Enterprise operation demand data linkage analysis method and system

Abstract

The invention relates to the technical field of enterprise operation management, in particular to an enterprise operation demand data linkage analysis method and system, and the method comprises the steps: obtaining a communication text, and carrying out the analysis to obtain information definition degree data; based on the keyword weight table, obtaining an emotion recognition result; obtaining project objective state data, and comparing the project objective state data with the corresponding emotion recognition result to obtain inconsistency degree data; calculating a channel reputation score based on the information definition degree data and the inconsistency degree data; outputting a project risk result according to the channel reputation score; and when the channel reputation score is lower than a preset first threshold value and it is detected that the information definition degree data is higher than a preset second threshold value, triggering a high-priority risk alarm, and outputting a project risk result. The method has the advantages that the problem of inaccurate project risk identification caused by information distortion and data sparsity in the prior art can be effectively solved, and the reliability and timeliness of risk judgment are remarkably improved.

Description

Technical Field

[0001] This application relates to the technical field of enterprise operation management, and more specifically, to a method and system for linking and analyzing enterprise operation demand data. Background Technology

[0002] In modern enterprise operations, especially in the field of large-scale software integration project management, timely and accurate identification of potential project delays and operational risks is crucial for ensuring project delivery certainty and maintaining corporate competitiveness. Traditional project risk management methods primarily rely on objective data recorded in project management systems (PMS), such as task completion status, milestone progress, and official delay markers. However, this objective data often fails to comprehensively and accurately reflect the actual problems within the project. Specifically, when project teams face performance evaluation pressure, human factors can distort information in internal communication channels. Project members or managers may strategically adjust their communication patterns and language on internal communication platforms to avoid negative evaluations; for example, they may avoid using directly negative terms and instead use vague, neutral, or positive expressions to mask actual difficulties or problems. This deliberate manipulation of information flow makes it difficult for text sentiment analysis-based risk identification systems to accurately capture true emotional tendencies and potential risks.

[0003] Given this combination of deep distortion and information scarcity, when a sudden, genuine piece of negative technical information finally emerges through official channels, existing linkage analysis methods often fail to correctly identify it as a key risk signal. Instead, because this information sharply contradicts long-term contaminated historical data patterns that exhibit neutral or positive trends, it may be mistakenly treated by the system as an isolated statistical outlier, thus being isolated, downweighted, or even ignored. This severely weakens the ability of linkage analysis systems to make accurate and reliable project risk assessments in an environment where the credibility of data sources is dynamically declining.

[0004] To address the aforementioned issues, existing technologies urgently need improvement. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this application provides a method and system for linking enterprise operation demand data for analysis. This method effectively solves the problem of inaccurate project risk identification caused by information distortion and data sparsity in existing technologies, and significantly improves the reliability and timeliness of risk assessment.

[0006] This application provides a method for linking and analyzing enterprise operational needs data, including: Obtain communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data; Based on a keyword weight table that includes negative emotion words, positive emotion words, and negative words, the communication text is matched word by word and the weights are accumulated to obtain the emotion recognition results. Obtain objective status data of the project, compare the objective status data of the project with the corresponding sentiment recognition results, and obtain the inconsistency data; Based on data on the clarity and inconsistency of information, calculate the channel reputation score for internal communication channels; Based on the channel reputation score, the sentiment recognition results are weighted and combined with the objective status data of the project to perform a comprehensive risk assessment and output the project risk results; When the channel reputation score is lower than the preset first threshold and the detected information clarity data is higher than the preset second threshold, a high-priority risk alert is triggered, and the project risk result is output.

[0007] The above solution can effectively solve the problem of inaccurate project risk identification caused by information distortion, data sparsity or human manipulation in existing technologies, and significantly improve the reliability and timeliness of risk assessment.

[0008] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, the channel reputation score of internal communication channels is calculated, including: When calculating channel reputation scores, monitor the planned completion dates of each task plan in the project management system. When the planned completion date changes, record a change log including the task identifier, the original planned completion date, the new planned completion date, and the time of the change. Based on the change log, calculate the instability score of the task plan changes. The instability score is obtained by accumulating the adjustment range and frequency of the planned completion date according to the preset adjustment weights. The instability score is compared with the preset plan instability threshold. When the instability score reaches the plan instability threshold, the channel reputation score is reduced according to the preset adjustment rules.

[0009] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, pre-defined language specificity scoring rules are established, including: Determine the type of the sender of the communication text. The sender type includes automated system entities and human entities. When the sending subject type is a human subject, determine whether the sender of the communication text belongs to the pre-marked core technical role, and detect whether the communication text contains preset architectural risk concept terms; For communication texts where the sender is a core technical figure and contains architectural risk concepts, a scoring gain process is applied to increase the information clarity score of the communication text according to a preset first gain coefficient. When the sending subject is an automated system subject, the communication text is subjected to score attenuation processing, and the information clarity score of the communication text is reduced according to the preset first attenuation coefficient. The weighted corrected information clarity data is obtained by summarizing the information clarity scores of each communication text after score attenuation or score gain processing.

[0010] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, after obtaining the inconsistency data, it also includes: Obtain the update time and update content of the objective status data of the project; Determine whether the objective status data of the project has been updated beyond a preset time threshold, or whether the updated content only involves the progress of non-critical tasks; Analyze communication texts to identify whether they contain internal problem indications pointing to technical obstacles, resource conflicts, or process defects within the project, and determine whether the nature of the internal problem indications is internal factors, external factors, or constructive discussion; When it is determined that the objective status data of the project has an update time exceeding the preset time threshold or the updated content only involves the progress of non-critical tasks, and the nature of the internal problem indication is an internal factor, the inconsistency data is increased according to the preset update weight rules. When it is determined that the update time of the objective status data of the project does not exceed the preset time threshold and the updated content involves the progress of key tasks, or the nature of the internal problem indication is an external factor or constructive discussion, the inconsistency data is reduced according to the preset update weight rules.

[0011] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, the following steps are also included before performing a comprehensive risk assessment: Extract communication texts from internal communication channels that are sent by core technical roles and contain preset core technical issue indicators. Count the number of times the corresponding core technical issue indicators are mentioned within a preset time window, and retrieve objective project status data to determine whether there is a corresponding solution. The degree to which the core technical problem remains unresolved is calculated based on the number of mentions and the number of missing solutions. For key tasks allocated to core technology resources, monitor the occurrence of micro-progress events or deliverables associated with the key tasks and calculate the degree of mismatch in resource progress. When the degree of unresolved issues or mismatch exceeds the preset judgment threshold, the risk weight in the comprehensive risk judgment is increased according to the preset deep risk gain rule.

[0012] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, the degree to which core technical issues remain unresolved is calculated, including: Extract technical concepts and system components from communication texts in which the core technical issues are first mentioned, and generate fingerprints of the core technical issues; For subsequent communication texts within a preset time window, generate communication text fingerprints and compare the similarity between the communication text fingerprints and the fingerprints of core technical issues. When the similarity reaches a preset fingerprint threshold, associate the communication text with the core technical issue and update the discussion scope of the core technical issue. Based on the updated scope of discussion of core technical issues, the number of times core technical issues are associated within a preset time window is counted. When a solution to a core technical problem is detected, the scope of the solution is extracted and compared with the scope of discussion of the core technical problem to obtain the solution coverage. When the scope of the solution does not cover the discussion of the core technical issues, the core technical issues will be marked as partially resolved and will remain unresolved. The degree to which core technical issues remain unresolved is calculated based on the number of associations and the coverage of solutions.

