A multi-party collaborative contribution quantitative calculation and abnormal evaluation data correction method

Abstract

This invention relates to a method for quantifying and calculating the contribution of multi-party collaborative tasks and correcting abnormal evaluation data. The method includes: dividing a collaborative task into multiple work stages and sub-tasks; ranking the work stages by importance among participating entities and calculating stage weights; recording the primary responsibility and assistance relationships of each participating entity in the sub-tasks using a structured task list; collecting contribution evaluation data from participating entities playing assistance roles; identifying abnormal evaluation data using statistical methods or artificial intelligence anomaly detection algorithms; removing, replacing, or predictively correcting abnormal evaluation data; calculating a comprehensive contribution based on stage weights, the number of primary responsibility sub-tasks, the number of assistance sub-tasks, and the corrected assistance evaluation score; and normalizing the comprehensive contribution to output a normalized contribution ratio. This method solves the problems of non-standardized contribution and inadequate handling of abnormal data in existing collaborative task data processing, achieving unified calculation of primary responsibility and assistance contributions.

Description

Technical Field

[0001] This invention relates to the fields of computer data processing, collaborative task management, and abnormal data detection, and in particular to a method for contribution quantification calculation and abnormal evaluation data correction for multi-party collaborative tasks. Background Technology

[0002] In collaborative projects involving multiple parties, such as engineering design, scheme development, joint research and development, and cross-unit collaborative projects, a complex task is typically completed by multiple participants. These tasks often include multiple work phases, such as preparation, overall design, scheme conception, form design, and deliverable production and submission. Each work phase can be further divided into multiple sub-tasks.

[0003] In actual collaboration, different participants play different roles in different sub-tasks. For a given sub-task, there is usually one main responsible party and one or more assisting parties. The main responsible party undertakes the main tasks such as task advancement and results integration, while the assisting parties undertake auxiliary tasks such as providing data, technical support, content collaboration, and verification.

[0004] 1. Existing methods typically employ manual presets or fixed weights, making it difficult to dynamically reflect the importance of each work stage based on the actual situation of different collaborative tasks;

[0005] 2. Outliers, malicious scoring, or data entry errors may exist in multi-subject evaluation data. Directly taking the average value will reduce the accuracy of the contribution calculation results. 3. Existing methods lack unified modeling of task structure, primary responsibility role, assisting role, and evaluation data, making it difficult to automate the contribution calculation process; Summary of the Invention The purpose of this invention is to provide a collaborative task contribution calculation method based on stage weights and anomaly score correction, which solves the problems in the existing collaborative task data processing process, such as the difficulty in dynamically determining stage weights, the susceptibility of assistance evaluation data to outliers, the difficulty in uniformly calculating primary responsibility contribution and assistance contribution, and the lack of standardized output of contribution results.

[0006] This invention models collaborative tasks in stages, collects stage importance ranking data, subtask division data, and assistance evaluation data, and combines anomaly detection algorithms to identify and correct the evaluation data, ultimately generating the comprehensive contribution and normalized contribution ratio of each participating entity.

[0007] The technical solution of this invention is as follows: A method for quantifying multi-party collaborative contributions and correcting outlier evaluation data includes: Step 1. Divide the collaborative task into multiple work phases and multiple sub-tasks; Step 2. Rank the work stages by importance through the participating entities and calculate the stage weights; Step 3. Record the primary responsibility and cooperation relationships of each participating entity in the sub-tasks using a structured task list; Step 4. Collect contribution evaluation data of the participating entities that assume the assisting role; Step 5. Identify abnormal evaluation data using statistical methods or artificial intelligence anomaly detection algorithms; Step 6. Remove, replace, or predict and correct abnormal evaluation data; Step 7. Calculate the overall contribution based on the stage weight, the number of primary responsibility sub-tasks, the number of assistance sub-tasks, and the revised assistance evaluation score; Step 8. Normalize the overall contribution and output the normalized contribution ratio.

[0008] Furthermore, specifically including: Step S1: Establish a collaborative task phase-subtask hierarchical data model Receive collaborative task data and divide a collaborative task into N task stages, which are sequentially numbered according to the execution flow or business process as follows: 1,2,…,n,…,N Where N represents the total number of task stages included in the collaborative task; Each task phase is further divided into several sub-tasks; the number of sub-tasks included in different task phases may vary. Let the nth task phase contain R n If there are n subtasks, then the subtasks in the nth task stage can be represented as: T n,1 ,T n,2 ,…T n,j ,…,T n,Rn Where: n represents the task stage number, and n = 1, 2, ..., N; R n T represents the number of subtasks contained in the nth task phase; n,j Let j represent the j-th subtask in the n-th task stage, where j = 1, 2, ..., R. n ; For each subtask T n,j Establish a structured subtask record; the structured subtask record shall include at least the following fields: Subtask number; Subtask name; Task stage number n; Subtask number j within the stage; Participant identifier; Participant type; Participant role; The participating entities include individuals, teams, departments, companies, or other collaborating units; the participating roles include primary responsible roles and assisting roles. Let the total number of participants be M, and let the participants be denoted as: m = 1, 2, ..., M; Step S2: Calculate stage weights based on sorted data Each participating entity ranks the importance of the N task phases; the ranking score is from 1 to N, where N represents the most important and 1 represents the least important. Let W n,m W represents the importance ranking score given by the m-th participating entity to the n-th task phase; n K represents the average ranking score of the nth task stage; n This represents the stage weight of the nth task stage; The average ranking score for the nth task stage is:

[0009] The average ranking scores for each stage are normalized to obtain the stage weights:

[0010] in:

