Pre-shift meeting data depth analysis and abnormity early warning method and system

Abstract

The invention relates to the technical field of pre-class meeting abnormity early warning, in particular to a pre-class meeting data deep analysis and abnormity early warning method and system, which dynamically matches standard pre-class meeting process data according to text feature data of a pre-class meeting initiating department, and accurately judges whether a pre-class meeting process is abnormal or not in combination with obtained real-time pre-class meeting process data. If so, executing a pre-class flow adjustment operation; according to the real-time meeting-before-class flow data, corresponding meeting-before-class flow index data is collected, and a meeting-before-class index evaluation model is combined to accurately evaluate meeting-before-class indexes to generate meeting-before-class index evaluation data; and judging whether the flow index in the before-class meeting process is abnormal or not by combining an evaluation threshold scientifically set through an improved intelligent optimization algorithm, and if so, giving out early warning, thereby realizing scientific analysis of before-class meeting data and accurate early warning of before-class meeting abnormity.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of pre-shift meeting abnormal early warning, in particular to a pre-shift meeting data deep analysis and abnormal early warning method and system. BACKGROUND

[0002] As an important link in enterprise production management, the standardization and execution effect of the pre-shift meeting process directly affect the efficiency and safety of subsequent production activities. With the continuous development of intelligent coal mines, the past non-standardized pre-shift meeting process has obviously been unable to integrate into the modern coal industry information transmission and automation operation technical system.

[0003] The existing pre-shift meeting abnormal early warning technology often analyzes and warns through past experience, and is difficult to dynamically adjust the process standard according to the pre-shift meeting initiating department, resulting in poor supervision and management effect of pre-shift meeting preparatory work. At the same time, the existing technology cannot timely identify the abnormalities in the pre-shift meeting process, and is also difficult to accurately evaluate various indicators of the pre-shift meeting, resulting in poor real-time and accuracy of pre-shift meeting abnormal early warning operation. SUMMARY

[0004] In view of the problems in the related art, the present application provides a pre-shift meeting data deep analysis and abnormal early warning method and system to overcome the above technical problems existing in the prior art.

[0005] To solve the above technical problems, the present application provides the following technical scheme: a pre-shift meeting data deep analysis and abnormal early warning method, comprising the following steps: S1, obtaining text feature data of a pre-shift meeting initiating department and corresponding real-time pre-shift meeting process data; S2, performing standard process matching processing according to the text feature data and different department pre-shift meeting process data to generate standard pre-shift meeting process data; S3, analyzing the pre-shift meeting process according to the standard pre-shift meeting process data and real-time pre-shift meeting process data to generate pre-shift meeting process determination data; If it is abnormal, perform pre-shift meeting process adjustment operation; If it is normal, go to S4; S4, collecting corresponding pre-shift meeting process indicator data according to real-time pre-shift meeting process data; S5, collecting historical data to construct a pre-shift meeting indicator evaluation model, evaluating the pre-shift meeting indicators according to the pre-shift meeting process indicator data and the pre-shift meeting indicator evaluation model to generate pre-shift meeting indicator evaluation data; S6, setting an evaluation threshold, determining whether the process indicators in the pre-shift meeting process are abnormal by comparing the pre-shift meeting indicator evaluation data with the evaluation threshold to generate process indicator abnormality determination data; If it is abnormal, the process indicator abnormality determination data is pushed to the pre-shift meeting abnormality early warning platform, and an early warning is issued. If it is normal, the pre-shift meeting abnormality early warning operation is ended.

[0006] Preferably, the specific steps of obtaining the text feature data of the pre-shift meeting initiating department and the corresponding real-time pre-shift meeting process data are as follows: S11, collecting the text feature data of the pre-shift meeting initiating department and the corresponding real-time pre-shift meeting process data online through the pre-shift meeting abnormality early warning platform . ; The real-time pre-shift meeting process data includes various processes to be performed by the pre-shift meeting initiating department. The processes include, but are not limited to, any one or more of personnel check-in verification, task arrangement, process monitoring, safety emphasis, and problem summary review.

[0007] Preferably, the specific steps of generating standard pre-shift meeting process data according to the text feature data and the pre-shift meeting process data of different departments are as follows: S21, establishing a pre-shift meeting process data set of different departments , wherein, denotes the pre-shift meeting process data of the i-th department, denotes the total number of pre-shift meeting process data of different departments. S22, sequentially matching the text feature data with each pre-shift meeting process data of different departments in the pre-shift meeting process data set of different departments according to department keywords through a unified cost search algorithm, searching for pre-shift meeting process data of different departments matched with the text feature data , and generating standard pre-shift meeting process data .

[0008] By obtaining the text feature data of the pre-shift meeting initiating department and intelligently matching the standard pre-shift meeting process data with the pre-shift meeting process data of different departments through an intelligent search algorithm, the accuracy of the matching result is ensured, and the efficiency and speed of the matching process are improved.

[0009] Preferably, the specific steps of analyzing the pre-shift meeting process according to the standard pre-shift meeting process data and the real-time pre-shift meeting process data, generating pre-shift meeting process determination data, and performing pre-shift meeting process adjustment operation if it is abnormal, or entering S4 if it is normal are as follows: S31, calculating the matching degree between the standard pre-shift meeting process data and the real-time pre-shift meeting process data to obtain pre-shift meeting process matching degree data ​The calculation formula is as follows: , in, Indicates the real-time pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions, Indicates the standard pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions; S32, set the matching threshold, and Comparing the value with the matching threshold; If the pre-shift meeting process matching data If the value is greater than or equal to the matching degree threshold, it means that the real-time pre-shift meeting process data meets the standard requirements, the pre-shift meeting process determination data is output as normal, and the process goes directly to S4; Otherwise, it means that the real-time pre-shift meeting process data does not meet the standard requirements, and the output pre-shift meeting process judgment data is abnormal. It is pushed to the pre-shift meeting abnormality warning platform through the Internet of Things communication network and the pre-shift meeting process adjustment operation is executed. After adjusting the real-time pre-shift meeting process data, return to S31.

