. The yellow threshold is 350-400 U / L (mild fatigue warning), the orange threshold is 400-450 U / L (moderate fatigue warning), and the red threshold is >450 U / L (severe fatigue warning, recommending reduction or cessation of training). Researchers confirm that the baseline range and warning thresholds do not need to be adjusted based on the athlete's training load.

[0053] Dynamic indicator warning: The warning trigger display unit monitors the athlete's creatine kinase data (380U / L) in real time. If it finds that the data exceeds the yellow threshold (350U / L), a yellow warning icon is triggered on the system front-end interactive interface. The abnormal indicator (creatine kinase), warning level (yellow), associated athlete, test time and other information are entered into the warning summary list on the system homepage.

[0054] Competitive Performance Benchmarking: The performance benchmarking unit compared the athlete's recent three key split times in the 100m freestyle (start reaction time 0.68s, turn speed 1.8m / s, sprint speed 2.2m / s; the system supports extended configuration of stroke frequency, stroke length, 50m split time, etc.) with the recent National Swimming Championships men's 100m freestyle qualifying record and world record. It was found that the start reaction time was 0.05s slower than the qualifying record, while the sprint speed was the same as the qualifying record.

[0055] S4, Intelligent Report Generation and Real-time Feedback This step utilizes the feedback layer module to generate and push the standardized evaluation report to mobile devices. Specifically: Corpus matching: Based on the analysis results of step S2 and the warning information and performance benchmarking results of step S3, the AI ​​intelligent report generation unit matches standardized evaluation corpus of the corresponding dimension from the professional evaluation corpus management unit, such as "creatine kinase value 380U / L, exceeding the individual yellow threshold, indicating a risk of mild muscle fatigue", "body fat percentage 12.5%, lower than the team average, meeting the specific requirements of swimming", "start reaction time 0.68s, slower than the national championship qualifying record by 0.05s, requiring strengthened start technique training", etc.

[0056] AI report generation: Using a triplet corpus matching algorithm based on index threshold classification, the matched standardized corpus is integrated with the trend chart, radar chart, scatter plot generated in step S2, as well as the abnormal cause analysis (elevated creatine kinase may be related to excessive recent interval training load, insufficient protein intake affects muscle recovery), and targeted improvement suggestions (reduce interval training intensity, increase protein intake to 2.0-2.2g / kg per day, strengthen starting technique training) to generate a monthly personalized comprehensive evaluation report for the athlete. After researchers perform simple editing and optimization of the report content, the report is exported as a PDF.

[0057] Real-time feedback push: The real-time feedback push unit pushes the yellow warning information and the monthly comprehensive evaluation report to the athlete's head coach and the athlete's bound mobile terminal through the WeChat official account and the self-developed competitive sports training APP. Based on the evaluation report and warning information, the coach immediately adjusts the athlete's training plan for the following week, reduces the interval training load, and increases guidance on protein supplementation and special training on starting techniques.

[0058] Upon completion of this assessment, the feedback data from the real-time feedback push unit and the coach's training adjustment records are synchronously synchronized to the data security storage unit, completing the closed-loop storage of the athlete's entire assessment data and providing data support for subsequent cross-cycle longitudinal analysis. Regarding the corpus feedback collection and self-learning update process, the specific steps are as follows: After the assessment report was pushed out, the athlete's head coach reviewed the report content through the APP. Paragraphs such as "creatine kinase value 380 U / L, exceeding the individual yellow threshold, indicating a mild risk of muscle fatigue" were not modified (marked as adopted), while the specific load values ​​in the paragraph "suggest reducing interval training intensity" were edited (marked as modified). The corpus feedback collection interface in the data access unit automatically recorded the above operation status and associated it with the corresponding corpus entry's triplet label (indicator code: CK, threshold range: 350-400 U / L, warning level: yellow) and stored it in the data security storage unit.

