Case handling center full-process management and control method based on big data

Abstract

The invention relates to the technical field of big data case handling management and control, and discloses a case handling center full-process management and control method based on big data. The method comprises the following steps: performing full-process management and control on a case handling center based on a corresponding initial data matrix in each business stage to obtain a data matrix after management and control in each stage; after the target stage is determined, performing management and control operation according to the managed and controlled data matrix, and measuring an actual data matrix through a data acquisition system; when the stage difference between the target stage and the actual data matrix is greater than or equal to a preset threshold value, determining a data deviation matrix according to the data matrix controlled by the target stage and an original data matrix corresponding to the actual data matrix; adjusting the controlled data matrix in the target stage based on the data deviation matrix to obtain an adjusted data matrix; and taking the adjusted data matrix as a new managed and controlled data matrix, and repeatedly executing the management and control operation until the phase difference is smaller than a preset threshold value. According to the method, through generation, comparison and adjustment of the data matrix, regular and orderly promotion of the case handling process is promoted.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of big data case handling management and control, in particular to a case handling center whole-process management and control method based on big data. BACKGROUND

[0002] In modern judicial practice, as the core hub of case handling, the standardization and efficiency of the process of the case handling center directly affect the quality and progress of the case handling. The traditional process management and control of the case handling center relies on manual operation and experience judgment, and the data collection, analysis and transmission of each business stage are often in a relatively independent state, lacking systematic integration and linkage. With the continuous increase in the number of cases and the increasing complexity of cases, the disadvantages of the traditional management and control mode gradually appear. For example, in the case acceptance stage, the entry of relevant information may be missing or incorrect, and these problems are difficult to be discovered in time in the subsequent investigation, examination and prosecution stages, resulting in obstruction of the process. In the evidence management link, due to the lack of real-time tracking and monitoring mechanism, the flow path of the evidence is not clear, and the risk of evidence loss or damage is easy to occur, which further affects the fair handling of the case.

[0003] The data sharing between different business stages has barriers, making it difficult for case handling personnel to fully grasp the overall situation of the case. When it is necessary to review or adjust the work of a certain stage, due to the lack of accurate and complete historical data support, it is often necessary to rely on subjective judgment, and it is difficult to form a scientific and reasonable decision. At the same time, the traditional management and control mode reacts slowly to the deviation in the process, and cannot take effective corrective measures in time, resulting in the accumulation of deviation and affecting the smoothness of the whole case handling process. Under such background, how to use advanced technical means to break through the limitations of the traditional management and control mode and realize the precise and dynamic management and control of the whole process of the case handling center has become a problem to be solved in the current judicial field. The rise of big data technology provides new possibilities for the solution of this problem. Through the integration and analysis of the massive data generated in the case handling process, it is expected to build a more efficient and standardized process management and control system, but how to deeply integrate big data technology with the actual business process of the case handling center still needs further exploration and practice. SUMMARY

[0004] The purpose of the present application is to provide a case handling center whole-process management and control method based on big data to solve the problems raised in the background art.

[0005] To achieve the above purpose, the present application provides a case handling center whole-process management and control method based on big data, which comprises: controlling the case-handling center based on the initial data matrix corresponding to each of the business stages to obtain a post-control data matrix of the case-handling center at each of the business stages; determining a target stage required to be controlled by the case-handling center, performing a process control operation based on a post-control data matrix corresponding to the target stage, and measuring an actual data matrix of the case-handling center through a data acquisition system; in a case where a stage difference between the target stage and the actual data matrix is greater than or equal to a preset difference threshold, determining a data deviation matrix according to the post-control data matrix corresponding to the target stage and an original data matrix corresponding to the actual data matrix; adjusting the post-control data matrix corresponding to the target stage based on the data deviation matrix to obtain an adjusted data matrix corresponding to the target stage; taking the adjusted data matrix corresponding to the target stage as a new post-control data matrix corresponding to the target stage, and returning to the step of performing the process control operation based on the post-control data matrix corresponding to the target stage until the stage difference between the target stage and the actual data matrix is less than the preset difference threshold.

[0006] Preferably, the controlling the case-handling center based on the initial data matrix corresponding to each of the business stages to obtain a post-control data matrix of the case-handling center at each of the business stages comprises: for any of the business stages, performing the process control operation of the case-handling center based on the initial data matrix corresponding to the business stage, and obtaining process data of the case-handling center by using a data acquisition system, and extracting an actual data matrix from the process data; determining a process consistency variance according to the actual data matrix; in a case where the process consistency variance is greater than or equal to a preset variance threshold, determining an actual data matrix corresponding to the actual data matrix based on a pre-constructed stage data relationship model; the stage data relationship model represents the corresponding relationship between the business stage and the data; iterating the actual data matrix by using a data matrix iteration model to obtain an iterated data matrix corresponding to the actual data matrix; taking the iterated data matrix as the initial data matrix, and returning to the step of performing the process control operation of the case-handling center based on the initial data matrix corresponding to the business stage until the process consistency variance is less than the preset variance threshold; taking the iterated data matrix as the post-control data matrix.