[0010] Scientifically calculate the matching degree between standard pre-shift meeting process data and real-time pre-shift meeting process data, and scientifically analyze whether the pre-shift meeting process is abnormal based on the calculation results to achieve accurate judgment of the pre-shift meeting process.

[0011] Preferably, the specific steps for collecting corresponding pre-shift meeting process indicator data based on real-time pre-shift meeting process data are as follows: S41. Collect the indicator data of each process corresponding to the real-time pre-shift meeting process data online through the pre-shift meeting abnormality warning platform to generate a pre-shift meeting process indicator data matrix as follows: , in, Indicates the first The first step in the process Class indicator data, Indicates the total number of items in the process corresponding to the real-time pre-shift meeting process data. Indicates the first The total number of categories of indicator data in the process; The indicator data includes but is not limited to personnel verification results, personnel check-in ratio, completion time of each process, task completion rate and task completion accuracy rate.

[0012] According to the real-time pre-shift meeting process data acquisition corresponding real-time pre-shift meeting process data for subsequent analysis to provide data support.

[0013] Preferably, the historical data is collected to construct a pre-shift meeting index evaluation model, and the pre-shift meeting index is evaluated according to the pre-shift meeting process index data and the pre-shift meeting index evaluation model to generate pre-shift meeting index evaluation data. The specific steps are as follows: S51, collecting the process index data and the corresponding index evaluation data in the past several pre-shift meeting processes through the pre-shift meeting abnormal early warning platform, to obtain a historical pre-shift meeting process index data set And the corresponding historical pre-shift meeting index evaluation data set , wherein, represents the process index data matrix in the past pre-shift meeting process, represents the index evaluation data matrix in the past pre-shift meeting process, represents the total number of historical pre-shift meeting process index data; S52, constructing a pre-shift meeting index evaluation model based on the historical pre-shift meeting process index data set and the corresponding historical pre-shift meeting index evaluation data set; S53, inputting the pre-shift meeting process index data in the pre-shift meeting process index data matrix into the pre-shift meeting index evaluation model to evaluate the pre-shift meeting index and generate a pre-shift meeting index evaluation data matrix As follows: , wherein, represents the index evaluation data of the th index data in the th process corresponding to the real-time pre-shift meeting process data.

[0014] The historical pre-shift meeting process index data and the corresponding historical pre-shift meeting index evaluation data are used to construct a pre-shift meeting index evaluation model, so that the model can better associate the pre-shift meeting process index data with the pre-shift meeting index evaluation data, and provide a good and reliable evaluation model for the evaluation of the pre-shift meeting index.