[0013] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, the degree of mismatch in resource progress is calculated, including: For each key task assigned to core technology resources, a list of micro-progress events or deliverables associated with the key task is pre-defined. The list of micro-progress events or deliverables includes code submissions, automated test pass records, interface integration completion records, and / or design document finalization. Within the preset monitoring period, the occurrence of micro-progress events or products is continuously monitored, the occurrence time and quantity of each event or product are recorded, and the actual progress score of the key task is calculated according to the preset event weight. Obtain the progress status of key tasks reported in the project management system, including task completion percentage or stage markers; The actual progress score is compared with the progress status, and the difference between the two is calculated. The difference values ​​are normalized to a range of 0 to 1 to obtain the degree of mismatch in resource progress.

[0014] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, before determining the degree of mismatch in resource progress, the following steps are also included: Obtain the resource allocation ratio for critical tasks; Within the preset monitoring period, the actual working time of resources related to key tasks on internal communication channels is counted, and the actual resource availability ratio is calculated. The actual resource availability ratio is compared with the resource allocation ratio to obtain the availability deviation value; The availability deviation value is normalized to a range of 0 to 1 according to the preset mapping rules, and then weighted and synthesized with the normalized difference value to calculate the degree of mismatch in resource progress.

[0015] Furthermore, based on the aforementioned data linkage analysis method for enterprise operational needs, before calculating the actual working hours of resources related to key tasks in internal communication channels, the following is also included: The system performs keyword and phrase matching on communication texts in internal communication channels and related to resource constraints or inter-task dependencies to obtain preliminary matching results. Perform contextual analysis on the communication texts corresponding to the initial matching results to identify the sender, discussion topic, and relevance to the task or resource; Based on the preliminary matching results and context analysis, the confidence score of the communication text is calculated; When the confidence score is higher than the preset confidence threshold, the communication text is confirmed as a task obstruction indication message, and the corresponding obstruction time is deducted from the resource's working time to update the actual working time of the resource within the preset monitoring period.

[0016] Furthermore, this application also discloses an enterprise operation needs data linkage analysis system, including: The text acquisition and parsing module is used to acquire communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data. The sentiment recognition module is used to match and accumulate weights word by word in the communication text based on a keyword weight table containing negative sentiment words, positive sentiment words, and negative words, in order to obtain the sentiment recognition result. The status acquisition and comparison module is used to acquire objective status data of the project, compare the objective status data of the project with the corresponding emotion recognition results, and obtain the inconsistency level data. The reputation calculation module is used to calculate the channel reputation score of internal communication channels based on information clarity data and inconsistency data; The risk calculation module is used to weight the sentiment recognition results based on the channel reputation score, and combine them with the objective status data of the project to perform a comprehensive risk assessment and output the project risk results. The alarm triggering module is used to trigger a high-priority risk alarm and output the project risk results when the channel reputation score is lower than a preset first threshold and the detected information clarity data is higher than a preset second threshold.

[0017] The above solution provides a system implementation plan that can effectively support the above methods, realize automated linkage analysis and risk warning of enterprise operation demand data, and improve operational efficiency and risk management capabilities.

[0018] In summary, the enterprise operation demand data linkage analysis method and system provided in this application effectively solves the problem of inaccurate project risk identification caused by information distortion, data sparsity, or human manipulation in existing technologies by combining the clarity of information in internal communication texts, sentiment recognition results, and objective project status data, and by introducing channel reputation scores for weighted processing. This significantly improves the reliability and timeliness of risk assessment, thus possessing the advantage of effectively solving the problem of inaccurate project risk identification caused by information distortion, data sparsity, or human manipulation in existing technologies, and significantly improving the reliability and timeliness of risk assessment. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a data linkage analysis method for enterprise operation needs provided in this application.

[0020] Figure 2 A flowchart of a data linkage analysis system for enterprise operation needs provided in this application.

[0021] In the diagram: 1. Text acquisition and parsing module; 2. Emotion recognition module; 3. Status acquisition and comparison module; 4. Reputation calculation module; 5. Risk calculation module; 6. Alarm triggering module. Detailed Implementation

[0022] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0023] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0024] Reference Figure 1 This application proposes a method for linking and analyzing enterprise operation needs data, including: S1000: Obtain communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data; S2000: Based on a keyword weight table containing negative emotion words, positive emotion words, and negative words, the communication text is matched word by word and weights are accumulated to obtain the emotion recognition result; S3000: Acquire objective state data of the project, compare the objective state data of the project with the corresponding emotion recognition results, and obtain inconsistency data; S4000: Calculates the channel reputation score for internal communication channels based on data on the clarity and inconsistency of information. S5000: Based on channel reputation scores, the sentiment recognition results are weighted and combined with objective project status data to perform a comprehensive risk assessment and output project risk results; S6000: When the channel reputation score is lower than the preset first threshold and the detected information clarity data is higher than the preset second threshold, a high-priority risk alarm is triggered, and the project risk result is output.

[0025] The pre-defined language specificity scoring rules refer to a set of standards used to quantify the clarity and directness of communication text content. These rules can be implemented using methods such as lexical density analysis, syntactic complexity assessment, technical terminology identification, or matching of specific risk indicator words. For example, they can be used to calculate the ratio of nouns, verbs, and non-descriptive words in the text, or to identify the presence of vague expressions or euphemisms. The primary purpose is to identify whether there is information ambiguity or avoidance in the communication text, thus providing a basis for information clarity data. Information clarity data refers to the quantitative assessment results of the clarity and directness of the communication text content. This can be represented using numerical scoring, rating scales, or Boolean notation. For example, text scores can be normalized to a range of 0 to 1, or divided into levels such as "clear," "moderate," and "vague." The main purpose is to reflect the specificity and operability of the information conveyed by the communication text, avoiding the obscuring of real issues due to linguistic ambiguity. The keyword weight table is a predefined set of words containing negative, positive, and negative sentiment words, each assigned a corresponding weight value. It can be constructed using methods such as manual annotation, machine learning training, or domain expert experience. For example, "delayed" might be assigned a negative weight, "successfully" a positive weight, and "not" a negative weight. Its primary purpose is to analyze the sentiment of communication text and capture the underlying emotional signals. The inconsistency data refers to a quantitative indicator of the difference between objective project status data and sentiment recognition results. This can be obtained through methods such as difference value calculation, deviation percentage, or matching score. For example, it calculates the deviation between project progress and team sentiment. Its main purpose is to reveal potential contradictions between objective facts and subjective feelings, thereby identifying potential risk signals. The channel reputation score is a comprehensive evaluation indicator of the reliability and authenticity of information from internal communication channels. It can be obtained through a weighted calculation based on historical data accuracy, clarity, inconsistency, and other auxiliary indicators (such as the score for instability due to changes in task plans). For example, it can be achieved by cumulatively evaluating the clarity of communication texts and their inconsistency with objective realities over a period of time. Its main purpose is to dynamically assess the trustworthiness of communication channels and provide a basis for risk assessment. The first threshold is a preset critical value used to determine whether the channel reputation score is in a low-trust state. It can be determined through empirical setting, historical data analysis, or risk tolerance assessment. For example, a channel reputation score below 0.3 is set as a low-trust state, primarily to identify that the reliability of the communication channel has fallen to a level requiring special attention.The second threshold is a preset critical value used to determine whether the information clarity data is in a high-definition state. It can be determined by means of experience setting, historical data analysis or information importance assessment. For example, information clarity data above 0.8 is set as a high-definition state. This is mainly to identify that even in low-reputation channels, there is still clear and important risk information.