[0007] Preferably, the actual data matrix is iterated by using the data matrix iteration model to obtain an iterated data matrix corresponding to the actual data matrix, comprising: a deviation matrix of a current iteration round is determined according to the actual data matrix and the initial data matrix; a calibration weight matrix corresponding to the current iteration round is obtained, and the deviation matrix of the current iteration round is adjusted by using the calibration weight matrix corresponding to the current iteration round to obtain adjustment information of the initial data matrix for the current iteration round; the initial data matrix is adjusted by using the adjustment information of the initial data matrix for the current iteration round to obtain the iterated data matrix.

[0008] Preferably, the calibration weight matrix comprises calibration weights for each process node in the case handling center; and the calibration weight matrix corresponding to the current iteration round is obtained, comprising: obtaining deviation change direction information, deviation change amplitude information and deviation change trend information corresponding to each process node in the current iteration round; generating an iteration state vector corresponding to each process node based on the deviation change direction information, the deviation change amplitude information and the deviation change trend information corresponding to each process node; querying calibration weights matched with each process node in a pre-constructed calibration weight table by taking the iteration state vector corresponding to each process node as an index; generating the calibration weight matrix corresponding to the current iteration round based on the calibration weights matched with each process node.

[0009] Preferably, after the step of performing full-process management and control on the case handling center based on the initial data matrix corresponding to each business stage to obtain a managed data matrix of the case handling center at each business stage, the method further comprises: determining a data fitting coefficient matrix corresponding to each business stage according to the initial data matrix and the managed data matrix corresponding to each business stage; constructing a mapping model according to each business stage and the data fitting coefficient matrix corresponding to each business stage; the mapping model represents a mapping relationship between the business stage and the data fitting coefficient matrix.

[0010] Preferably, the managed data matrix corresponding to the target stage is adjusted based on the data deviation matrix to obtain an adjusted data matrix corresponding to the target stage, comprising: determining a data fitting coefficient matrix for the target stage based on the mapping model; Adjust the data deviation matrix by using the data fitting coefficient matrix to obtain adjustment information of the post-control data matrix corresponding to the target stage; Adjust the post-control data matrix corresponding to the target stage by using the adjustment information to obtain an adjusted data matrix corresponding to the target stage.

[0011] Preferably, the actual data matrix is extracted from the process data, including normalizing the process data, mapping the attribute values of each process node to the 0-1 interval, and arranging the nodes in topological order to form a two-dimensional matrix structure.

[0012] Preferably, the process consistency variance determination method includes calculating the difference between the data value of each process node and the data mean, squaring the difference, and taking the arithmetic mean of all node square values as the process consistency variance.

[0013] Preferably, the pre-constructed calibration weight table is formed by the following steps: collecting deviation change samples of each process node in historical process control, classifying and labeling the direction, amplitude and trend information in the samples, counting the optimal calibration weight corresponding to each class of state vector, and establishing a mapping relationship table between state vectors and calibration weights.

[0014] Preferably, the step of constructing a mapping model includes: taking the business stage as the independent variable and the corresponding data fitting coefficient matrix as the dependent variable, using a multiple linear regression method to fit the functional relationship between the independent variable and the dependent variable, and forming the mapping model.

[0015] Compared with the prior art, the beneficial effects of the present application are: The case handling center full-process control method based on big data can realize seamless connection and real-time linkage of data in each business stage. In each business stage, the case handling center is controlled by the initial data matrix to obtain a post-control data matrix, so that the control of each stage has a clear data benchmark, avoiding the randomness of control caused by data ambiguity in the traditional mode. When the target stage is determined, the process control operation is performed according to the corresponding post-control data matrix, and combined with the actual data matrix measured by the data acquisition system, the differences between the target stage and the actual situation can be found in time. This difference detection mechanism can sensitively capture various problems that occur in the process of advancing the process, whether it is data entry deviation or operation link failure, which can be identified at the first time. When the stage difference reaches or exceeds the preset threshold, the post-control data matrix is adjusted by the data deviation matrix, which can make the control benchmark consistent with the actual situation. This dynamic adjustment mechanism can flexibly cope with various changes in the case handling process, ensuring that the process control always matches the actual business needs.