[0059] The adoption rate data is input as a supervisory signal to the corpus weight self-learning module at the end of this evaluation cycle. This module increases the weight of corpus entries with high adoption rates under the "CK-350~400U / L-yellow" triplet label, appropriately decreases the weight of corpus entries that will be modified, and sends the updated weight values ​​back to the intelligent report generation unit. In the next evaluation cycle (such as the monthly evaluation next month), when the same athlete or similar athletes have the same triplet label combination, the intelligent report generation unit will prioritize calling corpus entries with higher historical adoption rates to generate an evaluation report that better suits the coach's usage habits, thus achieving continuous improvement in corpus quality over the evaluation cycle.

[0060] It should be noted that the comprehensive evaluation system and method for competitive sports ability of the present invention is not only applicable to the comprehensive evaluation of swimmers, but can also be applied to the evaluation of athletes' abilities in various competitive sports such as track and field, ball games, and gymnastics. By simply adjusting the classification dimensions of the core indicator dictionary and norm database of competitive sports in the data standardization processing unit according to the specific characteristics of different sports, and updating the standard record database of the corresponding project in the performance benchmarking unit, personalized comprehensive evaluation of different competitive sports can be achieved. It has a wide range of application scenarios and promotional value.

[0061] The above description is merely a preferred embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A comprehensive evaluation system for competitive sports ability, characterized in that, This includes a data layer module, an analysis layer module, an early warning layer module, and a feedback layer module that form a closed-loop process and are linked sequentially, wherein: The data layer module is used to realize the automatic collection and integration of multi-source competitive sports-related data, standardized processing and secure storage, and to form standardized data; The analysis layer module performs multi-dimensional comprehensive analysis on the standardized data, generates visualization results, and enables data interoperability across test modules; The early warning layer module constructs an individual benchmark interval based on the athlete's individual historical data, establishes a personalized warning value system, realizes multi-level dynamic early warning of indicators, and supports real-time benchmarking of competitive performance with standard records; The feedback layer module calls standardized evaluation corpus based on analysis and early warning information to generate personalized intelligent evaluation reports, and completes real-time information push and closed-loop storage of feedback data through the mobile terminal interface.

2. The comprehensive evaluation system for competitive sports ability according to claim 1, characterized in that, The system has the following interconnected algorithm mechanism: (1) The weighted moving average mechanism based on the time decay coefficient λ is calculated as WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), which gives higher weight to recent data and dynamically constructs individual benchmark intervals. The value of λ ranges from 0.85 to 0.

95. (2) Based on the corpus precise matching mechanism of the intersection operation of indicator coding, threshold range, and warning level triplet labels, the evaluation corpus that best matches the athlete's current indicator status is retrieved from the standardized corpus; (3) Early warning reverse trigger supplementary collection mechanism: When an indicator triggers an orange or red warning, the system automatically pushes a designated psychological scale supplementary test task to the corresponding athlete, and the supplementary test results are fed back to the analysis layer module in real time to complete the re-analysis. (4) Feedback data self-iterative update mechanism: After the push is completed, the evaluation report, early warning record and training adjustment record will be back-synchronized to the storage module as the historical data input for the weighted moving average calculation of the next evaluation cycle, so as to realize the self-iterative update of the individual benchmark interval accumulated with the evaluation cycle.

3. The comprehensive evaluation system for competitive sports ability according to claim 1, characterized in that, The data layer module includes a data access unit, a data standardization processing unit, and a data security storage unit, wherein: The data access unit connects to laboratory equipment, imports standardized templates, and uses a questionnaire platform to automatically collect multi-source data. The questionnaire platform is used to automatically score POMS and PSQI psychological scales, and during the scoring process, it uses an iceberg profile deviation identification algorithm to output warnings in a priority-based manner. The data access unit also has a built-in corpus feedback collection interface to record the adoption status and modification content of each corpus paragraph in the evaluation report by coaches or researchers. The above adoption rate data is used as a supervision signal to be input into the corpus weight self-learning module of the feedback layer module. The corpus weight self-learning module dynamically adjusts the priority ranking of each corpus item under the same indicator status based on the historical adoption rate, realizing the self-iterative optimization of the corpus retrieval strategy as the evaluation cycle accumulates. The data standardization processing unit has a built-in dictionary of core indicators for competitive sports and an automatically updated norm database that is cross-categorized by gender, age group, sport, and training cycle stage. This database is used to achieve unified data encoding and standardized conversion to form the standardized data. The norm database is set with automatic update trigger conditions. When the amount of new data in any stratified subgroup reaches a preset threshold, the mean and standard deviation of that stratum are automatically recalculated. The data security storage unit adopts a distributed database architecture, encrypts sensitive data using the AES encryption algorithm, and sets up three-level role access management for researchers, coaches, and athletes.