[0026] This solution introduces a data-driven analysis mechanism to address the problem of risk signals being masked due to information distortion and communication shifts. The system first extracts internal communication text and quantifies the clarity of information based on language specificity scoring rules to identify ambiguous or evasive expressions. Then, it performs sentiment analysis on the text based on a keyword weight table, generating sentiment recognition results to supplement the understanding of the project status from a subjective perspective. The system then acquires objective project status data and compares it with the sentiment recognition results to obtain inconsistency data, revealing potential contradictions between subjective and objective perspectives. Based on this, the system calculates the channel reputation score and weights the sentiment recognition results accordingly, suppressing positive sentiment bias when reputation is low and improving risk perception capabilities. Finally, the system uses the weighted sentiment recognition results and objective status data together for project risk assessment, ensuring a comprehensive and reliable evaluation. If the channel reputation score falls below a first threshold and the information clarity exceeds a second threshold, the system will trigger a high-level risk alert, ensuring timely identification of key risks even when data quality deteriorates.

[0027] In one specific implementation of this embodiment: This solution describes a data linkage analysis system for enterprise operational needs deployed on a cloud server. The system continuously acquires communication text from internal instant messaging tools (such as WeChat Work and DingTalk) and objective data such as task progress and milestone status from project management systems (such as Jira and Confluence) via API interfaces. First, the text parsing module calls language specificity scoring rules based on natural language processing to analyze noun and verb density and vague words, generating an information clarity score from 0 to 100 for the communication text. For example, "Interface integration is blocked, expected to require an additional three days to resolve" might score 90, while "progressing smoothly" might only score 30. The sentiment recognition module then segments the text based on a keyword weight table containing thousands of words and calculates sentiment scores, such as -2 for "problem," +1 for "success," and -1 for "no," outputting a sentiment score range of -5 to +5. The status acquisition and comparison module periodically acquires objective status data such as task completion rate and delay markings, and compares it with sentiment recognition results from the same period to calculate the degree of inconsistency. For example, if the task completion rate is 90%, but the communication contains strong negative emotions, it is marked as inconsistent. Subsequently, the reputation calculation module uses a weighted formula to calculate the reputation score of the communication channel, such as: Channel Reputation = (A × Average Message Clarity) - (B × Cumulative Inconsistency) - (C × Text Quantity Reduction Rate), to adjust the sentiment recognition weights. The risk calculation module combines the reputation score and project status to determine the risk level. If the reputation score falls below the threshold and highly explicit negative text suddenly appears, such as "Core Module A Database Connection Failed," the alarm triggering module will immediately issue a high-priority risk alarm, notifying relevant personnel via email, SMS, etc., to ensure that sudden risks are responded to in a timely manner.

[0028] In another embodiment of this application, S4000 is further proposed: calculating the channel reputation score of the internal communication channel includes: S4100: When calculating channel reputation scores, monitor the planned completion dates of each task plan in the project management system. When the planned completion date changes, record a change log including the task identifier, the original planned completion date, the new planned completion date, and the time of the change. S4200: Based on the change log, calculate the instability score of the task plan changes. The instability score is obtained by accumulating the adjustment range and frequency of the planned completion date according to the preset adjustment weights. S4300: Compares the instability score with the preset plan instability threshold. When the instability score reaches the plan instability threshold, the channel reputation score is reduced according to the preset adjustment rules.

[0029] The project management system refers to a software platform used for planning, executing, monitoring, and controlling project progress, which can be implemented using tools such as Jira, Asana, and Microsoft Project. The change log refers to data entries automatically generated by the system or manually recorded when the planned completion date of a task changes. It may include the task's unique identifier, the planned completion date before the change, the planned completion date after the change, and the timestamp of the change. Its purpose is to provide a data foundation for subsequent quantitative analysis.

[0030] The instability score quantifies the impact of task plan changes on project stability. It reflects the degree of volatility in the task plan and aims to provide a comprehensive indicator for assessing its reliability. The adjustment range of the planned completion date refers to the difference in time span between the new and original dates each time the planned completion date changes. It can be measured in days, hours, or percentages and aims to reflect the impact of a single change on project progress. The adjustment frequency refers to the number of times a planned completion date changes within a specific time period. It can be counted as the number of changes per unit time and aims to reflect the stability or uncertainty of the task plan. The preset adjustment weight refers to the relative importance coefficients assigned to the adjustment range and frequency of the planned completion date when calculating the instability score. It can be set based on project type, risk appetite, or historical data analysis results, allowing the system to flexibly adjust the influence of different changing factors according to actual conditions. The plan instability threshold is a critical value used to determine whether changes in the task plan reach the threshold that triggers adjustments to the channel reputation score. It can be a preset value and aims to distinguish between normal fluctuations and abnormal changes requiring attention. Preset adjustment rules refer to the logic used to determine the specific method and extent of channel reputation score reduction when the instability score reaches the planned instability threshold. These rules can include fixed deductions, proportional deductions, or deductions based on a functional relationship of the instability score. The purpose is to ensure that the adjustment of channel reputation score is systematic and predictable.

[0031] This solution improves the accuracy of channel reputation score calculation by introducing a quantitative evaluation mechanism for task plan changes in the project management system. When calculating channel reputation scores, the system continuously monitors the planned completion date of tasks. Once a change occurs, a change log is immediately recorded, including the task identifier, original planned date, new planned date, and change time, ensuring that each adjustment is fully tracked. Subsequently, the system calculates an instability score based on the logs. This score comprehensively considers the magnitude and frequency of adjustments and is accumulated according to preset weights. For example, frequent small adjustments or a single large delay will significantly increase the instability score. When this score exceeds a preset threshold, the system determines that project stability is affected and lowers the channel reputation score of internal communication channels according to rules. This mechanism directly maps fluctuations at the project execution level to communication reputation assessment, serving as a powerful supplement to traditional methods for calculating the clarity and inconsistency of dependent information. Even if the communication text appears positive, if task plans are frequently adjusted, the system can identify potential risks in advance through changes in reputation scores, making project risk assessment more comprehensive and accurate.

[0032] In one specific implementation of this embodiment: Suppose a software development project is underway, and its project management system could use Jira. When calculating the channel reputation score for internal communication channels, the system can be configured with a backend service that continuously monitors the planned completion date of each task (e.g., developing a module, completing a test case) via Jira's API. Specifically, when the planned completion date of a task, such as the "User Interface Design" task, changes from the original planned "October 20, 2023" to "October 25, 2023," the system immediately generates and records a change log. This log may contain the task's unique identifier (e.g., Jira Task ID: UI-DESIGN-001), the original planned completion date of October 20, 2023, the new planned completion date of October 25, 2023, and the timestamp of the change (e.g., October 18, 2023, 14:30). The system then uses these change logs to calculate an instability score for the task's planned change. For example, for the "User Interface Design" task, the planned completion date adjustment might be 5 days. If the planned completion date of the task has changed twice in the past week, the adjustment frequency is 3 times. The system can preset adjustment weights, for example, setting the adjustment magnitude weight to 0.7 and the adjustment frequency weight to 0.3. The instability score can be calculated as (adjustment magnitude × 0.7) + (adjustment frequency × 0.3). The system then compares the calculated instability score with a preset instability threshold. For example, if the preset instability threshold is 50 points, when the instability score for the "User Interface Design" task reaches 55 points, it indicates that the changes have exceeded the acceptable range. At this point, the system will reduce the channel reputation score of the internal communication channel according to preset adjustment rules. The preset adjustment rules can stipulate that when the instability score exceeds the threshold, the channel reputation score is reduced by 5 points for every 10 points exceeding the threshold. In this way, by quantifying changes in the task plan, the system can dynamically adjust the channel reputation score, making it more accurately reflect the true state of project execution.