[0015] Preferably, an evaluation threshold is set, the pre-shift meeting index evaluation data and the evaluation threshold are used to determine whether the process index in the pre-shift meeting process is abnormal, to generate process index abnormality determination data. If it is abnormal, the process index abnormality determination data is pushed to the pre-shift meeting abnormal early warning platform and an early warning is issued. If it is normal, the specific steps of this pre-shift meeting abnormal early warning operation are as follows: S61, set evaluation threshold values corresponding to various types of index data in each process by improved intelligent optimization algorithm, obtain evaluation threshold matrix As follows: , Among them, represents the evaluation threshold value corresponding to the first type of index data in the first process in the real-time pre-shift meeting process data; S611, construct an evaluation threshold search set, set the current iteration number as , the maximum iteration number as , and the evaluation threshold search space dimension as ; Set the evaluation threshold value interval of each type of index data in each process, obtain the evaluation threshold value interval matrix As follows: , Among them, and respectively represent the lower limit and upper limit of the value of the first type of index data in the first process corresponding to the real-time pre-shift meeting process data; Take the evaluation threshold value interval matrix as the evaluation threshold value search space, randomly generate groups of evaluation threshold values in the evaluation threshold value search space, each group of evaluation threshold values corresponding to the initial position of an evaluation threshold search data in the evaluation threshold search set, obtain the initial position set of the evaluation threshold search set , wherein, represents the initial position of the evaluation threshold search data in the evaluation threshold search set; S612, evenly divide each evaluation threshold search data in the evaluation threshold search set into evaluation threshold search sub-sets, and calculate the fitness value of the evaluation threshold search data in each evaluation threshold search sub-set according to the fitness function formula, arrange the evaluation threshold search data in each evaluation threshold search sub-set in descending order of fitness value, select the evaluation threshold search data with the highest fitness value in each evaluation threshold search sub-set as the corresponding current optimal solution, obtain the first current optimal solution set , wherein, represents the current optimal solution in the evaluation threshold search sub-set; the fitness function formula is as follows: , Among them, Indicates the The evaluation threshold searches for the first The fitness value of the evaluation threshold search data, Indicates that through The evaluation threshold searches for the first The maximum misjudgment rate of whether each process indicator is abnormal is determined by the evaluation threshold corresponding to the evaluation threshold search data. Indicates the correction value; S613: The evaluation threshold search data in each evaluation threshold search subset are updated in the search space around the corresponding current optimal solution position; the position update formula is as follows: , in, Indicates the The evaluation threshold searches for the first The location after the evaluation threshold search data is updated, Indicates the The evaluation threshold searches for the first The current position of the evaluation threshold search data, Indicates the The evaluation threshold searches for the position of the current optimal solution in the subset, Indicates the impact factor of the current optimal solution on other evaluation threshold search data, Indicates a random number that obeys a uniform distribution between [0,1]; The current optimal solution in each evaluation threshold search subset is updated in the search space under the influence of the center position of the evaluation threshold search subset; the position update formula is as follows: , in, Indicates the The position after the current optimal solution in the evaluation threshold search subset is updated, Indicates the The evaluation threshold searches for the center position of the subset, represents the control factor; S614, calculating the fitness value of the evaluation threshold search data in each evaluation threshold search subset after the position is updated. If the fitness value of the evaluation threshold search data after the position is updated is greater than the original fitness value, the new position of the evaluation threshold search data is used to replace the original position; otherwise, the original position is retained; Arranging the evaluation threshold search data in each evaluation threshold search sub-set in descending order of fitness value, removing the evaluation threshold search data with the lowest fitness value in each evaluation threshold search sub-set from the corresponding evaluation threshold search sub-set, and introducing a position update strategy of the flood optimization algorithm to update the position of the removed evaluation threshold search data; the position update formula is as follows: , Wherein, represents the position of the evaluation threshold search data with the highest fitness value in the evaluation threshold search set, and both represent random numbers subject to uniform distribution between (0, 1), and represent the search upper limit and the search lower limit of the search space respectively; S615, judging whether the current iteration number is greater than or equal to the maximum iteration number , if the current iteration number is greater than or equal to the maximum iteration number , output the evaluation threshold search data with the highest fitness value to obtain the evaluation threshold matrix; otherwise, the current iteration number is increased by 1 and the process returns to S613; S62, sequentially comparing each pre-shift meeting index evaluation data in the pre-shift meeting index evaluation data matrix with the corresponding evaluation threshold in the evaluation threshold matrix; if each pre-shift meeting index evaluation data in the pre-shift meeting index evaluation data matrix is greater than or equal to the corresponding evaluation threshold in the evaluation threshold matrix, it indicates that each pre-shift meeting process index data in the pre-shift meeting process index data matrix is within a reasonable range, and the process index abnormality determination data is output as normal, the process index abnormality determination data is pushed to the pre-shift meeting abnormality early warning platform through the Internet of Things communication network for display, and the current pre-shift meeting abnormality early warning work is ended; otherwise, it indicates that there is pre-shift meeting process index data not within a reasonable range in the pre-shift meeting process index data matrix, and the process index abnormality determination data is output as abnormal, the pre-shift meeting index evaluation data less than the corresponding evaluation threshold is selected from the pre-shift meeting index evaluation data matrix, and the pre-shift meeting process index data corresponding to the selected pre-shift meeting index evaluation data is marked to generate a pre-shift meeting abnormal process index data set The process index abnormality determination data and the pre-shift meeting abnormal process index data set are pushed to a pre-shift meeting abnormality early warning platform through an Internet of Things communication network to be displayed, and an abnormality alarm is issued. wherein, represents the pre-shift meeting abnormal process index data of the total number of pre-shift meeting abnormal process index data.

[0016] By improving the intelligent optimization algorithm, the evaluation threshold values corresponding to various index data in each process are scientifically set, the optimal evaluation threshold values are accurately searched through multiple iteration optimization, a judgment standard for analyzing whether the process index is abnormal is provided, and in the algorithm process, the position update strategy of the flood optimization algorithm is introduced to improve the local search performance of the algorithm, effectively avoid the algorithm from falling into a local optimal solution, and accelerate the convergence speed of the algorithm; various indexes of the pre-shift meeting process are analyzed respectively through the pre-shift meeting index evaluation data and the corresponding evaluation threshold values, so that the accuracy of the analysis result is ensured, and the abnormal pre-shift meeting index is accurately locked.

[0017] The application also includes a pre-shift meeting data deep analysis and abnormality early warning system, which comprises a pre-shift meeting process data acquisition module, a standard pre-shift meeting process matching module, a pre-shift meeting process determination module, a pre-shift meeting process index acquisition module, a pre-shift meeting process index evaluation module and a process index abnormality determination module. The pre-shift meeting process data acquisition module acquires text feature data and corresponding real-time pre-shift meeting process data of a pre-shift meeting initiating department through a pre-shift meeting abnormality early warning platform; The standard pre-shift meeting process matching module matches the text feature data with the pre-shift meeting process data of different departments in sequence according to department keywords through a unified cost search algorithm to generate standard pre-shift meeting process data; The pre-shift meeting process determination module calculates the matching degree between the standard pre-shift meeting process data and the real-time pre-shift meeting process data to obtain pre-shift meeting process matching degree data, and compares the data with a preset matching degree threshold value, generates pre-shift meeting process determination data according to the numerical comparison result, and if it is normal, directly enters S4; otherwise, performs pre-shift meeting process adjustment work; The pre-shift meeting process index acquisition module acquires index data in each process corresponding to the real-time pre-shift meeting process data through the pre-shift meeting abnormality early warning platform to generate pre-shift meeting process index data; The pre-shift meeting process index evaluation module obtains historical pre-shift meeting process index data and corresponding historical pre-shift meeting index evaluation data by collecting process index data and corresponding index evaluation data in a plurality of past completed pre-shift meeting processes through a pre-shift meeting abnormal early warning platform, and constructs a pre-shift meeting index evaluation model. The process index abnormality determination module sets evaluation thresholds corresponding to various index data in each process by using an improved intelligent optimization algorithm, determines whether the process index in the pre-shift meeting process is abnormal in combination with the pre-shift meeting index evaluation data, generates process index abnormality determination data, and if the process index is abnormal, pushes the process index abnormality determination data to the pre-shift meeting abnormal early warning platform and issues a warning.