[0011] The above method converts the stage importance ranking data submitted by multiple participating entities into computable stage weight parameters. Step S3: Collect contribution data from the primary and assisting roles. For each subtask T n,j Record the roles of each participating entity in this subtask; S3.1 Contribution Data of the Primary Responsible Role For a certain subtask T n,j The participating entities that assume the primary responsibility receive a preset basic contribution value; S3.2 Assistance Role Evaluation Data For a certain subtask T n,j The participating entities that play an assisting role are evaluated by relevant evaluators for their assistance; the evaluation score ranges from 0 to 10, with a higher score indicating a greater contribution to the assistance. Suppose that the set of evaluation data received by a participating entity that assumes an assisting role is: Q={q1,q2,…,q t} Where: q i This represents the evaluation score given by the i-th evaluation subject; t represents the number of evaluation subjects participating in the evaluation. Before anomaly detection and correction, the original average assistance evaluation score was:

[0012] Anomaly detection and correction are performed on the evaluation dataset Q to obtain the corrected collaborative evaluation score Q. fixed ; Step S4: Perform anomaly detection and correction on the assisted evaluation data. The data used for evaluation were tested, cleaned, and corrected. S4.1 Historical Data Modeling Collect evaluation data from historical collaborative tasks and establish a normal collaborative evaluation data distribution model; Historical data is categorized and modeled according to the following dimensions: task type; task stage; sub-task type; participating entity type; participating role; historical evaluation habits of the evaluator; historical assisted evaluation score range; historical evaluation data fluctuation range. By using the historical data mentioned above, a baseline range for normal evaluation data is obtained, which is used to identify abnormal evaluation data that deviates significantly from the normal range. S4.2 Anomaly Evaluation Data Detection One or more of the following methods are used to identify anomalous evaluation data: standard deviation threshold detection; box plot detection; cluster analysis; isolated forest algorithm; anomaly detection model based on historical evaluation distribution; other machine learning anomaly detection algorithms; Anomaly detection was performed using the mean and standard deviation. For the evaluation dataset Q, calculate its mean and standard deviation:

[0013]

[0014] When a certain evaluation score Data is marked as anomalous if it meets one of the following conditions: or

[0015] That is, when the evaluation score is significantly higher or lower than 3 times the standard deviation of the mean, the evaluation score is identified as outlier data; S4.3 Correction of Abnormal Evaluation Data For evaluation data identified as anomalous, correct it using any of the following methods: remove the anomalous evaluation data and recalculate the average; replace the anomalous evaluation data with the median of the evaluation dataset; replace the anomalous evaluation data with the average evaluation score of historical similar tasks; replace the anomalous evaluation data with a reasonable evaluation score predicted by a machine learning model. The revised assistance evaluation score (Q) is obtained from the participating entities that assume the assistance role. fixed ; Q is used in subsequent contribution calculations. fixed Instead of directly using unprocessed Q raw ; S4.4 Evaluation Data Processing Log Output Generate an evaluation data processing log to record the evaluation data processing process; The log content includes: raw evaluation data; identified abnormal evaluation data; the basis for identifying abnormal evaluation data; the correction method used; the corrected evaluation data; and the final output Q. fixed ; By outputting evaluation data processing logs, the traceability of the calculation process is improved; Step S5: Calculate the overall contribution of participating entities After completing the stage weight calculation and assisting in the correction of evaluation data, the overall contribution of each participating entity is calculated. Let A m,n B represents the number of subtasks for which the m-th participating entity assumes a primary responsibility in the n-th task phase; m,n K represents the number of subtasks in which the m-th participating entity assumes a supporting role in the n-th task phase; n Q represents the stage weight of the nth task stage; m,n This represents the assistance evaluation score when the m-th participating entity assumes an assisting role in the n-th task phase; Among them, Q m,n The preferred score is the assist evaluation score after anomaly detection and correction; Then the overall contribution S of the m-th participating entity m for:

[0016] In this formula: 10×A m,n Q represents the basic contribution of the m-th participating entity in the sub-task where it assumes a primary responsibility in the n-th task phase; m,n ×B m,n K represents the contribution of the m-th participating entity in the sub-task of the assisting role in the n-th task phase; n This represents the weight of the nth task phase in the entire collaborative task; This calculation method takes into account: the importance of different task stages; the number of sub-tasks in which the participating entity assumes the primary responsibility; the number of sub-tasks in which the participating entity assumes the assisting role; and the evaluation score of the assisting work after anomaly detection and correction. Step S6: Output the normalized contribution ratio After calculating the overall contribution of all participating entities, the overall contribution of each participating entity is normalized. Let S be the value of S. m P represents the overall contribution of the m-th participating entity; m This represents the normalized contribution ratio of the m-th participating entity; but:

[0017] Finally, the normalized contribution ratio set is output:

[0018] The normalized contribution ratio is used as a standardized output parameter for collaborative task evaluation, project management analysis, resource scheduling, task settlement reference, performance analysis, or other downstream business systems.

[0019] A contribution quantification calculation and anomaly evaluation data correction system is provided to implement the multi-party collaborative contribution quantification calculation and anomaly evaluation data correction method as described above, comprising: The task data input module is used to input basic information about collaborative tasks, such as task stages, subtasks, participating entities, types of participating entities, and participating roles. Input methods include: table import; Excel import; web form entry; database interface import; and third-party project management system interface import. The phase-subtask modeling module is used to build a hierarchical data model of "project - task phase - subtask - participating entity - participating role" based on input data; This module is used to record the number R of subtasks in each task stage. n And generate a set of subtasks: T n,1 ,T n,2 ,…,T n,Rn The phase ranking and data collection module is used to collect data on the ranking of the importance of each task phase by each participating entity, and form W. n,m Data set; The stage weight calculation module is used to calculate the average ranking score W for each stage based on the ranking data submitted by each participating entity. n The stage weights K are generated through normalization. n ; The subtask contribution data collection module is used to record the participating entities and their roles in each subtask, and to collect the evaluation data of the evaluators on the participating entities that assume the assisting role. The abnormal evaluation data detection module is used to detect anomalies in the assisted evaluation data and identify evaluation data that deviates significantly from the normal range. This module employs standard deviation thresholding, box plots, cluster analysis, isolated forest algorithms, or other machine learning algorithms; The evaluation data correction module is used to remove, replace, or predict and correct abnormal evaluation data, and output the corrected assisted evaluation score Q. fixed ; The comprehensive contribution calculation module is used to calculate the contribution based on the stage weight K. n Number of main responsibility sub-tasks A m,n Number of sub-tasks to assist B m,n And assist in the evaluation score Q m,n Calculate the overall contribution S of each participating entity. m ; The normalization output module is used to calculate the normalized contribution ratio P based on the overall contribution of each participating entity. m It also outputs the contribution calculation results; The Reports and Logs module is used to generate contribution calculation reports and evaluation data processing logs; The report may include: stage weight data; sub-task division data; participating entity data; participating role data; original evaluation data; abnormal evaluation data; corrected evaluation data; overall contribution of each participating entity; and normalized contribution ratio of each participating entity.