[0033] In another embodiment of this application, a preset language specificity scoring rule is further proposed, including: A1: Determine the type of the sender of the communication text. The sender type includes automated system entities and human entities. A2: When the sending subject type is a human subject, determine whether the sender of the communication text belongs to the pre-marked core technical role, and detect whether the communication text contains preset architectural risk concept terms; A3: For communication texts where the sender is a core technical figure and contains architectural risk concepts, a scoring gain process is applied to increase the information clarity score of the communication text according to a preset first gain coefficient. A4: When the sending subject is an automated system subject, the score attenuation processing is applied to the communication text, and the information clarity score of the communication text is reduced according to the preset first attenuation coefficient. A5: The weighted corrected information clarity data is obtained by summarizing the information clarity scores of each communication text after score attenuation or score gain processing.

[0034] Among them, the pre-marked core technical roles refer to technical experts, architects, senior developers, etc., who are clearly identified and given special permissions or labels within an enterprise or project organization. They typically possess key technical details, system architecture, or core business logic, and their communication content has significant reference value for project risk assessment. The pre-defined architectural risk vocabulary refers to a set of professional terms or phrases related to potential technical risks such as system architecture, technical implementation, compatibility, performance bottlenecks, and security vulnerabilities, such as "interface incompatibility," "performance bottleneck," "data consistency issues," "system crash," and "security vulnerabilities." Scoring gain processing refers to the operation of increasing the clarity score of specific communication texts, which can be achieved by multiplying by a gain coefficient greater than 1. Scoring attenuation refers to the operation of lowering the clarity score of a specific communication text, which can be achieved by multiplying by an attenuation coefficient less than 1. The first gain coefficient is a preset value used to increase the clarity score of a specific communication text. This value is usually greater than 1, and its purpose is to quantify the magnitude of the gain processing. The first attenuation coefficient is a preset value used to decrease the clarity score of a specific communication text. This value is usually less than 1 and greater than 0. Weighted corrected clarity data refers to the data obtained by summarizing the clarity scores of communication texts from different sources or types after gaining or attenuating them. This data can more accurately reflect the true clarity of information in the communication channel, and its purpose is to provide a more reliable basis for risk assessment.

[0035] This application's solution, by refining the scoring rules for language specificity, can more accurately assess the clarity of information in communication texts, thereby improving the accuracy of risk assessment. Specifically, the solution first determines the type of sender of the communication text, distinguishing between automated system entities and human entities. This distinction is based on the differences in the roles and importance of different entities in information transmission. For example, automated systems typically send standardized, procedural notifications with relatively fixed information value; while human entities, especially those in core technical roles, may convey deeper technical insights or potential risks in their communication content.

[0036] In some preferred embodiments, the preset language specificity scoring rules can be implemented as follows: This solution integrates an enterprise user management system with an automation tool management platform to identify the type of sender of communication texts, thereby finely correcting the information clarity score. The system first determines whether the sender is a human subject (such as an employee account in the enterprise LDAP or Active Directory) or an automated system subject (such as a script or robot account). If it is a human subject, the system further determines whether it belongs to a core technical role, such as a user marked as a "technical expert" or "architect," and uses natural language processing technology to identify whether the communication text contains preset architectural risk concepts, such as "deadlock," "system crash," and "interface protocol change." If both conditions are met, the system applies a gain coefficient (e.g., 1.5) to the original information clarity score S, correcting it to S × 1.5. If the sender is an automated system, the system uses a decay coefficient (e.g., 0.8) to correct the score S′, i.e., S′ × 0.8. Finally, the system summarizes all scores after gain or decay processing, using methods such as average or weighted average, to generate an overall information clarity value for the communication channel, improving the accuracy and discriminative power of information assessment.

[0037] In another embodiment of this application, it is further proposed that, after S3000, the following is also included: S3100: Obtain the update time and update content of objective status data for the project; S3200: Determine whether the objective status data of the project has an update time exceeding the preset time threshold, or whether the updated content only involves the progress of non-critical tasks; S3300: Analyze communication texts to identify whether they contain indications of internal issues pointing to technical obstacles, resource conflicts, or process defects within the project, and determine whether the nature of the internal issue indications is internal factors, external factors, or constructive discussion; S3400: When it is determined that the objective status data of the project has an update time exceeding the preset time threshold or the updated content only involves the progress of non-critical tasks, and the nature of the internal problem indication is an internal factor, the inconsistency data is increased according to the preset update weight rules. S3500: When it is determined that the update time of the objective status data of the project does not exceed the preset time threshold and the updated content involves the progress of key tasks, or the nature of the internal problem indication is external factors or constructive discussion, the inconsistency data is reduced according to the preset update weight rules.

[0038] Among them, the preset time threshold refers to the upper limit of time used to judge whether the objective status data of the project is updated in a timely manner. It can be a fixed duration, such as 24 hours, a week, or an iteration cycle, or it can be dynamically adjusted according to the project type or stage; non-critical task progress refers to task updates that have little impact on the overall project progress or core deliverables, which can include routine administrative tasks, minor optimizations of non-core modules, or detailed supplements to completed tasks; internal problem indications refer to problems that exist within the project and can be directly controlled or resolved by the project team, which can include technical obstacles, resource conflicts, or process defects; internal factors refer to the root causes of project problems that lie within the project team itself, technical Factors related to the technology stack, resource allocation, or internal management processes can include insufficient technical personnel capabilities, development tool defects, or poor internal communication; external factors refer to factors whose root causes of project problems are located outside the project and beyond the direct control of the project team, such as third-party interface failures, changes in customer requirements, or adjustments in external policies; constructive discussions refer to proactive communication aimed at analyzing problems, proposing solutions, or optimizing workflows, such as technical solution discussions, risk avoidance strategy formulation, or experience sharing; preset update weight rules refer to the specific calculation methods or logic used to adjust inconsistent data, which can adopt linear weighting, exponential weighting, or dynamic adjustment strategies based on machine learning models.

[0039] This solution achieves dynamic correction of inconsistencies in project status data and improves the accuracy of risk assessment by deeply analyzing the update status of objective project status data and the indicative nature of internal issues in communication documents. The system first obtains the update time and content of project status data, assessing its timeliness and criticality. For example, if data has not been updated for a long time or only involves non-critical tasks, it indicates low reflective value. Subsequently, the system analyzes communication documents to identify internal issues such as technical obstacles and resource conflicts, and determines their nature, classifying them as internally controllable, externally uncontrollable, or constructive discussions. When status data updates are delayed and the issue is a controllable factor, the system increases the inconsistency level according to preset weights to strengthen risk perception; conversely, if data updates are timely or the issue is an external factor, the system decreases the inconsistency level to avoid misjudgment. By introducing a judgment mechanism based on data update quality and issue nature, the corrected inconsistency data more closely reflects the actual risk level, providing a reliable foundation for subsequent channel reputation weighting and comprehensive risk assessment.

[0040] In one specific implementation of this embodiment: This solution, based on the inconsistency data generated by comparing the objective status data of the project with the sentiment recognition results, further introduces dynamic analysis of data update status and problem nature to improve the accuracy of risk assessment. The system first obtains the update time and content of the objective status data of the project, such as "completed the first draft of the requirements analysis document" or "fixed known bugs." If a task's data has not been updated for more than a preset time threshold (e.g., 48 hours), or only involves non-critical task content (e.g., "meeting room reservation completed"), it is marked as having insufficient data quality. Next, the system analyzes internal communication texts to identify whether they contain indications of internal problems such as technical obstacles or resource conflicts, such as "database connection timeout" or "front-end framework compatibility issues." The system further determines the nature of these problems: if they are "developers are unfamiliar with the new framework" or "insufficient server resources," they are marked as internal factors; if they are "third-party service interruptions" or "customer data errors," they are marked as external factors; if the text discusses solutions, it is considered a constructive discussion. If the objective data is delayed or non-critically updated, and the problem is an internal factor, such as "insufficient test environment resources," the system increases the inconsistency data according to rules, such as increasing it by 10%. Conversely, if the status data is updated promptly and involves core progress, or if the communication issue is related to external factors or constructive content, such as "interface maintenance" or "deployment plan optimization," then the inconsistency level data is lowered, for example, by 5%. By combining the timeliness of the data with the nature of the issue, this approach makes the inconsistency level data closer to the actual risk status, providing a more reliable basis for subsequent risk assessments.