[0018] By the above technical solution, the present application provides a pre-shift meeting data deep analysis and abnormal early warning method and system, which has at least the following beneficial effects: 1. The present application dynamically matches standard pre-shift meeting process data according to text feature data of a pre-shift meeting initiating department, accurately determines whether the pre-shift meeting process is abnormal in combination with acquired real-time pre-shift meeting process data, and if the process is abnormal, performs pre-shift meeting process adjustment work; the pre-shift meeting index is accurately evaluated by a pre-shift meeting index evaluation model to generate pre-shift meeting index evaluation data, and the evaluation thresholds are scientifically set to determine whether the process index in the pre-shift meeting process is abnormal, and if the process is abnormal, a warning is issued to realize scientific analysis of pre-shift meeting data and accurate early warning of pre-shift meeting abnormalities.

[0019] 2. The present application matches standard pre-shift meeting process data by acquiring text feature data of a pre-shift meeting initiating department in combination with an intelligent search algorithm and different department pre-shift meeting process data, ensures the accuracy of the matching result, improves the efficiency and speed of the matching process, and scientifically calculates the matching degree between the real-time pre-shift meeting process data to scientifically analyze whether the pre-shift meeting process is abnormal to realize accurate determination of the pre-shift meeting process.

[0020] 3. The present application constructs a pre-shift meeting index evaluation model by using historical pre-shift meeting process index data and corresponding historical pre-shift meeting index evaluation data, so that the model can better associate the pre-shift meeting process index data and the pre-shift meeting index evaluation data, and provides a good and reliable evaluation model for the evaluation of the pre-shift meeting index.

[0021] 4、The present application sets the evaluation threshold of various index data in each process by the improved intelligent optimization algorithm, and through multiple iteration optimization, the optimal evaluation threshold is accurately searched, which provides a judgment standard for analyzing whether the process index is abnormal, and the flood optimization algorithm position updating strategy is introduced in the algorithm process, which improves the local search performance of the algorithm, effectively avoids the algorithm falling into local optimal solution, and speeds up the convergence speed of the algorithm. BRIEF DESCRIPTION OF DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can be obtained without creative labor.

[0023] Figure 1 The flow chart of the pre-shift meeting data depth analysis and abnormal early warning method provided by the present application; Figure 2 The module schematic diagram of the pre-shift meeting data depth analysis and abnormal early warning system provided by the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the application will be described clearly and completely in the following with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only some embodiments of the application, not all. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the application.

[0025] Embodiment one is as follows: In view of the problems that the existing pre-shift meeting abnormal early warning technology has poor supervision and management effect on pre-shift meeting front-end work, and the real-time performance and accuracy of completing pre-shift meeting abnormal early warning operation are poor. The present application provides a pre-shift meeting data depth analysis and abnormal early warning method, dynamically matches standard pre-shift meeting process data according to the text feature data of the pre-shift meeting initiating department, and accurately judges whether the pre-shift meeting process is abnormal in combination with the obtained real-time pre-shift meeting process data; through the pre-shift meeting index evaluation model, the pre-shift meeting index is accurately evaluated, the pre-shift meeting index evaluation data is generated, and whether the process index in the pre-shift meeting process is abnormal is judged in combination with the scientifically set evaluation threshold. If it is abnormal, an early warning is issued, the scientific analysis of pre-shift meeting data and the accurate early warning of pre-shift meeting abnormality are realized. Figure 1 As shown in the figure, the method comprises the following steps: S1, obtaining the text feature data of the pre-shift meeting initiating department and the corresponding real-time pre-shift meeting process data. S11, collecting the text feature data of the pre-shift meeting initiating department online through the pre-shift meeting abnormal early warning platform And the corresponding real-time pre-shift meeting process data ; Real-time pre-shift meeting process data includes various processes to be executed by the department initiating the pre-shift meeting; The process includes but is not limited to any one or more of personnel sign-in verification, task assignment, process monitoring, safety emphasis, and problem summary and review.

[0026] S2. Perform standard process matching based on text feature data and pre-shift meeting process data of different departments to generate standard pre-shift meeting process data. As a specific implementation plan of this method, the detailed plan of this step is as follows: S21. Establish a data set of pre-shift meeting processes for different departments ,in, Indicates the Pre-shift meeting process data corresponding to each department, Indicates the total number of pre-shift meeting process data for different departments; S22, through the unified cost search algorithm according to the part of the keyword text feature data Perform character feature matching with the pre-shift meeting process data of different departments in the pre-shift meeting process data set of different departments in turn, and search for the data that matches the text feature data. Match the pre-shift meeting process data of different departments, mark the data, and generate standard pre-shift meeting process data .

[0027] S3. Analyze the pre-shift meeting process based on the standard pre-shift meeting process data and the real-time pre-shift meeting process data to generate pre-shift meeting process judgment data. If it is abnormal, perform pre-shift meeting process adjustment operations; if it is normal, proceed to S4. As a specific implementation plan of this method, the detailed plan of this step is as follows: S31. Calculate the matching degree between the standard pre-shift meeting process data and the real-time pre-shift meeting process data to obtain the pre-shift meeting process matching degree data. The calculation formula is as follows: , in, Indicates the real-time pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions, Indicates the standard pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions; S32. Set the matching threshold and match the pre-shift meeting process matching data. Compare the value with the matching threshold; If the pre-shift meeting process matching data Greater than or equal to the matching degree threshold value, indicating that the real-time pre-shift meeting process data meets the standard requirements, outputting the pre-shift meeting process judgment data as normal, and directly entering S4; Otherwise, indicating that the real-time pre-shift meeting process data does not meet the standard requirements, outputting the pre-shift meeting process judgment data as abnormal, pushing to the pre-shift meeting abnormal early warning platform through the Internet of Things communication network and performing pre-shift meeting process adjustment work, adjusting the real-time pre-shift meeting process data and returning to S31.