[0020] The beneficial effects of this invention are as follows: 1. Improve the adaptability of stage weight calculation This invention dynamically generates stage weights K by collecting data on the importance of each participating entity to a task stage, and then averaging and normalizing the ranked data. n Compared to fixed stage weights, this invention can adapt to situations where the importance of stages varies in different collaborative tasks.

[0021] 2. Supports task models with different numbers of subtasks at different stages. This invention uses R n This indicates the number of subtasks included in the nth task stage, enabling the system to support collaborative task structures with different numbers of subtasks in different stages, thus improving the flexibility of task modeling.

[0022] 3. Adaptable to multi-entity collaborative scenarios such as individuals, teams, departments, and companies. This invention uses a modeling approach that combines participating entities and participating roles, making it applicable to collaborations between individuals as well as joint collaborations between teams, departments, companies, or organizations.

[0023] 4. Improve the reliability of data used for evaluation. This invention uses an anomaly detection and evaluation data correction mechanism to identify and process extreme evaluation data, erroneous data, or data that deviates significantly from the normal range, thereby reducing the impact of abnormal evaluation data on the contribution calculation results.

[0024] 5. Achieve unified calculation of primary responsibility contributions and assisting contributions. This invention uses A m,n This indicates the number of sub-tasks for which the primary responsibility is assumed, expressed via B. m,n It indicates the number of sub-tasks undertaken in the assisting role, and incorporates the primary responsibility basic contribution value and the assistance evaluation score into a unified formula to achieve unified quantification of primary responsibility contribution and assistance contribution.

[0025] 6. Improve the automation level of the contribution calculation process. This invention can automatically complete task data entry, stage weight calculation, evaluation data anomaly detection, evaluation data correction, comprehensive contribution calculation and normalization output through a computer system, reducing the manual statistical process.

[0026] 7. Improve the traceability of calculation results This invention can generate an evaluation data processing log, which records the original evaluation data, abnormal evaluation data, correction methods and final calculation results, facilitating subsequent verification and traceability.

[0027] 8. Improve the standardization of output results. This invention outputs the normalized contribution ratio P m This ratio can be used as a contribution parameter in a unified format for different types of collaborative task management systems or downstream business systems to call. Attached Figure Description

[0028] Figure 1 This is a flowchart of the method of the present invention; Figure 2 This is a schematic diagram of the hierarchical data model of a subtask and participating entities in the present invention. Figure 3 This is a flowchart of the abnormal evaluation data detection and correction process of the present invention. Detailed Implementation

[0029] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0030] The present invention achieves the above objectives in the following ways: 1. Divide collaborative tasks into multiple work phases and multiple sub-tasks; 2. Rank the importance of each work stage by the participating entities and calculate the stage weights; 3. Record the primary responsibility and collaborative relationships of each participating entity in the sub-tasks through a structured task list; 4. Collect contribution evaluation data of participating entities that assume the assisting role; 5. Identify abnormal evaluation data using statistical methods or artificial intelligence anomaly detection algorithms; 6. Remove, replace, or predict and correct abnormal evaluation data; 7. Calculate the overall contribution based on the stage weight, the number of primary responsibility sub-tasks, the number of assistance sub-tasks, and the revised assistance evaluation score; 8. Normalize the overall contribution and output the normalized contribution ratio.

[0031] A method for quantifying the collaborative contribution of multiple parties in a multi-party collaborative task and correcting outlier evaluation data includes: Step S1: Establish a collaborative task phase-subtask hierarchical data model The system receives collaborative task data and divides a collaborative task into N task stages, which are numbered sequentially according to the execution flow or business process as follows: 1,2,…,n,…,N Where N represents the total number of task stages included in the collaborative task.

[0032] Each task phase is further divided into several sub-tasks. The number of sub-tasks included in different task phases can vary.

[0033] Let the nth task phase contain R n If there are n subtasks, then the subtasks in the nth task stage can be represented as: T n,1 ,T n,2 ,…T n,j ,…,T n,Rn Where: n represents the task stage number, and n = 1, 2, ..., N; R n T represents the number of subtasks contained in the nth task phase; n,j Let j represent the j-th subtask in the n-th task stage, where j = 1, 2, ..., R. n .

[0034] For example, a collaborative task may consist of N=5 task phases, where: the first phase contains 2 subtasks, i.e., R1=2; the second phase contains 4 subtasks, i.e., R2=4; the third phase contains 3 subtasks, i.e., R3=3; the fourth phase contains 2 subtasks, i.e., R4=2; and the fifth phase contains 3 subtasks, i.e., R5=3.

[0035] The phase-subtask structure of this collaborative task can then be represented as: {T 1,1 ,T 1,2},{T 2,1 ,T 2,2 ,T 2,3 ,T 2,4},{T 3,1 ,T 3,2 ,T 3,3},{T 4,1 ,T 4,2},{T 5,1 ,T 5,2 ,T 5,3} The system targets each subtask T n,j Create a structured subtask record. The structured subtask record shall include at least the following fields: Subtask number; Subtask name; Task stage number n; Subtask number j within the stage; Participant identifier; Participant type; Participant role; The participating entities include individuals, teams, departments, companies, or other collaborating units; the participating roles include primary responsible roles and assisting roles.