4. The comprehensive evaluation system for competitive sports ability according to claim 1, characterized in that, The analysis layer module includes a comparative analysis unit, a visualization chart generation unit, and a cross-module data retrieval unit, wherein: The comparative analysis unit supports multi-dimensional comparisons of the same athlete across cycles, athletes from the same team / event across different categories of indicators, and provides customizable filtering conditions. The visualization chart generation unit automatically generates radar charts, trend charts, and scatter plots to draw reference lines based on the mean ± 2 standard deviations of group data of the same gender, age group, and sport in the norm database for key physiological indicators, so as to intuitively present the degree of deviation between individual athlete indicators and group level. The cross-module data call unit uses an indicator association mapping library to enable data interoperability between different test modules. The indicator association mapping library configures mapping rules for each monitoring indicator.

5. The comprehensive evaluation system for competitive sports ability according to claim 1, characterized in that, The early warning layer module includes a warning value setting unit, a performance benchmarking unit, and an early warning trigger display unit, wherein: The warning value setting unit calculates the historical average of individual athlete indicators using a weighted moving average method. A time decay coefficient λ, ranging from 0.85 to 0.95, is used to weight the historical data. The calculation formula is: WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), where n is the total number of data periods, i is the data period number, and Xi is the indicator value in the i-th period. Recent data receives higher weight. An adaptive warning coefficient is used for indicators with different physiological characteristics: for recovery-sensitive indicators, a weighted mean ± 1.5 times the weighted moving standard deviation is used as the yellow warning coefficient. Recovery-sensitive indicators include, but are not limited to, short-period fluctuating indicators such as creatine kinase and blood lactate. For slowly changing indicators, a weighted mean ± 2.0 times the weighted moving standard deviation is used as the yellow warning coefficient. Slowly changing indicators include, but are not limited to, long-period steady-state indicators such as hemoglobin and body fat percentage. This achieves adaptive matching of the warning threshold with the physiological characteristics of the indicators, and allows for setting red, orange, and yellow three-level warning thresholds for different types of indicators, with manual adjustment supported. The performance benchmarking unit compares athletes’ competitive results with the latest competition records, world records and event qualification records in real time. The warning triggering display unit triggers visual warnings through multi-color icons on the front-end interactive interface, and generates a summary list of warnings containing information such as abnormal indicators and warning levels on the system homepage. At the same time, the warning layer module has the ability to send reverse supplementary collection instructions: when any indicator triggers an orange or red warning, the system automatically identifies the corresponding athlete's identifier and the type of indicator that triggered the warning, and pushes a designated psychological scale retest task to the corresponding athlete through the questionnaire platform interface in the data access unit according to the preset "warning level-scale type" mapping rule; after the retest is completed, the retest results, along with the indicator identifier that triggered the warning, are fed back to the analysis layer module in real time, and the comparative analysis unit correlates and re-analyzes them with the original analysis results to update the current warning status, forming an active closed loop of warning triggering → supplementary collection → dynamic reassessment.