[0028] S4, collecting pre-shift meeting process index data corresponding to real-time pre-shift meeting process data. As follows: , Among them, indicates the first class index data in the first process corresponding to the real-time pre-shift meeting process data, indicates the total number of processes corresponding to the real-time pre-shift meeting process data, indicates the total number of classes of index data in the first process corresponding to the real-time pre-shift meeting process data; The index data includes but is not limited to personnel verification results, personnel check-in proportion, completion time of each process, task completion degree, and task completion accuracy.

[0029] S5, collecting historical data to construct a pre-shift meeting index evaluation model, evaluating the pre-shift meeting index according to the pre-shift meeting process index data and the pre-shift meeting index evaluation model, generating pre-shift meeting index evaluation data, which is a specific implementation scheme of the method. The detailed scheme is as follows: S51, collecting process index data and corresponding index evaluation data in a number of pre-shift meeting processes completed in the past through the pre-shift meeting abnormal early warning platform to obtain a historical pre-shift meeting process index data set and a corresponding historical pre-shift meeting index evaluation data set Among them, indicates the process index data matrix in the first pre-shift meeting process completed in the past, indicates the index evaluation data matrix in the first pre-shift meeting process completed in the past, indicates the total number of historical pre-shift meeting process index data; S52, constructing a pre-shift meeting index evaluation model based on the historical pre-shift meeting process index data set and the corresponding historical pre-shift meeting index evaluation data set; S521, construct an initial SVM model; set the penalty coefficient of the initial SVM model to And the kernel coefficient is , and select the radial basis kernel function as the kernel function of the initial SVM model; the radial basis kernel function is as follows: , in, represents the radial basis kernel function, and Represents the eigenvector corresponding to any two sample data, Represents a vector and vector The Euclidean distance between S522. Set a training data ratio, and divide the historical pre-shift meeting process indicator dataset and the corresponding historical pre-shift meeting indicator evaluation dataset according to the training data ratio to obtain a historical pre-shift meeting process indicator training dataset, a historical pre-shift meeting indicator evaluation training dataset, a historical pre-shift meeting process indicator test dataset, and a historical pre-shift meeting indicator evaluation test dataset; S523. Set a training error threshold and a maximum number of training times, input the historical pre-shift meeting process indicator training data set as training data and the historical pre-shift meeting indicator evaluation training data set as training label data into the initial SVM model for training, and continuously adjust the initial SVM model parameters according to the training results until the training error is less than the training error threshold or the number of training times is greater than the maximum number of training times, thereby obtaining a trained SVM model; S524. Set an accuracy threshold, input the historical pre-shift meeting process indicator test data set as test data and the historical pre-shift meeting indicator evaluation test data set as test label data into the trained SVM model for testing, calculate the accuracy of the test results, and if the accuracy of the test results is greater than or equal to the accuracy threshold, obtain the pre-shift meeting indicator evaluation model; otherwise, return to S523 and retrain until the accuracy of the test results is greater than or equal to the accuracy threshold; S53. Input the pre-shift meeting process indicator data in the pre-shift meeting process indicator data matrix into the pre-shift meeting indicator evaluation model to evaluate the pre-shift meeting indicators and generate a pre-shift meeting indicator evaluation data matrix. as follows: , in, Indicates the first The first step in the process Indicator evaluation data for class indicator data.