[0041] Another embodiment of this application further proposes that, before performing a comprehensive risk assessment, the method includes: B1: Extract communication texts from internal communication channels that are sent by core technical roles and contain preset core technical issue indicators. Within a preset time window, count the number of times the corresponding core technical issue indicators are mentioned and retrieve objective project status data to determine whether there is a corresponding solution. B2: Calculate the degree of continued unresolved nature of core technical issues based on the number of mentions and the number of missing solutions; B3: For key tasks allocated to core technology resources, monitor the occurrence of micro-progress events or deliverables associated with the key tasks and calculate the degree of mismatch in resource progress. B4: When the degree of unresolved issues or mismatch exceeds the preset judgment threshold, the risk weight in the comprehensive risk judgment is increased according to the preset deep risk gain rule.

[0042] The preset core technical issue indicator words refer to a set of predefined keywords or phrases that directly or indirectly indicate technical obstacles, defects, or risks in the project. These can be obtained through methods such as summarizing expert experience, analyzing historical failure reports, or extracting information using natural language processing techniques. The preset time window refers to a specific period used to count the number of times the core technical issue indicator words are mentioned and to monitor the occurrence of micro-progress events or deliverables. This can be set as a fixed duration or dynamically adjusted duration. The degree of continued unresolved issues refers to an indicator measuring the frequency of mentions of core technical issues within the preset time window and the severity of the lack of effective solutions. This can be calculated using a weighted combination of the number of mentions and the number of times solutions are missing. Micro-progress events or deliverables refer to granular, quantifiable events or deliverables that directly reflect the actual progress of a task during its execution. These can be manifested as code submissions, passed test cases, finalized design documents, or successful interface integration. The degree of mismatch in resource progress refers to an indicator measuring the difference between the actual progress and reported progress of key tasks allocated to core technical resources. This can be calculated using the difference between the actual progress score and the reported progress status. The preset judgment threshold refers to the critical value used to determine whether the degree of persistent unresolved issues or mismatches has reached the point where risk weights need to be increased. It can be determined through expert experience, historical data statistical analysis, or machine learning model training. The preset deep risk gain rule refers to the rule used to adjust the risk weights in the comprehensive risk assessment when the degree of persistent unresolved issues or mismatches exceeds the preset judgment threshold. It can be implemented using linear gain, exponential gain, or a step-by-step gain based on risk level.

[0043] In one specific implementation of this embodiment: Before conducting a comprehensive risk assessment, this solution identifies hidden critical technical risks through a deep risk assessment process. Taking a software development project as an example, even though the project management system shows smooth progress, the project manager avoids using words like "problem" or "delay" in communication, exhibiting vague language. In this case, the system extracts text sent by core technical roles (such as "architect" or "senior developer") from internal communication platforms like WeChat or DingTalk, scanning for pre-defined key technical issue indicators such as "memory leak," "deadlock," and "interface incompatibility." Within a set time window (e.g., the past 7 days), the system counts the frequency of these problematic terms and searches the project management system or code repository to determine if relevant solutions exist, such as defect fix records or task closures. If "memory leak" is mentioned 5 times without corresponding fix records, the level of unresolved issues is considered high. Simultaneously, the system monitors micro-level progress events of critical tasks, such as code commits, test pass rates, and document finalization. If a task is marked as 80% complete in the project management system but has no code commits or test records in the past week, the system will consider its resource progress to be significantly mismatched. Ultimately, when the degree of unresolved issues or mismatch between resource progress and the threshold (e.g., 0.7) is exceeded, the system will increase the overall risk level according to preset rules (e.g., risk weighting is increased by 20%). Even if the overall sentiment is positive, it can accurately identify technical bottlenecks and progress distortions, ensuring that key risks are exposed and responded to in a timely manner.

[0044] In another embodiment of this application, the sub-step of B2, calculating the degree of continued unresolved nature of the core technical problem, includes: B21: Extract technical concepts and system components from communication texts where core technical issues are first mentioned, and generate fingerprints of core technical issues; B22: Generate communication text fingerprints for subsequent communication texts within a preset time window, and compare the similarity between the communication text fingerprints and the core technical issue fingerprints. When the similarity reaches a preset fingerprint threshold, associate the communication text with the core technical issue and update the discussion scope of the core technical issue. B23: Based on the updated scope of discussion on core technical issues, count the number of times core technical issues are associated within a preset time window; B24: When a solution to a core technical problem is detected, the scope of the solution is extracted and compared with the scope of discussion of the core technical problem to obtain the solution coverage. B25: When the scope of the solution does not cover the discussion of core technical issues, mark the core technical issues as partially resolved and keep them unresolved. B25: Calculate the degree to which core technical issues remain unresolved based on the number of associations and the coverage of solutions.

[0045] The core technical issue fingerprint refers to the identifier extracted from the communication text in which the core technical issue is first mentioned. This identifier represents the core content and characteristics of the issue and can be implemented using keyword vectors, semantic vectors generated by topic models, or deep learning-based text embeddings. Its purpose is to establish a quantifiable digital representation of the core technical issue for subsequent comparison and tracking. The preset time window refers to the time interval used to limit the analysis of communication texts and the association with the issue. This can be a duration, such as 7 days, 30 days, or a project iteration cycle. Its purpose is to ensure that the continuous evaluation of the core technical issue is conducted within a relevant and manageable timeframe. The communication text fingerprint refers to the digital representation generated after processing subsequent communication texts within the preset time window. This fingerprint can use the same or compatible generation methods as the core technical issue fingerprint, such as keyword vectors or text embeddings. Its purpose is to transform subsequent communication texts into a form that can be quantitatively compared with the core technical issue fingerprint. A preset fingerprint threshold is a critical value used to judge the similarity between the fingerprint of a communication text and the fingerprint of the core technical issue. It can be a floating-point number between 0 and 1, such as 0.7 or 0.8. Its purpose is to set a standard to determine whether subsequent communication texts are sufficiently relevant to the core technical issue, thus classifying them as discussions of that issue. The scope of core technical issue discussion refers to the dynamic set of technical concepts, system components, impact boundaries, and related sub-issues involved in the core technical issue over time and with the relevance of communication texts. It can be a list containing keywords, conceptual entities, or semantic topics, reflecting the breadth and depth of understanding and discussion of the core technical issue within the project team. The solution coverage refers to the set of technical concepts, system components, or subsets of issues that a proposed solution to the core technical issue can address or affect. It can be a list or set with a structure similar to the scope of core technical issue discussion, quantifying the actual boundaries of the solution's impact. Solution coverage is an assessment of the degree to which the solution addresses the core technical issue by comparing the solution coverage with the scope of core technical issue discussion. It can be a percentage value, a Boolean symbol, or a multi-level classification.