[0016] After multiple cycle adjustments, until the stage difference is less than a preset threshold, so that each business stage of the case handling center can be promoted under accurate control. The data of each stage can form an organic whole, mutually confirm and support each other, eliminate the data island phenomenon, and enable the case handling personnel to comprehensively and accurately master the case information. This control method can timely correct the deviation in the process, avoid the accumulation of problems, keep the case handling process in a standardized and smooth state at all times, reduce the time waste and resource loss caused by the unsmooth process, and enable the case handling to proceed at a reasonable pace. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 A working principle diagram of the case handling center whole-process control method based on big data; Figure 2 A flowchart of the data matrix iterative model; Figure 3 A flowchart of the calibration weight matrix generation method; Figure 4 A flowchart of the target stage data matrix adjustment method. DETAILED DESCRIPTION

[0018] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0019] Please refer to Figures 1-4 The present application provides a whole-process control method for a case handling center based on big data, which comprises.

[0020] In each business stage, based on the initial data matrix corresponding to each business stage, the whole-process control of the case handling center is performed to obtain the controlled data matrix of the case handling center in each business stage; Determine the target stage required for control of the case handling center, perform process control operation based on the controlled data matrix corresponding to the target stage, and measure the actual data matrix of the case handling center through the data acquisition system; In the case where the stage difference between the target stage and the actual data matrix is greater than or equal to a preset difference threshold, determine the data deviation matrix according to the controlled data matrix corresponding to the target stage and the original data matrix corresponding to the actual data matrix; Based on the data bias matrix, the post-control data matrix corresponding to the target stage is adjusted to obtain an adjusted data matrix corresponding to the target stage; The adjusted data matrix corresponding to the target stage is taken as a new post-control data matrix corresponding to the target stage, and the step of performing process control operation based on the post-control data matrix corresponding to the target stage is returned until the stage difference between the target stage and the actual data matrix is less than the preset difference threshold.

[0021] In the process of the whole-process management and control of the case handling center based on big data, for any business stage, based on the initial data matrix corresponding to the business stage, the process control operation of the case handling center is performed, and the process data of the case handling center is obtained by using the data acquisition system, and the actual data matrix is extracted from the process data. For example, assuming that the business stage is the case acceptance stage, the initial data matrix includes data dimensions such as case type, acceptance time, and acceptance personnel and corresponding numerical values. When performing the process control operation, information input, material review and other operations are performed according to the preset case acceptance specification, and the data acquisition system synchronously records various data in the whole process, such as actual input case information, review time, review result, etc. When extracting the actual data matrix from these process data, the process data needs to be normalized to map the attribute values of each process node to the 0-1 interval, and then arranged in the order of node topology to form a two-dimensional matrix structure. Like the business proficiency of the acceptance personnel, which is originally presented in percentage, it is converted to a value between 0 and 1 after normalization, and then combined with the normalized data of other nodes in order to form the actual data matrix.

[0022] According to the actual data matrix, the process consistency variance is determined. The determination method of the process consistency variance is to calculate the difference between the data value of each process node and the data mean, square the difference, and take the arithmetic mean of all node square values. Taking the actual data matrix of the case acceptance stage as an example, assuming that there are 5 process nodes, and their data values are 0.2, 0.3, 0.5, 0.4, and 0.6 after normalization, first calculate the mean of these data as (0.2+0.3+0.5+0.4+0.6) ÷ 5 = 0.4, then the difference between each node data value and the mean is -0.2, -0.1, 0.1, 0, and 0.2, and the square of the difference is 0.04, 0.01, 0.01, 0, and 0.04. Take the arithmetic mean of these square values, i.e. (0.04+0.01+0.01+0+0.04) ÷ 5 = 0.02, and this 0.02 is the process consistency variance corresponding to the actual data matrix.

[0023] In the case that the process consistency variance is greater than or equal to a preset variance threshold, an actual data matrix corresponding to the actual data matrix is determined based on a pre-constructed stage data relationship model, wherein the stage data relationship model represents a corresponding relationship between a business stage and data. Assuming that the preset variance threshold is 0.015, and the process consistency variance 0.02 calculated above is greater than the threshold, the stage data relationship model is called at this time. The model is constructed by analyzing a large number of historical case acceptance stage data, summarizing the correlation between different data dimensions in the case acceptance stage, and summarizing the data change rule. The actual data matrix extracted at the present time is analyzed by the model to further clarify its specific characteristics and the corresponding actual data matrix in the business stage.

[0024] The actual data matrix is iterated by using the data matrix iteration model to obtain an iteration data matrix corresponding to the actual data matrix. Specifically, according to the actual data matrix and the initial data matrix, a deviation matrix of the current iteration round is determined. For example, the value corresponding to the case acceptance time in the initial data matrix is 0.3, and the value in the actual data matrix is 0.5, so the value at this position in the deviation matrix is 0.5-0.3=0.2, and so on to calculate the entire deviation matrix.