[0036] Let the total number of participants be M, and let the participants be denoted as: m=1,2,…,M Using the above method, the system constructs collaborative tasks into a hierarchical data model of "project - task phase - sub-task - participating entity - participating role". This hierarchical data model is used to subsequently count the number of sub-tasks for which each participating entity assumes a primary responsibility role, the number of sub-tasks for which they assume a supporting role, and the corresponding assistance evaluation data in different task phases.

[0037] Step S2: Calculate stage weights based on sorted data Each participating entity ranks the importance of the N task phases. The ranking score is from 1 to N, where N represents the most important and 1 represents the least important.

[0038] Let W n,m W represents the importance ranking score given by the m-th participating entity to the n-th task phase; nK represents the average ranking score of the nth task stage; n This represents the stage weight of the nth task stage.

[0039] The average ranking score for the nth task stage is:

[0040] The average ranking scores for each stage are normalized to obtain the stage weights:

[0041] in:

[0042] In this way, the system converts the stage importance ranking data submitted by multiple participating entities into computable stage weight parameters, avoiding the problem of insufficient adaptability caused by using fixed weights.

[0043] Step S3: Collect contribution data from the primary and assisting roles. For each subtask T n,j The system records the role of each participating entity in the subtask.

[0044] S3.1 Contribution Data of the Primary Responsible Role For a certain subtask T n,j The participating entity that assumes the primary responsibility receives a preset basic contribution value. For example, this basic contribution value is set to 10.

[0045] In other implementations, the base contribution value can also be adjusted to other values ​​based on the task type, task complexity, or system preset parameters.

[0046] S3.2 Assistance Role Evaluation Data For a certain subtask T n,j The participating entities that play a supporting role will have their assistance evaluated by relevant evaluators. The evaluation score ranges from 0 to 10, with higher scores indicating greater contributions.

[0047] Suppose that the set of evaluation data received by a participating entity that assumes an assisting role is: Q={q1,q2,…,q t} in: q i This represents the evaluation score given by the i-th evaluator; t represents the number of evaluation entities participating in the evaluation.

[0048] Before anomaly detection and correction, the original average assistance evaluation score was:

[0049] To avoid abnormal evaluation data from affecting the final calculation results, this invention further performs anomaly detection and correction on the evaluation data set Q to obtain the corrected assisted evaluation score Q. fixed .

[0050] Step S4: Perform anomaly detection and correction on the assisted evaluation data. The system includes an anomaly detection and scoring correction module, which is used to detect, clean, and correct the data used for evaluation.

[0051] This module can be implemented using statistical methods, machine learning algorithms, or a combination of both.

[0052] S4.1 Historical Data Modeling The system can collect evaluation data from historical collaborative tasks and establish a normal collaborative evaluation data distribution model.

[0053] Historical data can be categorized and modeled according to the following dimensions: task type; task stage; sub-task type; participating entity type; participating role; historical evaluation habits of the evaluator; historical assisted evaluation score range; and historical evaluation data fluctuation range.

[0054] Using the historical data mentioned above, the system can obtain the baseline range of normal evaluation data, which can be used to identify abnormal evaluation data that deviates significantly from the normal range.

[0055] S4.2 Anomaly Evaluation Data Detection The system can use one or more of the following methods to identify abnormal evaluation data: standard deviation threshold detection; box plot detection; cluster analysis; isolated forest algorithm; anomaly detection model based on historical evaluation distribution; and other machine learning anomaly detection algorithms.

[0056] In this invention, the system uses the mean and standard deviation for anomaly detection.

[0057] For the evaluation dataset Q, calculate its mean and standard deviation:

[0058]

[0059] When a certain evaluation score Data is marked as anomalous if it meets one of the following conditions: or

[0060] When the evaluation score is significantly higher or lower than three times the standard deviation of the mean, the system will identify the evaluation score as abnormal data.

[0061] S4.3 Correction of Abnormal Evaluation Data For evaluation data identified as anomalous, the system can correct it in any of the following ways: recalculate the average value after removing the anomalous evaluation data; replace the anomalous evaluation data with the median of the evaluation data set; replace the anomalous evaluation data with the average evaluation score of historical similar tasks; or replace the anomalous evaluation data with a reasonable evaluation score predicted by a machine learning model.

[0062] The revised assistance evaluation score (Q) is obtained from the participating entities that assume the assistance role. fixed ; Q is used in subsequent contribution calculations. fixed Instead of directly using unprocessed Q raw .

[0063] In one specific implementation, the following pseudocode can be used for exception detection and correction: Python # scores is the list of scores collected during this evaluation assistance. import numpy as np Q = np.array(scores) mean = Q.mean() std = Q.std() # Mark outliers: Scores that are more than 3 standard deviations from the mean anomalies = Q[(Q>mean + 3 std) | (Q <mean - 3 std)] # Correction method: Recalculate the average after removing outliers Q_fixed = Q[(Q<= mean + 3 std)&(Q>= mean - 3 std).mean() The code above is only one example implementation. In other implementations, median replacement, historical mean replacement, or machine learning model prediction replacement can also be used.

[0064] S4.4 Evaluation Data Processing Log Output The system can generate evaluation data processing logs to record the processing of evaluation data.

[0065] The log content may include: raw evaluation data; identified abnormal evaluation data; the basis for identifying abnormal evaluation data; the correction method used; the corrected evaluation data; and the final output Q. fixed .

[0066] By outputting evaluation data processing logs, the traceability of the system's calculation process can be improved.

[0067] Step S5: Calculate the overall contribution of participating entities After completing the stage weight calculation and assisting in the correction of evaluation data, the system calculates the overall contribution of each participating entity.

[0068] Let A m,n B represents the number of subtasks for which the m-th participating entity assumes a primary responsibility in the n-th task phase; m,n K represents the number of subtasks in which the m-th participating entity assumes a supporting role in the n-th task phase; n Q represents the stage weight of the nth task stage; m,n This represents the assistance evaluation score when the m-th participant assumes an assisting role in the n-th task phase.