6. The comprehensive evaluation system for competitive sports ability according to claim 1, characterized in that, The feedback layer module includes a professional evaluation lexicon management unit, an intelligent report generation unit, and a real-time feedback push unit, wherein: The professional evaluation corpus management unit stores standardized evaluation expressions according to the dimensions of physiology, psychology, nutrition, and sports injury, and supports manual addition and automatic updating of the corpus. The intelligent report generation unit generates an evaluation report based on a triplet corpus matching algorithm through the following steps: Step 1, Triple Tag Retrieval: Each standardized corpus in the professional evaluation corpus is labeled with a triple tag consisting of an indicator code, a threshold range, and a warning level; the intelligent report generation unit uses the current athlete's indicator code, the range in which the current indicator value falls, and the current warning level as query keys, and determines the optimal matching corpus set that simultaneously meets the three conditions through triple intersection operation; Step 2, Dynamic Adjustment of Corpus Weights: Receive the output results from the performance benchmarking unit. When the gap between an athlete's performance and the qualifying record exceeds a preset threshold, dynamically increase the weight of improvement suggestions in the report; when the performance is within the qualifying range, increase the weight of positive evaluations; simultaneously, based on the historical adoption rate ranking output by the corpus weight self-learning module, further adjust the priority of similar corpus items. Step 3, Report Integration and Output: The optimal matching corpus is structurally integrated with the data charts, anomaly analysis, and targeted improvement suggestions generated by the analysis layer module to generate a personalized comprehensive evaluation report with four dimensions: physiological, psychological, nutritional, and competitive performance. The report supports format export and manual editing and optimization. The real-time feedback push unit pushes reports and warning information to coaches and athletes through the official account and APP mobile terminal interface. After the real-time feedback push unit completes the push, it reverse-synchronizes the current evaluation report, warning record and coach training adjustment record to the data security storage unit. The feedback data is used as historical data input for the calculation of the individual benchmark interval weighted moving average in the next evaluation cycle.

7. The comprehensive evaluation system for competitive sports ability according to claim 6, characterized in that, The intelligent report generation unit also works with the corpus weight self-learning module to achieve continuous iterative optimization of the corpus retrieval strategy. Specifically, the data access unit has a built-in corpus feedback collection interface to record the three types of operation states of coaches or researchers in the evaluation report: adoption, modification, and deletion. The above operation states are associated with the corresponding corpus entries' indicator codes, threshold ranges, and warning level triple tags and stored. The above adoption rate statistics are periodically input into the corpus weight self-learning module as a supervision signal. The corpus weight self-learning module dynamically adjusts the priority ranking weight of the corresponding corpus entries in triplet matching retrieval based on the historical adoption rate of each corpus entry under the same triplet label conditions, and sends the updated weight value back to the intelligent report generation unit, forming a self-learning closed loop of "report generation → coach feedback → weight update → retrieval optimization", realizing the continuous iterative optimization of the corpus retrieval strategy as the evaluation cycle accumulates.

8. A method for comprehensive evaluation of competitive athletic ability, characterized in that, The evaluation method for the comprehensive evaluation system of competitive sports ability as described in any one of claims 1-7 includes the following steps: S1. Multi-source data integration and standardization: Collect multi-dimensional competitive sports-related data from athletes, complete unified encoding and standardized conversion, and then perform encrypted and hierarchical storage. S2. Multi-dimensional data comprehensive analysis: Performs multi-dimensional comparative analysis on standardized data, generates visual professional charts, and enables cross-module data interoperability. S3, Multi-level dynamic early warning and performance benchmarking: Based on the athlete's historical data, establish individual benchmark intervals and three-level early warning thresholds, monitor and trigger visual early warnings in real time, and benchmark the competitive performance against the standard records in real time; S4, intelligent report generation and real-time feedback: Based on analysis and early warning information, it calls standardized evaluation corpus, automatically generates a comprehensive evaluation report and pushes it to the designated mobile terminal, forming a closed loop of scientific research-training feedback.

9. The method for comprehensive evaluation of competitive athletic ability according to claim 8, characterized in that, In step S1, the process of multi-source data integration and standardization includes multi-source data acquisition, standardization transformation, and encrypted hierarchical storage, wherein: In multi-source data acquisition, physiological and biochemical data and body composition data are automatically synchronized by connecting to laboratory equipment via the HL7 protocol, structured data are imported using standardized Excel / CSV templates, and psychological scale data are collected and automatically scored using a built-in questionnaire platform. Standardization conversion is achieved through a dictionary of core indicators for competitive sports and a classification norm database; The encrypted hierarchical storage adopts a distributed database architecture, uses the AES encryption algorithm to encrypt and store sensitive data, and allocates data access scope based on three-level role permissions.