[0025] The calibration weight matrix corresponding to the current iteration round is obtained, and the deviation matrix of the current iteration round is adjusted by using the calibration weight matrix to obtain the adjustment information of the current iteration round for the initial data matrix. The calibration weight matrix includes calibration weights for each process node in the case handling center. When obtaining the matrix, first, the deviation change direction information, the deviation change amplitude information, and the deviation change trend information corresponding to each process node in the current iteration round are obtained. For example, the deviation of a certain process node increases compared with the last iteration, which is the deviation change direction information; the deviation increases by 0.1, which is the deviation change amplitude information; the deviation has been increasing for three consecutive iterations, which is the deviation change trend information. Based on these information, an iteration state vector corresponding to each process node is generated, and each iteration state vector is used as an index to query the calibration weight matching each process node in the pre-constructed calibration weight table, and then the calibration weight matrix corresponding to the current iteration round is generated.

[0026] The initial data matrix is adjusted by using the adjustment information of the current iteration round for the initial data matrix to obtain the iteration data matrix. For example, according to the adjustment information, the value of a certain process node in the initial data matrix is adjusted from 0.3 to 0.35, and other nodes are also modified according to the corresponding adjustment information to form the iteration data matrix.

[0027] The data matrix after iteration is taken as the initial data matrix, and the initial data matrix corresponding to the business stage is returned to perform the steps of the process management and control operation of the case handling center. The above process is repeated until the process consistency variance is less than the preset variance threshold. When the process consistency variance is less than the preset variance threshold, the data matrix after iteration is taken as the data matrix after management and control.

[0028] Suppose that after multiple iterations, the process consistency variance is reduced to 0.01, which is less than the preset variance threshold 0.015. At this time, the data matrix after iteration becomes the data matrix after management and control under the business stage, and the whole process management and control process from the initial data matrix to the data matrix after management and control under the business stage is completed. In the whole process, each iteration is an optimization of the data matrix. By continuously reducing the process consistency variance, the data matrix after management and control can more accurately reflect the actual situation of the business stage, thereby realizing effective management and control of the business stage of the case handling center.

[0029] For other business stages, such as the case investigation stage and the case trial stage, the same method is used for processing. Each business stage has its corresponding initial data matrix. Through a series of steps such as performing process management and control operations, extracting actual data matrices, calculating process consistency variances, and iterating data matrices, the data matrix after management and control under each business stage is obtained, ensuring that the process of the case handling center in each business stage can be accurately and effectively controlled, making the whole case handling process more standardized and orderly. In the case investigation stage, the initial data matrix may contain data such as investigation methods, investigation progress, and information of persons involved in the case. After the above process, the data matrix after management and control can accurately reflect the actual progress and data status of the investigation stage, providing reliable data support for subsequent case handling. Similarly, in the case trial stage, the data matrix after management and control obtained by this method can clearly present the key data in the trial process, ensuring the smooth progress of the trial work.

[0030] In the multiple iteration process, the data matrix iteration model continuously adjusts the initial data matrix according to the actual situation, so that the matrix after each iteration is more in line with the actual needs of the business stage. The introduction of the calibration weight matrix takes into account the specific deviation of each process node, making the adjustment process more targeted and improving the accuracy and efficiency of the data matrix iteration. The stage data relationship model provides a basis for determining the actual data matrix, ensuring that when the process consistency variance is large, the data matrix matching the business stage can be accurately found.

[0031] Embodiment 2: The process of iterating the actual data matrix to obtain the data matrix after iteration corresponding to the actual data matrix using the data matrix iteration model can be divided into the following steps: According to the actual data matrix and the initial data matrix, the deviation matrix of the current iteration round is determined. For example, in the case of a trial business stage, the initial data matrix contains the benchmark values of indicators such as trial duration, participants, and audio integrity, and the actual data matrix is the actual operation data of this stage obtained through the data acquisition system. If the value of the “trial duration” indicator in the initial data matrix is 2 hours (normalized to 0.5), and the value of this indicator in the actual data matrix is 2.5 hours (normalized to 0.625), then the corresponding value of this indicator in the deviation matrix is 0.125; if the value of the “audio integrity” indicator in the initial data matrix is 0.9, and the value of this indicator in the actual data matrix is 0.85, then the corresponding value of this indicator in the deviation matrix is -0.05. In this way, the deviation values of all indicators are calculated to form the deviation matrix of the current iteration round.