[0069] Among them, Q m,n The preferred score is the assist evaluation score after anomaly detection and correction.

[0070] Then the overall contribution S of the m-th participating entity m for:

[0071] In this formula: 10×A m,n Q represents the basic contribution of the m-th participating entity in the sub-task where it assumes a primary responsibility in the n-th task phase; m,n ×B m,n K represents the contribution of the m-th participating entity in the sub-task of the assisting role in the n-th task phase; n This represents the weight of the nth task stage in the entire collaborative task.

[0072] This calculation method can simultaneously consider: the importance of different task stages; the number of sub-tasks for which the participating entity assumes a primary responsibility role; the number of sub-tasks for which the participating entity assumes a supporting role; and the evaluation score of the supporting work after anomaly detection and correction.

[0073] Step S6: Output the normalized contribution ratio After calculating the overall contribution of all participating entities, the system normalizes the overall contribution of each participating entity.

[0074] Let S be the value of S. m P represents the overall contribution of the m-th participating entity; m This represents the normalized contribution ratio of the m-th participating entity.

[0075] but:

[0076] Finally, the system outputs a set of normalized contribution ratios:

[0077] The normalized contribution ratio can be used as a standardized output parameter for collaborative task evaluation, project management analysis, resource scheduling, task settlement reference, performance analysis, or other downstream business systems.

[0078] VI. System Implementation Method The method of this invention can be implemented using a computer system. This system may include the following modules.

[0079] 1. Task Data Input Module This is used to input basic information about collaborative tasks, including task stages, subtasks, participating entities, participating entity types, and participating roles.

[0080] Input methods can include: table import; Excel import; web form entry; database interface import; and third-party project management system interface import.

[0081] 2. Phase-Subtask Modeling Module This is used to build a hierarchical data model based on input data, consisting of "project - task phase - subtask - participating entity - participating role".

[0082] This module is used to record the number R of subtasks in each task stage. n And generate a set of subtasks: T n,1 ,T n,2 ,…,T n,Rn 3. Stage-based sorting and data acquisition module This is used to collect data on the ranking of the importance of each participating entity to the task phase, and to form W. n,m Data set.

[0083] 4. Stage Weight Calculation Module Used to calculate the average ranking score W at each stage based on the ranking data submitted by each participating entity. n The stage weights K are generated through normalization. n .

[0084] 5. Subtask Contribution Data Collection Module This is used to record the participants and their roles in each subtask, and to collect evaluation data from the evaluators on the participants who assume the assisting role.

[0085] 6. Anomaly Evaluation Data Detection Module Used to detect anomalies in assisted evaluation data and identify evaluation data that deviates significantly from the normal range.

[0086] This module can employ standard deviation thresholding, box plots, cluster analysis, isolated forest algorithms, or other machine learning algorithms.

[0087] 7. Evaluation Data Correction Module Used to remove, replace, or predict and correct outlier evaluation data, and output the corrected auxiliary evaluation score Q. fixed .

[0088] 8. Overall Contribution Calculation Module Used according to stage weight K n Number of main responsibility sub-tasks A m,n Number of sub-tasks to assist B m,n And assist in the evaluation score Q m,n Calculate the overall contribution S of each participating entity. m .

[0089] 9. Normalized Output Module Used to calculate the normalized contribution ratio P based on the overall contribution of each participating entity. m It also outputs the contribution calculation results.

[0090] 10. Reporting and Log Module Used to generate contribution calculation reports and evaluation data processing logs.

[0091] The report may include: stage weight data; sub-task division data; participating entity data; participating role data; original evaluation data; abnormal evaluation data; corrected evaluation data; overall contribution of each participating entity; and normalized contribution ratio of each participating entity. VII. Specific Implementation Methods The invention will now be described in conjunction with a specific collaborative task.

[0093] Suppose that four participants, A, B, C, and D, collaborate to complete a project. These participants can be individuals, teams, departments, or companies. The project is divided into five phases, each containing several sub-tasks.

[0094] The five task phases are: preparation phase; overall design phase; scheme conception phase; form design phase; and deliverable production and submission phase.

[0095] Each participating entity ranked the importance of the above five stages. The average ranking of the stages was as follows:

[0096] therefore:

[0097] The weights for each stage are: ; ; ; ; ; In a certain subtask, a participant who plays an assisting role receives evaluation scores from four evaluators, A, B, C, and D, which are 6, 8, 5, and 5 respectively. The original average assistance evaluation score is:

[0098] The system performed anomaly detection on the evaluation dataset and found no evaluation data exceeding the preset anomaly threshold. Therefore, the corrected evaluation score is Q=6. Assume that entity A undertakes 1 primary subtask and 2 supporting subtasks during the preparation phase. Then: A A,1 =1; B A,1 =2;Q A,1 =6; Entity A's contribution during the preparation phase is as follows:

[0099] That is: 0.20×(10×1+6×2)=0.20×(10+12)=4.4; The system continues to calculate the contribution of entity A in the overall design, scheme conception, form design, deliverable production and submission stages in the same way, and adds up the results of the five stages to obtain the comprehensive contribution of entity A.

[0100] Assume the final calculation yields:

[0101] The total contribution is: 18.2 + 15.6 + 13.7 + 14.8 = 62.3 Then the normalized contribution ratio of subject A is: P A =18.2 / 62.3≈0.29 The normalized contribution ratios of other participating entities can be calculated in the same way.

[0102] This normalized contribution ratio can be used as a standardized contribution parameter output by the system for subsequent collaborative task evaluation, resource allocation reference, project management analysis, or other business system calls.

[0103] Figure 1 illustrates the overall process of the present invention, including: inputting collaborative task data; establishing a hierarchical data model of collaborative task stages, sub-tasks, and participating entities; collecting stage importance ranking data; calculating stage weights; collecting contribution data of the primary and assisting roles; performing anomaly detection on the assistance evaluation data; correcting the abnormal evaluation data; calculating the comprehensive contribution of participating entities; and outputting the normalized contribution ratio and processing log.