10. The method for comprehensive evaluation of competitive athletic ability according to claim 8, characterized in that, In step S2, the multi-dimensional data comprehensive analysis process includes multi-dimensional comparative analysis, generation of professional visualization charts, and cross-module data interoperability, among which: Multi-dimensional comparative analysis includes longitudinal analysis of the same athlete across cycles, horizontal analysis of athletes from the same team / same event, and cross-analysis of different categories of indicators. It also allows users to customize filtering conditions such as "only the most recent N tests" and "specific training phase". The visualization features professional charts including radar charts, trend charts, and scatter plots, and draws reference lines for the mean ± 2 standard deviations for key physiological indicators; Cross-module data interoperability includes the rehabilitation assessment module calling FMS / Y balance test results and the dietary analysis module linking body composition data.

11. The method for comprehensive evaluation of competitive athletic ability according to claim 8, characterized in that, In step S3, the process of multi-level dynamic early warning and performance benchmarking includes, The weighted moving average method is used to assign weights to historical data using a time decay coefficient λ, which ranges from 0.85 to 0.

95. The calculation formula is: WMA=Σ(λ^(ni)×Xi) / Σλ^(ni), where n is the total number of data periods, i is the data period number, and Xi is the indicator value of the i-th period. Recent data is given higher weights. Based on this, an individual benchmark interval is constructed by combining the weighted moving standard deviation. Three-level warning thresholds of red, orange, and yellow are set for different types of indicators and manual adjustment is supported. When an indicator exceeds a threshold, an alert is triggered via a multi-colored icon in the front-end UI, and a summary list of alerts is generated on the system homepage. The system benchmarks athletes' competitive performance in real time against the latest competition records, world records, and event qualification records. When an indicator triggers an orange or red warning, a supplementary data collection sub-step is executed: the system pushes a designated psychological scale supplementary test task to the corresponding athlete through the questionnaire platform according to the "warning level - scale type" mapping rule. After the supplementary test is completed, the results, along with the indicator that triggered the warning, are fed back to the multi-dimensional comparative analysis process in step S2 in real time to complete the correlation reanalysis and update the current warning status, thus realizing a closed loop of "warning trigger → supplementary data collection → dynamic reassessment".

12. The method for comprehensive evaluation of competitive athletic ability according to claim 8, characterized in that, In step S4, the process of intelligent report generation and real-time feedback includes, Based on the indicator types of the analysis and early warning information, standardized evaluation corpora are matched for physiological, psychological, nutritional, and sports injury dimensions; By using a triplet corpus matching algorithm based on index threshold classification, combined with data charts, anomaly analysis, and targeted improvement suggestions, personalized or team-wide comprehensive evaluation reports are generated. The reports support format export and manual editing optimization. The system considers overall competitive performance during the report generation process. When the gap between an athlete's competitive performance and the qualifying record exceeds a preset threshold, the system dynamically increases the weight of improvement suggestions in the report. When the performance falls within the acceptable range, the weight of positive evaluation corpora is increased to achieve dynamic and personalized adjustments to the report content; The assessment report and early warning information are pushed to the bound mobile terminals of coaches and athletes through WeChat official accounts and apps. After the push is completed, the assessment report, early warning record and coach training adjustment record are back-synchronized to the data security storage unit. The feedback data is used as the historical data input for the weighted moving average calculation of step S3 in the next assessment cycle, realizing the self-iterative update of the individual benchmark interval as the assessment cycle accumulates. At the same time, the coach's adoption status and modification record of the report corpus paragraphs are synchronized to the corpus weight self-learning module, which serves as the supervision signal input for the dynamic adjustment of corpus priority in the next cycle, forming a complete closed loop of assessment-feedback-learning-reassessment.