[0032] The calibration weight matrix corresponding to the current iteration round is obtained, and the deviation matrix of the current iteration round is adjusted using the calibration weight matrix corresponding to the current iteration round to obtain the adjustment information of the initial data matrix for the current iteration round. The calibration weight matrix includes calibration weights for each process node in the case handling center, and the matrix is obtained through the following steps: The deviation change direction information, deviation change amplitude information, and deviation change trend information corresponding to each process node in the current iteration round are obtained. Taking the “evidence entry” process node as an example, if the deviation value of the last round is 0.08 and the deviation value of the current round is 0.1, the deviation change direction is positive; the deviation change amplitude is 0.02; if the deviation values of the last three rounds are 0.04, 0.08, and 0.1 respectively, the deviation change trend is continuously increasing. For the “suspect information checking” node, if the deviation value of the last round is -0.05 and the deviation value of the current round is -0.03, the deviation change direction is negative and decreasing; the deviation change amplitude is 0.02; the deviation values of the last three rounds are -0.07, -0.05, and -0.03 respectively, and the deviation change trend is continuously approaching zero.

[0033] Based on the deviation change direction information, deviation change amplitude information, and deviation change trend information corresponding to each process node, an iteration state vector corresponding to each process node is generated. For example, the deviation change direction of the “evidence entry” node is positive, the amplitude is 0.02, and the trend is continuously increasing, which can be converted into a vector (1, 0.02, 1), where 1 represents positive, 0.02 is the amplitude value, and 1 represents the trend of increasing; the deviation change direction of the “suspect information checking” node is negative and decreasing, the amplitude is 0.02, and the trend is approaching zero, which can be converted into a vector (-1, 0.02, 0), where -1 represents negative, 0.02 is the amplitude value, and 0 represents the trend of approaching zero.

[0034] The iteration state vector corresponding to each process node is taken as an index to query the calibration weight matched with each process node in the pre-constructed calibration weight table. The pre-constructed calibration weight table is established by collecting deviation change samples of each process node in historical process management and control, classifying and labeling the direction, amplitude and trend information in the samples, and counting the optimal calibration weight corresponding to each class of state vector. If the iteration state vector (1, 0.02, 1) of the "evidence entry" node corresponds to the optimal calibration weight 0.8 in the historical samples, then the calibration weight of this node in this round is 0.8; the iteration state vector (-1, 0.02, 0) of the "suspect information checking" node corresponds to the historical optimal calibration weight 0.5, and the calibration weight of this node in this round is 0.5. The calibration weights of all process nodes are arranged in order to generate the calibration weight matrix corresponding to the current iteration round.

[0035] The deviation matrix is adjusted by using the calibration weight matrix corresponding to the current iteration round to obtain adjustment information. For example, the deviation value of the "evidence entry" node in the deviation matrix is 0.1, and the calibration weight is 0.8, so the adjustment value of the node in the adjustment information is 0.1x0.8=0.08; the deviation value of the "suspect information checking" node is -0.03, and the calibration weight is 0.5, so the adjustment value is -0.03x0.5=-0.015. The adjustment values of all nodes together constitute the adjustment information of the initial data matrix for the current iteration round.

[0036] The initial data matrix is adjusted by using the adjustment information of the initial data matrix for the current iteration round to obtain the iteration data matrix. The value of the "evidence entry" index in the initial data matrix is 0.7, which is adjusted to 0.78 according to the adjustment information 0.08; the initial value of the "suspect information checking" index is 0.8, which is adjusted to 0.785 according to the adjustment information -0.015. In the same way, the values of all indexes in the initial data matrix are adjusted to form the iteration data matrix.

[0037] After completing this iteration, the iteration data matrix is taken as a new initial data matrix, and the step of performing the case handling center process management and control operation based on the initial data matrix corresponding to the business stage is returned to reacquire the actual data matrix and calculate the process consistency variance. If the process consistency variance is still greater than or equal to the preset variance threshold, the above iteration process is repeated until the process consistency variance is less than the preset variance threshold. In multiple iterations, the calibration weight matrix of each round is dynamically generated according to the deviation change of each process node in the round, so that the adjustment information is more in line with the actual data fluctuation characteristics, the deviation between the actual data matrix and the initial data matrix is gradually reduced, and finally the iteration data matrix meeting the management and control requirements is obtained.

[0038] Taking a complete iteration cycle as an example, assuming that in the case transfer stage, the process consistency variance of the initial data matrix after the first iteration is 0.05 (the preset variance threshold is 0.03), the second round of iteration is required. In the second round of iteration, the bias matrix is calculated according to the new actual data matrix and the data matrix after the first iteration, the calibration weight matrix is regenerated according to the change direction, amplitude and trend of the bias of each node in the second round, the adjustment information is updated, and then the data matrix after the second iteration is obtained. If the process consistency variance at this time is reduced to 0.02, which is less than the preset threshold, the iteration is stopped, and the data matrix after the iteration is the data matrix after management and control of the business stage.