[0104] Figure 2 illustrates the hierarchical structure of a collaborative task, where one collaborative task comprises N task phases, and the nth task phase includes R. n Each subtask is associated with a participating entity, participating role, and evaluation data.

[0105] Figure 3 illustrates the system's process for anomaly detection and correction of assisted evaluation data, including collecting raw evaluation data, calculating the mean and standard deviation, identifying abnormal evaluation data, correcting abnormal evaluation data, and outputting the corrected evaluation score.

[0106] Key technical points 1. Stage-Subtask-Participating Entity Hierarchical Data Modeling Mechanism The collaborative task is divided into N task phases, and the nth task phase is further divided into R phases. n Sub-tasks: T n,1 ,T n,2 ,…,T n,Rn Among them, the number of subtasks R in different task stages n They can be different. Each subtask is associated with a participating entity and its participating role.

[0107] 2. Data modeling mechanism based on the primary responsibility and contribution of participating roles The structured subtask record records the roles of participating entities in each subtask, including primary and supporting roles. It also counts the number of primary subtasks (A) for each participating entity at each stage. m,n And the number of sub-tasks B m,n .

[0108] 3. Stage weighting mechanism based on sorted data The average ranking score W is calculated by ranking the importance of each task phase from 1 to N based on the participating entities. n The stage weights K are generated using a normalization formula. n :

[0109] This mechanism converts the stage ranking data submitted by multiple entities into computable stage weight parameters.

[0110] 4. A contribution calculation mechanism that combines the primary responsibility's basic contribution value with assistance evaluation scores. A preset basic contribution value is set for the participants who assume the primary responsibility role, preferably 10; evaluation data from 0 to 10 is collected for the participants who assume the assisting role, and the corrected assistance evaluation score is obtained through anomaly detection and correction.

[0111] 5. Mechanism for detecting and correcting abnormal evaluation data Anomaly detection is performed on the assistance evaluation dataset, and abnormal evaluation data are removed, replaced by the median, replaced by the historical mean, or corrected by the model prediction to generate a corrected assistance evaluation score.

[0112] 6. Formula for calculating overall contribution The overall contribution of participating entities is calculated using the following formula:

[0113] Among them, Q m,n The preferred score is the assist evaluation score after anomaly detection and correction.

[0114] 7. Normalized Contribution Ratio Output Mechanism The normalized contribution ratio of participating entities is calculated using the following formula:

[0115] This method generates standardized contribution ratio parameters for subsequent system calls.

[0116] 8. Computer System Implementation Methods This invention can be implemented through a computer system comprising a task data input module, a stage-subtask modeling module, a stage sorting and acquisition module, a stage weight calculation module, a subtask contribution data acquisition module, an anomaly evaluation data detection module, an evaluation data correction module, a comprehensive contribution calculation module, a normalization output module, and a report log module.

[0117] XI. Alternative Implementation Methods Without changing the core idea of ​​the invention, the invention can also have the following alternative embodiments: The primary responsibility base contribution value is not limited to 10, but can also be set to other values ​​according to task type, task difficulty or system parameters; The range of scores for the evaluation is not limited to 0 to 10; scores of 0 to 5, 0 to 100, or other scoring scales may also be used. In the calculation of stage weights, in addition to simple average, weighted average, extreme value removal average or other normalization methods can also be used; Anomaly detection methods are not limited to 3 times the standard deviation; box plots, isolation forests, clustering algorithms, neural network models, or other anomaly detection algorithms can also be used. The methods for correcting outlier evaluation data are not limited to removing outliers; they can also include replacing with the median, historical mean, or model prediction. Participants are not limited to companies or organizations; they can also be individuals, teams, departments, project groups, institutions, or other collaborating entities. The roles involved are not limited to primary and secondary roles, but can also be expanded to include review roles, verification roles, support roles, coordinating roles, etc., depending on the task type. The output results are not limited to the normalized contribution ratio report, but can also be output through API interface, database interface, spreadsheet file or visualization interface; The method can be applied to multi-entity collaborative task scenarios such as engineering collaboration, design collaboration, solution collaboration, joint research and development, cross-unit project management, and individual collaborative task management.

[0118] Key technical content: 1. A contribution quantification calculation method for multi-party collaborative tasks, the core of which is to divide the collaborative task into multiple task stages, and further divide each task stage into several sub-tasks, forming a hierarchical data model of "collaborative task - task stage - sub-task - participating subject - participating role".

[0119] 2. A method for calculating stage weights based on stage importance ranking data. Each participating entity ranks the task stages according to their importance. The system calculates the average ranking score based on the ranking data submitted by each participating entity, and normalizes the average ranking score to obtain the stage weight of each task stage.

[0120] 3. A method for calculating primary and secondary contributions based on participating roles. The system records the participating roles of each participant in each subtask through structured subtask records, and calculates the number of subtasks in which each participant assumes a primary role and the number of subtasks in which they assume a secondary role at different task stages.

[0121] 4. A method for anomaly detection and correction of assistance evaluation data. The system identifies anomalies in the evaluation data obtained by participating entities that assume the assistance role, and removes, replaces, or predictively corrects the identified anomaly evaluation data to obtain a corrected assistance evaluation score.

[0122] 5. A method for calculating the comprehensive contribution by integrating stage weights, the number of sub-tasks for the primary responsible role, the number of sub-tasks for assisting roles, and the revised assistance evaluation score. The comprehensive contribution of the participating entities can be calculated according to the following formula:

[0123] Among them, S m K represents the overall contribution of the m-th participating entity. n A represents the stage weight of the nth task stage. m,n B represents the number of subtasks for which the m-th participating entity assumes a primary responsibility in the n-th task phase. m,n Q represents the number of subtasks in which the m-th participating entity assumes a supporting role in the n-th task phase. m,n This represents the modified assistance evaluation score when the m-th participant assumes an assisting role in the n-th task phase.