[0030] S6, set the evaluation threshold, the pre-shift meeting index evaluation data and the evaluation threshold to determine whether the process indicators in the pre-shift meeting process are abnormal, generate process indicator abnormality determination data, if abnormal, push the process indicator abnormality determination data to the pre-shift meeting abnormality early warning platform, and issue a warning; if normal, end the pre-shift meeting abnormality early warning work, as a specific implementation scheme of the method, the detailed scheme is as follows: S61, set the evaluation threshold of each type of index data in each process by the improved intelligent optimization algorithm, and obtain the evaluation threshold matrix As follows: , Among them, represents the evaluation threshold corresponding to the type index data in the process corresponding to the real-time pre-shift meeting process data; S611, construct an evaluation threshold search set, set the current iteration number as , the maximum iteration number as , and the evaluation threshold search space dimension as ; Set the evaluation threshold value interval of each type of index data in each process, and obtain the evaluation threshold value interval matrix As follows: , Among them, and respectively represent the lower limit and upper limit of the value of the type index data in the process corresponding to the real-time pre-shift meeting process data; Take the evaluation threshold value interval matrix as the evaluation threshold search space, randomly generate groups of evaluation thresholds in the evaluation threshold search space, each group of evaluation thresholds corresponding to the initial position of an evaluation threshold search data in the evaluation threshold search set, and obtain the initial position set of the evaluation threshold search set , wherein represents the initial position of the evaluation threshold search data in the evaluation threshold search set; S612, evenly divide each evaluation threshold search data in the evaluation threshold search set into The evaluation threshold search subsets are formed, and the fitness value of the evaluation threshold search data in each evaluation threshold search subset is calculated according to the fitness function formula. The evaluation threshold search data in each evaluation threshold search subset are arranged from large to small according to the fitness value, and the evaluation threshold search data with the highest fitness value in each evaluation threshold search subset is selected as the corresponding current optimal solution to obtain the first current optimal solution set. ,in, Indicates the The current optimal solution in the subset is searched using an evaluation threshold; the fitness function formula is as follows: , in, Indicates the The evaluation threshold searches for the first The fitness value of the evaluation threshold search data, Indicates that through The evaluation threshold searches for the first The maximum misjudgment rate of whether each process indicator is abnormal is determined by the evaluation threshold corresponding to the evaluation threshold search data. Indicates the correction value; S613: The evaluation threshold search data in each evaluation threshold search subset are updated in the search space around the corresponding current optimal solution position; the position update formula is as follows: , in, Indicates the The evaluation threshold searches for the first The location after the evaluation threshold search data is updated, Indicates the The evaluation threshold searches for the first The current position of the evaluation threshold search data, Indicates the The evaluation threshold searches for the position of the current optimal solution in the subset, Indicates the impact factor of the current optimal solution on other evaluation threshold search data, Indicates a random number that obeys a uniform distribution between [0,1]; The current optimal solution in each evaluation threshold search subset is updated in the search space under the influence of the center position of the evaluation threshold search subset. The position update formula is as follows: , in, Indicates the The position after the current optimal solution in the evaluation threshold search subset is updated, representing the center position of the evaluation threshold search sub-set, representing the control factor; S614, calculate the fitness value of the evaluation threshold search data in each evaluation threshold search sub-set after position updating, if the fitness value of the evaluation threshold search data after position updating is greater than the original fitness value, replace the original position with the new position of the evaluation threshold search data; otherwise, keep the original position; Arrange the evaluation threshold search data in each evaluation threshold search sub-set in descending order of fitness value, remove the evaluation threshold search data with the lowest fitness value from the corresponding evaluation threshold search sub-set, and perform position updating on the removed evaluation threshold search data using the flood optimization algorithm position updating strategy; the position updating formula is as follows: , wherein, represents the position of the evaluation threshold search data with the highest fitness value in the evaluation threshold search set, and both represent random numbers uniformly distributed between (0, 1), and represent the search upper limit and the search lower limit of the search space, respectively; S615, judge whether the current iteration number is greater than or equal to the maximum iteration number , if the current iteration number is greater than or equal to the maximum iteration number , output the evaluation threshold search data with the highest fitness value to obtain the evaluation threshold matrix; otherwise, increase the current iteration number by 1 and return to S613; S62, perform numerical comparison between each pre-shift meeting index evaluation data in the pre-shift meeting index evaluation data matrix and the corresponding evaluation threshold in the evaluation threshold matrix; If each pre-shift meeting index evaluation data in the pre-shift meeting index evaluation data matrix is greater than or equal to the corresponding evaluation threshold in the evaluation threshold matrix, it indicates that each pre-shift meeting process index data in the pre-shift meeting process index data matrix is within a reasonable range, and the process index abnormality determination data is output as normal, the process index abnormality determination data is pushed to the pre-shift meeting abnormality early warning platform for display through the Internet of Things communication network, and the current pre-shift meeting abnormality early warning work is ended; Otherwise, it indicates that there is pre-shift meeting process index data in the pre-shift meeting process index data matrix that is not within a reasonable range, and the process index abnormality determination data For the abnormal, the pre-shift meeting index evaluation data less than the corresponding evaluation threshold is screened out from the pre-shift meeting index evaluation data matrix, and the pre-shift meeting process index data corresponding to the screened pre-shift meeting index evaluation data is data identified to generate a pre-shift meeting abnormal process index data set The process index abnormality determination data and the pre-shift meeting abnormal process index data set are pushed to the pre-shift meeting abnormal early warning platform through the Internet of Things communication network for display, and an abnormal alarm is issued. Among them, represents the pre-shift meeting abnormal process index data, represents the total number of pre-shift meeting abnormal process index data.

[0031] Embodiment two is as follows: Please refer to Figure 2 A pre-shift meeting data deep analysis and abnormal early warning system, comprising a pre-shift meeting process data acquisition module, a standard pre-shift meeting process matching module, a pre-shift meeting process determination module, a pre-shift meeting process index acquisition module, a pre-shift meeting process index evaluation module, and a process index abnormality determination module. The pre-shift meeting process data acquisition module acquires text feature data and corresponding real-time pre-shift meeting process data of the pre-shift meeting initiating department online through the pre-shift meeting abnormal early warning platform; The standard pre-shift meeting process matching module matches the text feature data with the pre-shift meeting process data of each different department according to the department keywords through a uniform cost search algorithm to generate standard pre-shift meeting process data; The pre-shift meeting process determination module calculates the matching degree between the standard pre-shift meeting process data and the real-time pre-shift meeting process data to obtain pre-shift meeting process matching degree data, and compares the data with a preset matching degree threshold value, generates pre-shift meeting process determination data according to the numerical comparison result, and if it is normal, directly enters S4; otherwise, perform pre-shift meeting process adjustment work; The pre-shift meeting process index acquisition module acquires index data in each process corresponding to the real-time pre-shift meeting process data online through the pre-shift meeting abnormal early warning platform to generate pre-shift meeting process index data; The pre-shift meeting process index evaluation module acquires process index data and corresponding index evaluation data in a number of pre-shift meeting processes completed in the past through the pre-shift meeting abnormal early warning platform to obtain historical pre-shift meeting process index data and corresponding historical pre-shift meeting index evaluation data, and constructs a pre-shift meeting index evaluation model, evaluates the pre-shift meeting index in combination with the pre-shift meeting process index data, and generates pre-shift meeting index evaluation data; The flow indicator abnormality determination module sets the evaluation threshold corresponding to various indicator data in each flow through an improved intelligent optimization algorithm, and determines whether the flow indicator in the pre-shift meeting process is abnormal in combination with the pre-shift meeting indicator evaluation data, generates flow indicator abnormality determination data, and if it is abnormal, pushes the flow indicator abnormality determination data to the pre-shift meeting abnormality early warning platform and issues a warning.