[0039] In the entire iteration process, the process nodes of different business stages are different, and the calculation of the bias matrix and the calibration weight matrix will also be adjusted accordingly. For example, in the case filing stage, the process nodes may include "reporting information registration", "jurisdiction authority audit", "case filing approval" and the like, the bias change characteristics of each node are different, and the generated calibration weight is also different, so as to ensure that the iteration process can adapt to the management and control requirements of different business stages.

[0040] In example 3, after the step of performing whole-process management and control on the case handling center based on the initial data matrix corresponding to each business stage to obtain the data matrix after management and control of the case handling center in each business stage, the data fitting coefficient matrix corresponding to each business stage is determined according to the initial data matrix and the data matrix after management and control corresponding to each business stage. For example, in the case acceptance stage, the initial data matrix includes data dimensions such as reporting time, completeness of reporter information, case type classification, and the data matrix after management and control is the corresponding data after multiple rounds of iteration optimization. The determination of the data fitting coefficient matrix needs to be obtained by calculating the mapping relationship of each corresponding element in the initial data matrix and the data matrix after management and control. Assuming that the value of the "reporting time" dimension in the initial data matrix is 0.3 (after normalization), and the value of this dimension in the data matrix after management and control is 0.35, then the fitting coefficient of this dimension can be calculated as 1.12 through the difference ratio of the two and the historical data correlation; the initial value of the "completeness of reporter information" dimension is 0.8, and the value after management and control is 0.82, and the fitting coefficient is calculated as 1.03. In this way, the data fitting coefficient matrix of the case acceptance stage is formed. For the case investigation stage, the initial data matrix may involve dimensions such as investigation means selection, evidence collection progress, and investigation range of involved personnel, and the data matrix after management and control is the optimized data after process consistency verification. Similarly, the data fitting coefficient matrix of this stage is calculated through the mapping relationship of the initial value and the value after management and control of each dimension, such as the initial value of the "evidence collection progress" dimension is 0.5, and the value after management and control is 0.58, and the fitting coefficient is 1.16.

[0041] According to the business stages and the data fitting coefficient matrix corresponding to each business stage, a mapping model is constructed, which represents the mapping relationship between the business stages and the data fitting coefficient matrix. When constructing the mapping model, the business stages are taken as the independent variables, and the corresponding data fitting coefficient matrix is taken as the dependent variable. The function relationship between the independent variables and the dependent variables is fitted by using the multiple linear regression method. For example, the business stages of case acceptance, investigation, examination and prosecution, and trial are respectively assigned quantitative identifiers, such as 1, 2, 3, and 4, as the values of the independent variables; and the coefficients of each dimension in the data fitting coefficient matrix corresponding to each business stage are taken as the values of the dependent variables. By collecting a large number of historical business stages and their corresponding data fitting coefficient matrix samples, the regression coefficients and constant terms are obtained by substituting them into the multiple linear regression formula for calculation, and a function relationship formula is formed. Assuming that after calculation, the function relationship of the coefficients of each dimension in the data fitting coefficient matrix of the case acceptance stage (identifier 1) and the stage identifier is: coefficient = 0.05 x stage identifier + 0.98, and the function relationship of the case investigation stage (identifier 2) is: coefficient = 0.07 x stage identifier + 0.92, and so on. The function relationships of all business stages are integrated to form a complete mapping model.

[0042] When adjusting the data matrix after management and control corresponding to the target stage based on the data deviation matrix to obtain the adjusted data matrix corresponding to the target stage, first, the data fitting coefficient matrix for the target stage is determined based on the mapping model. For example, if the target stage is the examination and prosecution stage (quantitative identifier 3), the fitting coefficients of each data dimension are calculated by querying the function relationship corresponding to the stage in the mapping model, such as the fitting coefficient of the “case file transfer integrity” dimension is 1.08, and the fitting coefficient of the “suspect rights notification record” dimension is 1.05. The coefficients are integrated to form the data fitting coefficient matrix of the examination and prosecution stage.