[0124] 6. A method for normalizing the overall contribution of each participating entity and outputting a normalized contribution ratio. The normalized contribution ratio can be used as a standardized data parameter for collaborative task evaluation, project management analysis, resource scheduling, performance analysis, or other downstream business systems.

[0125] Positive significance 1. Improve the objectivity and accuracy of collaborative task contribution calculation. This invention calculates the comprehensive contribution of participating entities by combining multiple dimensions such as stage weight, number of sub-tasks of the primary responsible role, number of sub-tasks of the assisting role, and revised assistance evaluation score, avoiding reliance on human experience or a single evaluation indicator for judgment, and improving the objectivity and accuracy of contribution calculation results.

[0126] 2. Improve the modeling capabilities for complex collaborative tasks. This invention divides collaborative tasks into multiple task phases and allows different task phases to have different numbers of sub-tasks. This can more realistically reflect the actual working structure of complex tasks such as engineering design, scheme preparation, joint research and development, and cross-departmental collaboration, and improve the system's adaptability to complex collaborative tasks.

[0127] 3. Reduce the impact of outlier evaluation data on calculation results Before the evaluation data is included in the contribution calculation, this invention performs anomaly detection and correction on the evaluation data. It can identify extreme scores, abnormal scores, data with input errors, or data that deviates significantly from the normal range, thereby reducing the interference of abnormal data on the final contribution calculation results.

[0128] 4. Achieve unified quantification of primary and secondary contributions. This invention incorporates the contributions of those who assume primary responsibility and those who assume auxiliary responsibility into a unified calculation model, enabling the system to simultaneously reflect the contributions of both primary and auxiliary collaborative work, thereby improving the completeness of the evaluation of multi-party collaborative tasks.

[0129] 5. Improve the standardization and traceability of contribution calculation results. This invention outputs the normalized contribution ratio of each participating entity through normalization processing, and can generate an evaluation data processing log to record the original evaluation data, abnormal evaluation data, correction methods and final output results, which facilitates subsequent business system calls, verification and traceability.

[0130] The above-described embodiments are merely one implementation of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A method for quantifying multi-party collaborative contributions and correcting anomaly evaluation data, characterized in that, include: Step 1. Divide the collaborative task into multiple work phases and multiple sub-tasks; Step 2. Rank the work stages by importance according to the participating entities and calculate the stage weights; Step 3. Record the primary responsibility and collaborative relationships of each participating entity in the sub-tasks using a structured task list; Step 4. Collect contribution evaluation data of the participating entities that assume the assisting role; Step 5. Identify abnormal evaluation data using statistical methods or artificial intelligence anomaly detection algorithms; Step 6. Remove, replace, or predict and correct abnormal evaluation data; Step 7. Calculate the overall contribution based on the stage weight, the number of primary responsibility sub-tasks, the number of assistance sub-tasks, and the revised assistance evaluation score; Step 8. Normalize the overall contribution and output the normalized contribution ratio.