[0046] This solution improves the accuracy of assessing the continued unresolved nature of problems by introducing a refined quantitative mechanism for core technical issues and their solutions. The system first extracts technical concepts and system components from the initial communication texts where the core technical issue first appears, generating a unique core technical issue fingerprint to represent the essential characteristics of the problem and support subsequent automatic tracking. Within a preset time window, the system generates communication text fingerprints for subsequent communication texts and compares them with the core technical issue fingerprint. When the similarity exceeds a threshold, the relevant text is included in the discussion scope of that issue. This mechanism ensures that even with changes in language or expression, the system can still comprehensively track the ongoing discussion of the same issue, overcoming the limitations of traditional keyword-based statistical methods. Based on the dynamically updated discussion scope, the system counts the frequency of the issue's mentions to measure its activity. Simultaneously, the system extracts the coverage of identified solutions and compares it with the issue's discussion scope to determine whether the solutions fully address the key points of the problem. If the coverage is incomplete, the system marks the issue as "partially resolved" and maintains its unresolved status to prevent incorrect judgments about the effectiveness of the solution. Finally, the system combines discussion frequency and solution coverage to calculate the continued unresolved nature of the core technical issue. This mechanism, combined with traditional assessment methods based on the number of mentions and missing solutions, improves the accuracy of risk identification. When a more accurate level of unresolved issues is used in comprehensive risk assessment, the system can promptly identify deep-seated risks that appear to have been addressed but are not actually resolved, avoiding underreporting or misjudgment and ensuring that enterprises can still conduct effective risk control and assessment in complex project environments.

[0047] In some preferred embodiments, the degree to which the core technical problem remains unsolved can be specifically implemented as follows: This solution utilizes fingerprint comparison and coverage analysis mechanisms to achieve precise quantification of the persistent unresolved nature of core technical issues. Taking "database connection pool overflow causing service interruption" as an example, the system first extracts technical concepts such as "database connection pool," "overflow," and "service interruption," along with their associated system components (e.g., "database system" and "application service"), from the initial communication text mentioning the issue. It then generates a core technical issue fingerprint based on TF-IDF vectorization or a Word2Vec model. Subsequently, within a set time window (e.g., the past 7 days), the system continuously collects new communication texts, such as "application service logs show abnormal database connection counts," and generates communication text fingerprints for each. The system determines relevance by calculating the cosine similarity between fingerprints. For example, if the similarity is 0.85 and higher than the threshold of 0.7, the text is associated with the core technical issue, and the scope of the discussion is dynamically expanded to include "abnormal connection counts."

[0048] Next, the system counts the number of times the issue is associated within the time window, for example, 15 times. Simultaneously, the system monitors for available solutions, such as "the maximum number of database connection pools has been adjusted," and extracts their coverage. If the coverage only includes "maximum connection configuration adjustment" and does not cover the core issues such as "service interruption" or "abnormal connection count," the system determines that the solution is only partially resolved, and the issue remains unresolved. Finally, the system combines the number of associations (e.g., 15) with the solution coverage (partial coverage) and calculates the degree of continued unresolved nature of the core technical issue using a preset weighted formula, quantifying the true risk status of the issue when it is frequently mentioned but not completely resolved.

[0049] In another embodiment of this application, the sub-step of B3, calculating the degree of mismatch in resource progress, includes: B31: For each key task assigned to core technology resources, a list of micro-progress events or deliverables associated with the key task is pre-defined. The list of micro-progress events or deliverables includes code submissions, automated test pass records, interface integration completion records, and / or design document finalization. B32: Within the preset monitoring period, continuously monitor the occurrence of micro-progress events or products, record the occurrence time and quantity of each event or product, and calculate the actual progress score of the key task according to the preset event weight. B33: Obtain the progress status of key tasks reported in the project management system, including task completion percentage or stage markers; B34: Compare the actual progress score with the progress status and calculate the difference between the two; B39: Normalize the difference values ​​to a range of 0 to 1 to obtain the degree of mismatch in resource progress.

[0050] The micro-level progress event or deliverable list refers to a predefined series of specific, quantifiable activities or results that directly reflect the actual progress of each critical task. These can include, but are not limited to, code submissions, automated test pass records, interface integration completion records, finalized design documents, improved unit test coverage, integration test case execution results, performance test report generation, user interface prototype confirmation, database structure change records, API documentation updates, technical review meeting minutes, defect repair records, and deployment script completion. Its purpose is to capture the true progress of the task at a fine-grained level, avoiding information bias that may arise from relying solely on macro-level reports. The pre-defined event weight refers to the numerical value assigned to each event or deliverable in the micro-level progress event or deliverable list to measure its relative importance in assessing the overall progress of the critical task. It can be determined using expert scoring, historical data analysis, or dynamic adjustments based on task type and complexity. Its purpose is to ensure that the contribution of different types or importance of progress events to the actual progress score is reasonably reflected. The preset monitoring cycle refers to the time interval at which the system continuously monitors the occurrence of micro-level progress events or products. It can be set to a daily, weekly, or dynamically adjusted cycle according to the project stage. Its purpose is to ensure that the tracking of task progress is timely and continuous, so as to capture progress changes in real time and discover potential problems in a timely manner.

[0051] In one specific implementation of this embodiment: This solution quantifies the micro-level progress events of key tasks to determine whether their actual progress matches the reported progress, thereby identifying project risks arising from resource execution deviations. Taking the "core module development" task as an example, the system first sets up a list of micro-level progress events or deliverables, including: Git code commits, Jenkins automated tests passing, API integration confirmation, and Confluence document finalization status. During the weekly monitoring cycle, the system continuously collects records of each event by integrating Git hooks, parsing Jenkins logs, identifying integration signals in emails or logs, and monitoring document status changes.

[0052] Each event is assigned a preset weight, such as code submission 0.1, test pass 0.3, interface integration 0.4, and document finalization 0.2. If 10 submissions, 2 tests, and 1 integration are recorded in a week, the system will calculate the actual progress score as: (10 × 0.1) + (2 × 0.3) + (1 × 0.4) = 2.0. Simultaneously, the system reads the task progress status reported by the project manager from a project management system (such as Jira), for example, "70% complete". To compare the two, the system uses a mapping function to convert the score of 2.0 into an actual completion rate of 60%.

[0053] The system then calculates the difference between the two as 10%, and normalizes this difference to 0.1, representing the degree of mismatch in resource progress. If this value exceeds a preset threshold, the system determines that the task report is inflated, indicating potential progress deviations or risks. This mechanism ensures that even when communication data is positive or ambiguous, the system can still assess the true status of task execution based on technical output.

[0054] In another embodiment of this application, it is further proposed that, prior to B39, the following is also included: B35: Obtain the resource allocation ratio for critical tasks; B36: Within the preset monitoring period, the actual working time of resources related to key tasks on internal communication channels is counted, and the actual resource availability ratio is calculated. B37: Compare the actual resource availability ratio with the resource allocation ratio to obtain the availability deviation value; B38: Normalize the availability deviation value to a range of 0 to 1 according to the preset mapping rules, and weight and synthesize it with the normalized difference value to calculate the degree of mismatch in resource progress.

[0055] Among these, the resource allocation ratio refers to the share of resources reserved in the project plan for specific key tasks, which can be expressed as a percentage of time invested in resources, the number of human resources, or the budget ratio. Its purpose is to clarify the resource support that the task should receive during the planning phase. The pre-set monitoring period refers to a fixed time period used to collect and analyze actual resource investment data, which can be one day, one week, or one month. Its purpose is to ensure the periodicity and consistency of data collection. Internal communication channels refer to the platforms for daily communication and collaboration among project team members, which may include internal instant messaging tools, discussion forums in project collaboration software, or email systems. Their purpose is to obtain informal but authentic information about resource activities. Actual working hours refer to the total time resources are actually invested in key tasks within the pre-set monitoring period. This can be statistically analyzed based on communication records, meeting minutes, or task logs, and its purpose is to reflect the true investment of resources. The actual resource availability ratio refers to the percentage of resources available during the pre-set monitoring period. The first concept, defining the proportion of time actually spent on critical tasks within a monitoring period relative to the total available time, aims to quantify the actual efficiency of resource allocation. The second concept, availability deviation, refers to the difference between the actual resource availability ratio and the resource allocation ratio, revealing the degree of deviation between actual and planned resource input. The third concept, pre-defined mapping rules, are mathematical functions or algorithms that transform the original data range to a target range (e.g., 0 to 1), such as linear mapping, logarithmic mapping, or the sigmoid function, aiming to unify the dimensions of different indicators for easier subsequent comprehensive calculations. The fourth concept, normalization, transforms data of different dimensions or ranges into a unified range, typically 0 to 1, aiming to eliminate the influence of dimensions and make different indicators comparable. The fifth concept, weighted synthesis, assigns different weights to different indicators based on their importance and then combines these weighted indicators for calculation, aiming to comprehensively consider the impact of multiple factors on the final result, making the evaluation more comprehensive.