[0032] Those skilled in the art can understand that all or part of the steps in the above-mentioned embodiment methods can be completed by programs instructing related hardware, therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a computer program product in the form of being implemented on one or more computer usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer usable program codes.

[0033] In the description of the present specification, the description referring to the terms "one embodiment", "an example", "a specific example" and the like means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are contained in at least one embodiment or example of the invention. In the present specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0034] The above-disclosed preferred embodiments of the invention are only used to help explain the invention. The preferred embodiments do not describe all the details, nor limit the invention to the specific embodiments. Obviously, many modifications and variations can be made according to the content of the present specification. The present specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can well understand and utilize the invention.

Claims

1. A method for in-depth analysis and abnormal warning of pre-shift meeting data, characterized by: The steps include: S1. Obtain text feature data of the department initiating the pre-shift meeting and the corresponding real-time pre-shift meeting process data; S2. Perform standard process matching processing based on the text feature data and the pre-shift meeting process data of different departments to generate standard pre-shift meeting process data; S3. Analyze the pre-shift meeting process based on the standard pre-shift meeting process data and the real-time pre-shift meeting process data to generate pre-shift meeting process determination data; If there is an abnormality, the pre-shift process adjustment work will be carried out; If normal, go to S4; S4. Collect corresponding pre-shift meeting process indicator data based on real-time pre-shift meeting process data; S5. Collect historical data to build a pre-shift meeting indicator evaluation model, evaluate the pre-shift meeting indicators based on the pre-shift meeting process indicator data and the pre-shift meeting indicator evaluation model, and generate pre-shift meeting indicator evaluation data; S6. Setting an evaluation threshold, comparing the pre-shift meeting indicator evaluation data with the evaluation threshold to determine whether a process indicator during the pre-shift meeting is abnormal, and generating process indicator abnormality determination data; If it is abnormal, the abnormal determination data of the process indicator will be pushed to the abnormal warning platform before the shift meeting and an early warning will be issued; If it is normal, the abnormal warning operation before this shift will be completed.

2. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 1 is characterized in that: The acquisition of text feature data of the department initiating the pre-shift meeting and the corresponding real-time pre-shift meeting process data includes: Collect text feature data of the department initiating the pre-shift meeting online through the pre-shift meeting abnormal warning platform And the corresponding real-time pre-shift meeting process data .

3. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 2 is characterized in that: The S2 comprises the following steps: S21. Establish a data set for pre-shift meeting processes in different departments ,in, Indicates the Pre-shift meeting process data corresponding to each department, Indicates the total number of pre-shift meeting process data for different departments; S22, using a unified cost search algorithm to search the text feature data according to the department keywords Sequentially perform character feature matching with the pre-shift meeting process data of each different department in the pre-shift meeting process data set of the different departments, and search for the text feature data that matches the text feature data. Match the pre-shift meeting process data of different departments, mark the data, and generate standard pre-shift meeting process data .

4. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 3 is characterized in that: The S3 includes the following steps: S31, calculate the matching degree between the standard pre-shift meeting process data and the real-time pre-shift meeting process data, and obtain the pre-shift meeting process matching degree data. The calculation formula is as follows: ， in, Indicates the real-time pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions, Indicates the standard pre-shift meeting process data corresponding to The eigenvector in The values ​​in the dimensions; S32, set the matching threshold, and Performing a numerical comparison with the matching threshold; If the pre-shift meeting process matching data If the value is greater than or equal to the matching threshold, it means that the real-time pre-shift meeting process data meets the standard requirements, the pre-shift meeting process determination data is output as normal, and the process goes directly to S4; Otherwise, it means that the real-time pre-shift meeting process data does not meet the standard requirements, and the output pre-shift meeting process judgment data is abnormal. It is pushed to the pre-shift meeting abnormality warning platform through the Internet of Things communication network and the pre-shift meeting process adjustment operation is executed. After adjusting the real-time pre-shift meeting process data, return to S31.

5. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 4 is characterized in that: The collected corresponding pre-shift meeting process indicator data includes: The pre-shift meeting abnormality warning platform collects the indicator data of each process corresponding to the real-time pre-shift meeting process data online to generate the pre-shift meeting process indicator data matrix .

6. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 5 is characterized in that: The S5 comprises the following steps: S51. Collect the process indicator data and corresponding indicator evaluation data of several previously completed pre-shift meetings online through the pre-shift meeting abnormality warning platform to obtain the historical pre-shift meeting process indicator data set. And the corresponding historical pre-class meeting indicator evaluation dataset ,in, Indicates the completed The process indicator data matrix during the pre-shift meeting, Indicates the completed The indicator evaluation data matrix during the pre-shift meeting, Indicates the total number of historical pre-shift meeting process indicator data; S52, constructing a pre-shift meeting indicator evaluation model based on the historical pre-shift meeting process indicator dataset and the corresponding historical pre-shift meeting indicator evaluation dataset; S53, inputting the pre-shift meeting process indicator data in the pre-shift meeting process indicator data matrix into the pre-shift meeting indicator evaluation model to evaluate the pre-shift meeting indicators, and generating a pre-shift meeting indicator evaluation data matrix .

7. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 6 is characterized in that: The S6 comprises the following steps: S61. Use the improved intelligent optimization algorithm to set the evaluation thresholds corresponding to various indicator data in each process, and obtain the evaluation threshold matrix ; S62, the pre-shift meeting indicator evaluation data matrix The evaluation data of each pre-shift meeting indicator in the evaluation threshold matrix are numerically compared with the corresponding evaluation threshold in turn; If the pre-shift meeting indicator evaluation data matrix If the evaluation data of each pre-shift meeting indicator in the matrix are greater than or equal to the corresponding evaluation threshold in the evaluation threshold matrix, it means that the process indicator data of each pre-shift meeting in the pre-shift meeting process indicator data matrix are within a reasonable range, and the process indicator abnormality judgment data is output. If the result is normal, the abnormality determination data of the process indicators is pushed to the abnormality warning platform before the shift meeting through the Internet of Things communication network for display, and the abnormality warning operation before the shift meeting is ended; Otherwise, it means that the pre-shift meeting process indicator data matrix contains pre-shift meeting process indicator data that is not within a reasonable range, and the process indicator abnormality judgment data is output. If it is abnormal, the pre-shift meeting indicator evaluation data that is less than the corresponding evaluation threshold is screened out from the pre-shift meeting indicator evaluation data matrix, and the pre-shift meeting process indicator data corresponding to the screened pre-shift meeting indicator evaluation data is marked to generate a pre-shift meeting abnormal process indicator data set. , pushing the process indicator abnormality determination data and the abnormal process indicator data set of the pre-shift meeting to the pre-shift meeting abnormality warning platform for display through the Internet of Things communication network, and issuing an abnormality alarm; in, Indicates the Before each shift, abnormal process indicator data will be collected. Indicates the total number of abnormal process indicator data in the pre-shift meeting.

8. The method for in-depth analysis and abnormal warning of pre-shift meeting data according to claim 7 is characterized in that: The S61 includes the following steps: S611, build an evaluation threshold search set, set the current number of iterations to , the maximum number of iterations is And the evaluation threshold search space dimension is ; Set the evaluation threshold value intervals corresponding to various indicator data in each process to obtain the evaluation threshold value interval matrix as follows: ， in, and Respectively represent the first The first step in the process The lower and upper limits of the values ​​corresponding to the class indicator data; The evaluation threshold value interval matrix is ​​used as the evaluation threshold search space, and a random Each group of evaluation thresholds corresponds to the initial position of an evaluation threshold search data in the evaluation threshold search set, and the initial position set of the evaluation threshold search set is obtained. ,in, Indicates the first An evaluation threshold searches for the initial position of the data; S612: Evenly divide each evaluation threshold search data in the evaluation threshold search set into The evaluation threshold search subsets are formed, and the fitness value of the evaluation threshold search data in each evaluation threshold search subset is calculated according to the fitness function formula. The evaluation threshold search data in each evaluation threshold search subset are arranged from large to small according to the fitness value, and the evaluation threshold search data with the highest fitness value in each evaluation threshold search subset is selected as the corresponding current optimal solution to obtain the first current optimal solution set. ,in, Indicates the The current optimal solution in the subset is searched using an evaluation threshold; the fitness function formula is as follows: ， in, Indicates the The evaluation threshold searches for the first The fitness value of the evaluation threshold search data, Indicates that through The evaluation threshold searches for the first The maximum misjudgment rate of whether each process indicator is abnormal is determined by the evaluation threshold corresponding to the evaluation threshold search data. Indicates the correction value; S613: The evaluation threshold search data in each evaluation threshold search subset are updated in the search space around the corresponding current optimal solution position; the position update formula is as follows: ， in, Indicates the The evaluation threshold searches for the first The location after the evaluation threshold search data is updated, Indicates the The evaluation threshold searches for the first The current position of the evaluation threshold search data, Indicates the The evaluation threshold searches for the position of the current optimal solution in the subset, Indicates the impact factor of the current optimal solution on other evaluation threshold search data, Indicates a random number that obeys a uniform distribution between [0,1]; The current optimal solution in each evaluation threshold search subset is updated in the search space under the influence of the center position of the evaluation threshold search subset; the position update formula is as follows: ， in, Indicates the The position after the current optimal solution in the evaluation threshold search subset is updated, Indicates the The evaluation threshold searches for the center position of the subset, represents the control factor; S614, calculating the fitness value of the evaluation threshold search data in each evaluation threshold search subset after the position is updated. If the fitness value of the evaluation threshold search data after the position is updated is greater than the original fitness value, the new position of the evaluation threshold search data is used to replace the original position; otherwise, the original position is retained; The evaluation threshold search data in each evaluation threshold search subset are arranged from large to small according to the fitness value, and the evaluation threshold search data with the lowest fitness value in each evaluation threshold search subset is removed from the corresponding evaluation threshold search subset. The flood optimization algorithm position update strategy is introduced to update the position of the removed evaluation threshold search data; the position update formula is as follows: ， in, Indicates the position of the evaluation threshold search data with the highest fitness value in the evaluation threshold search set, and Both represent random numbers that are uniformly distributed between (0,1). and Respectively represent the search upper limit and search lower limit of the search space; S615: Determine the current number of iterations Is it greater than or equal to the maximum number of iterations? , if the current number of iterations Greater than or equal to the maximum number of iterations , then the evaluation threshold search data with the highest fitness value is output to obtain the evaluation threshold matrix; otherwise, the current number of iterations Add 1 and return to S613.

9. A system for implementing the method for in-depth analysis of pre-shift meeting data and abnormality warning according to any one of claims 1 to 8.

Images