[0043] The data fitting coefficient matrix is used to adjust the data deviation matrix to obtain the adjustment information of the data matrix after management and control corresponding to the target stage. The data deviation matrix is determined according to the original data matrix corresponding to the data matrix after management and control of the target stage and the actual data matrix when the stage difference between the target stage and the actual data matrix is greater than or equal to the preset difference threshold. Assuming that the value of the “case file transfer integrity” dimension in the data matrix after management and control of the examination and prosecution stage is 0.9, and the original value of this dimension in the actual data matrix is 0.82, then the deviation value of this dimension is 0.08; the value of the “suspect rights notification record” dimension after management and control is 0.95, and the actual original value is 0.88, and the deviation value is 0.07, forming a data deviation matrix. The corresponding coefficients in the data fitting coefficient matrix are used to adjust the deviation matrix, such as the adjustment value of the “case file transfer integrity” dimension is 0.08 x 1.08 = 0.0864, and the adjustment value of the “suspect rights notification record” dimension is 0.07 x 1.05 = 0.0735. These adjustment values collectively constitute the adjustment information.

[0044] Adjust the post-control data matrix corresponding to the target stage by using the adjustment information to obtain the adjusted post-control data matrix corresponding to the target stage. For example, the value of "case file transfer integrity" in the post-control data matrix of the examination and prosecution stage is 0.9, and according to the adjustment information 0.0864, it is adjusted to 0.9-0.0864=0.8136; the value of "suspect rights notification record" is 0.95, which is adjusted to 0.95-0.0735=0.8765 according to the adjustment information 0.0735, and other data dimensions are adjusted in the same way to form the adjusted post-control data matrix.

[0045] The adjusted post-control data matrix is taken as the new post-control data matrix corresponding to the target stage, and the step of performing process control operation based on the post-control data matrix corresponding to the target stage is returned. The actual data matrix is measured again through the data acquisition system, and the stage difference is calculated. If the stage difference is still greater than or equal to the preset difference threshold at this time, the above process of determining the data fitting coefficient matrix based on the mapping model, adjusting the data deviation matrix, and generating the adjusted post-control data matrix is repeated. For example, after the new post-control data matrix performs the process control, the stage difference between the original data corresponding to the actual data matrix and the new post-control data matrix is still not up to standard after calculation, then the data fitting coefficient matrix of the target stage is obtained again by calling the mapping model, the data deviation matrix is recalculated and adjusted, until the stage difference is less than the preset difference threshold.

[0046] In the whole process, the data fitting coefficient matrix of different target stages will be different due to the pertinence of the mapping model. For example, the data fitting coefficient matrix of the trial stage is different from that of the examination and prosecution stage. The fitting coefficients of the dimensions such as "trial record integrity" and "judgment document standardization" are determined according to the historical data characteristics of the stage through the mapping model, so as to ensure that the adjustment information can adapt to the business characteristics of different stages.

[0047] In the case of extracting the actual data matrix from the process data, the process data needs to be normalized to map the attribute values of each process node to the interval of 0-1, and arranged in the order of node topology to form a two-dimensional matrix structure. For example, in the case receiving stage of the case handling center, the process data contains information of multiple process nodes such as case number, receiving time, material integrity, and entry personnel number. Among them, the “receiving time” is recorded in hours, and it is assumed that the historical data of the receiving time in this stage shows that the longest is 10 hours and the shortest is 0 hour, and the actual receiving time of a case is 3 hours. In the normalization process, it is mapped to 0.3 by calculating (3-0) / (10-0)=0.3; “material integrity” is determined by checking the number of missing items, if the total number of items is 20, and 3 items are missing in a case, then the completeness is (20-3) / 20=0.85, which is directly used as the normalized attribute value of the node; “entry personnel number” is a classification data, which is converted to multiple 0-1 values by one-hot encoding, such as the personnel with number 3 corresponding to the vector [0, 0, 1, 0, …], and the position of the 1 is taken as the normalized representation of the node. Arrange these normalized attribute values in the order of node topology in the case receiving stage, i.e. case number → receiving time → material integrity → entry personnel number, to form a 4-row 1-column two-dimensional matrix. The matrix is the actual data matrix of the case in the receiving stage.

[0048] The method for determining the process consistency variance includes: calculating the difference between the data value of each process node and the data mean, and taking the arithmetic mean of the square values of all nodes as the process consistency variance. Take the case interrogation stage as an example, which contains 4 process nodes such as interrogation duration, synchronous audio and video quality, interrogation record integrity, and personnel qualification. The actual data matrix of each node after normalization is [0.6, 0.7, 0.5, 0.8]. First, calculate the mean of the data values of the 4 nodes, i.e. (0.6+0.7+0.5+0.8) / 4=0.65. Then calculate the difference between each node data value and the mean: 0.6-0.65=-0.05; 0.7-0.65=0.05; 0.5-0.65=-0.15; 0.8-0.65=0.15. Square these differences to get (-0.05)²=0.0025; (0.05)²=0.0025; (-0.15)²=0.0225; (0.15)²=0.0225. Finally, take the arithmetic mean of these square values, i.e. (0.0025+0.0025+0.0225+0.0225) / 4=0.0125, which is the process consistency variance of the case interrogation stage.