[0056] In one specific implementation of this embodiment: A critical task, "Module A Development," is allocated 50% of the resources in the project management system, meaning this task should theoretically require half of the relevant resources' working time. The system first obtains this 50% resource allocation. Then, within a preset monitoring period, such as one week, the system continuously monitors communication text on internal instant messaging tools (e.g., DingTalk, WeChat Work) and project collaboration platforms (e.g., Jira, Confluence). Using natural language processing technology, the system can identify discussions related to the "Module A Development" task and calculate the actual working time of the resources involved in these discussions. For example, if a developer actively communicates in groups related to "Module A Development" for 20 hours during the week, and their total working time is 40 hours, then their actual working time for this task is 20 hours. By summing the actual working time of all relevant resources, the actual resource availability ratio can be calculated. For example, if all relevant resources have invested a total of 100 hours, while theoretically 200 hours should be invested, then the actual resource availability ratio is 50%. Next, the system compares the actual resource availability ratio (e.g., 50%) with the previously obtained resource allocation ratio (e.g., 50%) to calculate the availability deviation value. If they are equal, the deviation value is 0; if the actual availability ratio is lower than the allocation ratio, the deviation value is positive, and vice versa. For example, if the actual availability ratio is 40% and the allocation ratio is 50%, the availability deviation value is -10%. Finally, the system normalizes this availability deviation value to a range of 0 to 1 according to a preset mapping rule. For example, a linear mapping rule can be set to map the deviation value from -100% to 100% to 0 to 1. Simultaneously, the system obtains the normalized difference value previously calculated based on micro-progress events and reported progress. Assume this difference value is 0.7 (indicating that actual progress is significantly lower than reported progress). The system will then perform a weighted summation of the normalized availability deviation value (e.g., 0.2, indicating insufficient resource allocation) and this 0.7 difference value. For example, we can set the weight of resource availability deviation to 0.4 and the weight of progress difference value to 0.6. Then the final resource progress mismatch degree is 0.2 * 0.4 + 0.7 * 0.6 = 0.08 + 0.42 = 0.5. This final value of 0.5 comprehensively reflects the mismatch between the two dimensions of task progress and resource input, providing a more comprehensive basis for project risk assessment.

[0057] In another embodiment of this application, it is further proposed that B36 includes: C1: Perform keyword and phrase matching on communication texts in internal communication channels that are related to resource obstruction or inter-task dependencies to obtain preliminary matching results; C2: Perform contextual analysis on the communication text corresponding to the initial matching results to identify the sender, discussion topic, and relevance to the task or resource; C3: Calculate the confidence score of the communication text based on the preliminary matching results and context analysis; C4: When the confidence score is higher than the preset confidence threshold, the communication text is confirmed as a task obstruction indication message, and the corresponding obstruction time is deducted from the resource's working time to update the actual working time of the resource within the preset monitoring period.

[0058] Contextual analysis refers to the in-depth extraction of semantic and structured information from the communication text corresponding to the initial matching results. This can be achieved using Natural Language Processing (NLP) techniques, such as named entity recognition, dependency parsing, and topic modeling. Its purpose is to accurately locate key information related to resource obstruction or task dependence from massive amounts of communication information and verify its authenticity and validity. Confidence scoring is a quantitative assessment of whether the communication text truly reflects resource obstruction. It can be achieved by comprehensively considering multiple dimensions, including the accuracy of keyword matching, the strength of relevance identified by contextual analysis, the role weight of the sender, and the reliability of the information source, and using weighted summation or machine learning models for calculation. Its purpose is to provide an objective basis for subsequent decision-making regarding whether to deduct resource working time and avoid misjudgments.

[0059] This application's solution employs in-depth analysis of communication texts from internal communication channels to more accurately calculate the actual working time of resources within a preset monitoring period. Specifically, it first performs keyword and phrase matching on the communication texts related to resource obstruction or inter-task dependencies, quickly filtering out potential problem information and obtaining preliminary matching results. This rapid filtering of preliminary matching results allows subsequent in-depth analysis to focus on highly relevant texts, improving processing efficiency. Based on this, contextual analysis is performed on the communication texts corresponding to the preliminary matching results to identify the sender, discussion topic, and relevance to the task or resource. This contextual analysis is necessary because it compensates for the shortcomings of simple keyword matching, more accurately determining the authenticity and importance of information by comprehensively considering the source, content, and connection to specific tasks or resources. Based on the preliminary matching results and contextual analysis, the system further calculates a confidence score for the communication text, which is a quantitative assessment of the information's reliability. When the confidence score exceeds a preset confidence threshold, the system confirms the communication text as an indication of task obstruction. The introduction of confidence scoring and threshold judgment enables the system to filter out low-confidence information, ensuring that only verified obstruction information is adopted. Once obstruction is confirmed, the system deducts the corresponding obstruction time from the resource's working time to update the resource's actual working time within the preset monitoring period. In this way, the actual working time of resources can more accurately reflect the effective investment time of resources in actual projects, avoiding inflated availability due to implicit obstruction. This corrected actual working time provides more accurate data input when calculating the actual resource availability ratio, making the assessment of the mismatch between resource progress and actual availability more accurate and reliable, effectively solving the problem of assessment bias caused by ignoring obstruction information in internal communication channels in traditional methods.

[0060] Reference Figure 2 Another embodiment of this application further proposes an enterprise operation demand data linkage analysis system, including: Text acquisition and parsing module 1 is used to acquire communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data; The emotion recognition module 2 is used to match and accumulate weights word by word in the communication text based on a keyword weight table containing negative emotion words, positive emotion words and negative words, so as to obtain the emotion recognition result; The status acquisition and comparison module 3 is used to acquire objective status data of the project, compare the objective status data of the project with the corresponding emotion recognition results, and obtain the inconsistency level data. Reputation calculation module 4 is used to calculate the channel reputation score of internal communication channels based on information clarity data and inconsistency data; Risk calculation module 5 is used to weight the sentiment recognition results based on the channel reputation score, and perform a comprehensive risk judgment in combination with the objective status data of the project, and output the project risk result; Alarm triggering module 6 is used to trigger a high-priority risk alarm and output project risk results when the channel reputation score is lower than a preset first threshold and the detected information clarity data is higher than a preset second threshold.

[0061] This application's solution, through the effective collaboration of multiple functional modules, constructs a data linkage analysis system capable of dynamically adjusting to adapt to the operational needs of enterprises in environments characterized by data distortion and information sparsity. The above descriptions are merely embodiments of this application and are 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 method for linking and analyzing enterprise operational needs data, characterized in that, include: Obtain communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data; Based on a keyword weight table containing negative emotion words, positive emotion words, and negative words, the communication text is matched word by word and weights are accumulated to obtain the emotion recognition result. Obtain objective state data of the project, compare the objective state data of the project with the corresponding emotion recognition results, and obtain inconsistency data; Based on the information clarity data and the inconsistency data, calculate the channel reputation score for the internal communication channels; Based on the channel reputation score, the sentiment recognition result is weighted and combined with the objective status data of the project to perform a comprehensive risk assessment and output the project risk result; When the channel reputation score is lower than a preset first threshold and the information clarity data is detected to be higher than a preset second threshold, a high-priority risk alarm is triggered, and the project risk result is output.