2. The method for quantifying multi-party collaborative contributions and correcting anomaly evaluation data according to claim 1, characterized in that, Specifically, it includes: Step S1: Establish a collaborative task phase-subtask hierarchical data model Receive collaborative task data and divide a collaborative task into N task stages, which are sequentially numbered according to the execution flow or business process as follows: 1,2,…,n,…,N Where N represents the total number of task stages included in the collaborative task; Each task phase is further divided into several sub-tasks; the number of sub-tasks included in different task phases may vary. Let the nth task phase contain R n If there are n subtasks, then the subtasks in the nth task stage can be represented as: T n,1 ,T n,2 ,…T n,j ,…,T n,Rn Where: n represents the task stage number, and n = 1, 2, ..., N; R n T represents the number of subtasks contained in the nth task phase; n,j Let j represent the j-th subtask in the n-th task stage, where j = 1, 2, ..., R. n ; For each subtask T n,j Establish a structured subtask record; the structured subtask record shall include at least the following fields: Subtask number; Subtask name; Task stage number n; Subtask number j within the stage; Participant identifier; Participant type; Participant role; The participating entities include individuals, teams, departments, companies, or other collaborating units; the participating roles include primary responsible roles and assisting roles. Let the total number of participants be M, and let the participants be denoted as: m = 1, 2, ..., M; Step S2: Calculate stage weights based on sorted data Each participating entity ranks the importance of the N task phases; the ranking score is from 1 to N, where N represents the most important and 1 represents the least important. Let W n,m W represents the importance ranking score given by the m-th participating entity to the n-th task phase; n K represents the average ranking score of the nth task stage; n This represents the stage weight of the nth task stage; The average ranking score for the nth task stage is: The average ranking scores for each stage are normalized to obtain the stage weights: in: The above method converts the stage importance ranking data submitted by multiple participating entities into computable stage weight parameters. Step S3: Collect contribution data from the primary and assisting roles. For each subtask T n,j Record the roles of each participating entity in this subtask; S3.1 Contribution Data of the Primary Responsible Role For a certain subtask T n,j The participating entities that assume the primary responsibility receive a preset basic contribution value; S3.2 Assistance Role Evaluation Data For a certain subtask T n,j The participating entities that play an assisting role are evaluated by relevant evaluators for their assistance; the evaluation score ranges from 0 to 10, with a higher score indicating a greater contribution to the assistance. Suppose that the set of evaluation data received by a participating entity that assumes an assisting role is: Q={q1,q2,…,q t } Where: q i This represents the evaluation score given by the i-th evaluation subject; t represents the number of evaluation subjects participating in the evaluation. Before anomaly detection and correction, the original average assistance evaluation score was: Anomaly detection and correction are performed on the evaluation dataset Q to obtain the corrected collaborative evaluation score Q. fixed ; Step S4: Perform anomaly detection and correction on the assisted evaluation data. The data used for evaluation were tested, cleaned, and corrected. S4.1 Historical Data Modeling Collect evaluation data from historical collaborative tasks and establish a normal collaborative evaluation data distribution model; Historical data is categorized and modeled according to the following dimensions: task type; task stage; sub-task type; participating entity type; participating role; historical evaluation habits of the evaluator; historical assisted evaluation score range; historical evaluation data fluctuation range. By using the historical data mentioned above, a baseline range for normal evaluation data is obtained, which is used to identify abnormal evaluation data that deviates significantly from the normal range. S4.2 Anomaly Evaluation Data Detection One or more of the following methods are used to identify anomalous evaluation data: standard deviation threshold detection; box plot detection; cluster analysis; isolated forest algorithm; anomaly detection model based on historical evaluation distribution; other machine learning anomaly detection algorithms; Anomaly detection was performed using the mean and standard deviation. For the evaluation dataset Q, calculate its mean and standard deviation: When a certain evaluation score Data is marked as outlier if it meets one of the following conditions: or That is, when the evaluation score is significantly higher or lower than 3 times the standard deviation of the mean, the evaluation score is identified as outlier data; S4.3 Correction of Abnormal Evaluation Data For evaluation data identified as anomalous, correct it using any of the following methods: remove the anomalous evaluation data and recalculate the average; replace the anomalous evaluation data with the median of the evaluation dataset; replace the anomalous evaluation data with the average evaluation score of historical similar tasks; replace the anomalous evaluation data with a reasonable evaluation score predicted by a machine learning model. The revised assistance evaluation score (Q) is obtained from the participating entities that assume the assistance role. fixed ; Q is used in subsequent contribution calculations. fixed Instead of directly using unprocessed Q raw ; S4.4 Evaluation Data Processing Log Output Generate an evaluation data processing log to record the evaluation data processing process; The log content includes: raw evaluation data; identified abnormal evaluation data; the basis for identifying abnormal evaluation data; the correction method used; the corrected evaluation data; and the final output Q. fixed ; By outputting evaluation data processing logs, the traceability of the calculation process is improved; Step S5: Calculate the overall contribution of participating entities After completing the stage weight calculation and assisting in the correction of evaluation data, the overall contribution of each participating entity is calculated. Let A m,n B represents the number of subtasks for which the m-th participating entity assumes a primary responsibility in the n-th task phase; m,n K represents the number of subtasks in which the m-th participating entity assumes a supporting role in the n-th task phase; n Q represents the stage weight of the nth task stage; m,n This represents the assistance evaluation score when the m-th participating entity assumes an assisting role in the n-th task phase; Among them, Q m,n The preferred score is the assist evaluation score after anomaly detection and correction; Then the overall contribution S of the m-th participating entity m for: In this formula: 10×A m,n Q represents the basic contribution of the m-th participating entity in the sub-task where it assumes a primary responsibility in the n-th task phase; m,n ×B m,n K represents the contribution of the m-th participating entity in the sub-task of the assisting role in the n-th task phase; n This represents the weight of the nth task phase in the entire collaborative task; This calculation method takes into account: the importance of different task stages; the number of sub-tasks in which the participating entity assumes the primary responsibility; the number of sub-tasks in which the participating entity assumes the assisting role; and the evaluation score of the assisting work after anomaly detection and correction. Step S6: Output the normalized contribution ratio After calculating the overall contribution of all participating entities, the overall contribution of each participating entity is normalized. Let S be the value of S. m P represents the overall contribution of the m-th participating entity; m This represents the normalized contribution ratio of the m-th participating entity; but: Finally, output the set of normalized contribution ratios: The normalized contribution ratio is used as a standardized output parameter for collaborative task evaluation, project management analysis, resource scheduling, task settlement reference, performance analysis, or other downstream business systems.

3. A contribution quantification calculation and anomaly evaluation data correction system, characterized in that, The method for implementing the multi-party collaborative contribution quantification calculation and anomaly evaluation data correction method as described in any one of claims 1 and 2 includes: The task data input module is used to input basic information about collaborative tasks, such as task stages, subtasks, participating entities, types of participating entities, and participating roles. Input methods include: table import; Excel import; web form entry; database interface import; and third-party project management system interface import. The phase-subtask modeling module is used to build a hierarchical data model of "project - task phase - subtask - participating entity - participating role" based on input data; This module is used to record the number R of subtasks in each task stage. n And generate a set of subtasks: T n,1 ,T n,2 ,…,T n,Rn The phase ranking and data collection module is used to collect data on the ranking of the importance of each task phase by each participating entity, and form W. n,m Data set; The stage weight calculation module is used to calculate the average ranking score W for each stage based on the ranking data submitted by each participating entity. n The stage weights K are generated through normalization. n ; The subtask contribution data collection module is used to record the participating entities and their roles in each subtask, and to collect the evaluation data of the evaluators on the participating entities that assume the assisting role. The abnormal evaluation data detection module is used to detect anomalies in the assisted evaluation data and identify evaluation data that deviates significantly from the normal range. This module employs standard deviation thresholding, box plots, cluster analysis, isolated forest algorithms, or other machine learning algorithms; The evaluation data correction module is used to remove, replace, or predict and correct abnormal evaluation data, and output the corrected assisted evaluation score Q. fixed ; The comprehensive contribution calculation module is used to calculate the contribution based on the stage weight K. n Number of main responsibility sub-tasks A m,n Number of sub-tasks to assist B m,n And assist in the evaluation score Q m,n Calculate the overall contribution S of each participating entity. m ; The normalization output module is used to calculate the normalized contribution ratio P based on the overall contribution of each participating entity. m It also outputs the contribution calculation results; The Reports and Logs module is used to generate contribution calculation reports and evaluation data processing logs; The report may include: stage weight data; sub-task division data; participating entity data; participating role data; original evaluation data; abnormal evaluation data; corrected evaluation data; overall contribution of each participating entity; and normalized contribution ratio of each participating entity.

Images