2. The enterprise operation demand data linkage analysis method according to claim 1, characterized in that, The calculation of the channel reputation score for internal communication channels includes: When calculating the channel reputation score, the planned completion date of each task plan in the project management system is monitored. When the planned completion date changes, a change log is recorded, which includes the task identifier, the original planned completion date, the new planned completion date, and the time of the change. Based on the change log, an instability score for the task plan changes is calculated. The instability score is obtained by accumulating the adjustment range and frequency of the planned completion date according to a preset adjustment weight. The instability score is compared with a preset planned instability threshold. When the instability score reaches the planned instability threshold, the channel reputation score is reduced according to a preset adjustment rule.

3. The enterprise operation demand data linkage analysis method according to claim 1 or 2, characterized in that, The preset language specificity scoring rules include: Determine the type of the sender of the communication text, where the sender type includes automated system entities and human entities; When the sending entity type is a human entity, determine whether the sender of the communication text belongs to a pre-marked core technology role, and detect whether the communication text contains preset architectural risk concept terms; The communication text in which the sender is a core technical player and contains the architectural risk concept vocabulary is subjected to scoring gain processing to increase the information clarity score of the communication text according to a preset first gain coefficient; When the sending entity is an automated system entity, the communication text is subjected to score attenuation processing, and the information clarity score of the communication text is reduced according to a preset first attenuation coefficient. The weighted corrected information clarity data is obtained by summarizing the information clarity scores of each communication text after score attenuation or score gain processing.

4. The enterprise operation demand data linkage analysis method according to claim 1, characterized in that, After obtaining the inconsistency data, the process also includes: Obtain the update time and update content of the objective status data of the project; Determine whether the objective status data of the project has an update time exceeding a preset time threshold, or whether the update content only involves the progress of non-critical tasks; Analyze the communication text to identify whether it contains internal problem indications pointing to technical obstacles, resource conflicts, or process defects within the project, and determine whether the nature of the internal problem indications is internal factors, external factors, or constructive discussion; When it is determined that the objective status data of the project has an update time exceeding a preset time threshold or the updated content only involves the progress of non-critical tasks, and the nature of the internal problem indication is an internal factor, the inconsistency data is increased according to the preset update weight rules. When it is determined that the update time of the objective status data of the project does not exceed the preset time threshold and the updated content involves the progress of key tasks, or the nature of the internal problem indication is an external factor or constructive discussion, the inconsistency data is reduced according to the preset update weight rules.

5. The enterprise operation demand data linkage analysis method according to claim 1, characterized in that, Before performing the comprehensive risk assessment, the following steps are also included: Extract communication texts from internal communication channels that are sent by core technical roles and contain preset core technical issue indicators. Count the number of times the corresponding core technical issue indicators are mentioned within a preset time window, and retrieve objective project status data to determine whether there is a corresponding solution. Based on the number of mentions and the number of missing solutions, the degree to which the core technical problem remains unresolved is calculated; For key tasks allocated to core technology resources, monitor the occurrence of micro-progress events or products associated with the key tasks and calculate the degree of mismatch in resource progress. When the degree of continued unresolved issues or the degree of mismatch exceeds a preset judgment threshold, the risk weight in the comprehensive risk judgment is increased according to a preset deep risk gain rule.

6. The enterprise operation demand data linkage analysis method according to claim 5, characterized in that, The calculation of the degree to which the core technical problem remains unresolved includes: Extract technical concepts and system components from the communication texts in which the core technical issues are first mentioned, and generate a fingerprint of the core technical issues; A communication text fingerprint is generated for subsequent communication texts within a preset time window, and the communication text fingerprint is compared with the fingerprint of the core technical issue. When the similarity reaches a preset fingerprint threshold, the communication text is associated with the core technical issue and the discussion scope of the core technical issue is updated. Based on the updated scope of discussion of core technical issues, the number of times the core technical issues are associated within the preset time window is counted. When a solution to the core technical problem is detected, the scope of the solution is extracted, and the scope of the solution is compared with the scope of discussion of the core technical problem to obtain the solution coverage. When the scope of the solution does not cover the scope of the discussion of the core technical issues, the core technical issues are marked as partially resolved and remain unresolved. The degree to which the core technical problem remains unresolved is calculated based on the number of associations and the coverage of the solutions.

7. The enterprise operation demand data linkage analysis method according to claim 5, characterized in that, The degree of mismatch in the progress of computing resources includes: For each key task assigned to core technology resources, a list of micro-progress events or deliverables associated with the key task is preset. The list of micro-progress events or deliverables includes code submission, automated test pass record, interface integration completion record, and / or design document finalization. Within the preset monitoring period, the occurrence of micro-progress events or products is continuously monitored, the occurrence time and quantity of each event or product are recorded, and the actual progress score of the key task is calculated according to the preset event weight. Obtain the progress status of the key tasks reported in the project management system, including the task completion percentage or stage marker; The actual progress score is compared with the progress status, and the difference between the two is calculated. The difference value is normalized to a range of 0 to 1 to obtain the degree of mismatch in resource progress.

8. The enterprise operation demand data linkage analysis method according to claim 7, characterized in that, Before obtaining the degree of mismatch in the resource progress, the method further includes: Obtain the resource allocation ratio for the key tasks; Within a preset monitoring period, the actual working time of resources related to the key tasks on internal communication channels is statistically analyzed, and the actual resource availability ratio is calculated. The actual resource availability ratio is compared with the resource allocation ratio to obtain the availability deviation value; The availability deviation value is normalized to a range of 0 to 1 according to a preset mapping rule, and then weighted and synthesized with the normalized difference value to calculate the degree of mismatch in resource progress.

9. The enterprise operation demand data linkage analysis method according to claim 8, characterized in that, The statistics, before including the actual working hours of resources related to the key tasks in internal communication channels, also include: The system performs keyword and phrase matching on communication texts in internal communication channels and related to resource constraints or inter-task dependencies to obtain preliminary matching results. Perform context analysis on the communication text corresponding to the preliminary matching results to identify the sender, discussion topic, and relevance to the task or resource; Based on the preliminary matching results and the context analysis, the confidence score of the communication text is calculated; When the confidence score is higher than the preset confidence threshold, the communication text is confirmed as a task obstruction indication message, and the corresponding obstruction time is deducted from the resource's working time to update the actual working time of the resource within the preset monitoring period.

10. A data linkage analysis system for enterprise operational needs, characterized in that, include: The text acquisition and parsing module is used to acquire communication texts from internal communication channels, parse the communication texts according to preset language specificity scoring rules, and obtain information clarity data. The sentiment recognition module is used to match each word of the communication text and accumulate weights based on a keyword weight table containing negative sentiment words, positive sentiment words and negative words, so as to obtain the sentiment recognition result. The status acquisition and comparison module is used to acquire objective status data of the project, compare the objective status data of the project with the corresponding emotion recognition results, and obtain inconsistency data. The reputation calculation module is used to calculate the channel reputation score of the internal communication channel based on the information clarity data and the inconsistency data. The risk calculation module is used to weight the sentiment recognition results based on the channel reputation score, and perform a comprehensive risk judgment in combination with the objective status data of the project, and output the project risk result; The alarm triggering module is used to trigger a high-priority risk alarm and output the project risk result when the channel reputation score is lower than a preset first threshold and the information clarity data is detected to be higher than a preset second threshold.

Images