[0049] In the case evidence review stage, the process nodes include evidence number, evidence collection time, evidence relevance, storage record integrity, etc. The actual data matrix after normalization is [0.9, 0.7, 0.8, 0.6, 0.7]. The average value of the data is (0.9+0.7+0.8+0.6+0.7) / 5=0.74. The difference between each node and the average value is 0.16, -0.04, 0.06, -0.14, -0.04, and the square is 0.0256, 0.0016, 0.0036, 0.0196, 0.0016. The arithmetic mean is (0.0256+0.0016+0.0036+0.0196+0.0016) / 5=0.0104, which is the process consistency variance of this stage.

[0050] When processing the process data of the case transfer stage, it is assumed that there are 6 process nodes including transfer document specification, case item list integrity, receiving time, electronic file synchronization rate, etc. The normalized actual data matrix is [0.85, 0.9, 0.7, 0.8, 0.75, 0.8]. The average value is (0.85+0.9+0.7+0.8+0.75+0.8) / 6=0.8. The difference between each node and the average value is 0.05, 0.1, -0.1, 0, -0.05, 0, and the square is 0.0025, 0.01, 0.01, 0, 0.0025, 0. The arithmetic mean is (0.0025+0.01+0.01+0+0.0025+0) / 6≈0.00417, which is the process consistency variance of this stage.

[0051] In the above process, normalization ensures that process node data of different types and different magnitudes can be compared on the same scale, and process consistency variance quantifies the deviation of each node data from the overall average, reflecting the stable state of the process. Whether it is case reception, interrogation, evidence review or transfer stage, the actual data matrix is extracted in the same way and the process consistency variance is calculated, providing a quantitative basis for subsequent judgment of whether the process needs to be adjusted. When the process consistency variance is greater than or equal to the preset variance threshold, the data matrix iteration optimization is started; when it is less than the preset variance threshold, it is considered that the current process data is in a stable state, and the corresponding iteration data matrix can be used as the control data matrix.

[0052] In specific operation, the number and attributes of process nodes in different business stages differ, and the specific calculation method of normalization needs to be adjusted flexibly according to the type of node data. For example, for time data, the difference between the maximum value and the minimum value is used as the denominator; for count data, the total number of items is used as the denominator; for classification data, hot encoding is used for conversion. The calculation logic of process consistency variance remains the same, and all follow the formula:

[0053] wherein S2 represents the process consistency variance, xi represents the normalized data value of the i-th process node, μ represents the arithmetic mean of the normalized data values of all process nodes, n represents the number of process nodes, and ∑ represents the summation operation on the calculation results of all nodes. Through this formula, the process data consistency of different business stages can be quantified uniformly, and a standardized judgment index is provided for the whole-process control.

[0054] In the case registration process of the case handling center, nodes such as “party information input”, “case type division”, “jurisdiction verification” are involved. The deviation change samples of the process in the past year are collected, wherein the “party information input” node has 1000 samples, each sample contains the deviation direction of the node in a certain control (positive deviation means the data value is higher than the standard, negative deviation means the data value is lower than the standard), deviation amplitude (such as 0.02, 0.05, etc. normalized difference value), deviation trend (such as continuous 3 times deviation increase, first increase and then decrease, etc.). These samples are classified and labeled, the deviation direction is divided into “positive” and “negative”, the deviation amplitude is divided into three categories according to 0-0.03, 0.03-0.07, 0.07 and above, and the deviation trend is divided into “continuous increase”, “continuous decrease” and “fluctuation change”. By combining these classifications, a number of state vectors are formed, such as (positive, 0-0.03, continuous increase), (negative, 0.03-0.07, fluctuation change), etc. The optimal calibration weight corresponding to each type of state vector in the historical sample is calculated, that is, the value of the process consistency variance decreases fastest after applying the weight. Assuming that the state vector (positive, 0-0.03, continuous increase) has the best adjustment effect when using a calibration weight of 0.6 in the historical sample, the state vector is associated with the weight of 0.6 and recorded. All process node samples are processed in this way, and finally a calibration weight table covering various state vectors of each node is formed.

[0055] The step of constructing the mapping model comprises: taking the business stage as the independent variable, taking the corresponding data fitting coefficient matrix as the dependent variable, adopting the multivariate linear regression method to fit the functional relationship between the independent variable and the dependent variable, and forming the mapping model. For example, the business stages of the case handling center include four stages of "case acceptance", "evidence collection", "examination and prosecution", and "court preparation", which are represented by variables x = 1, x = 2, x = 3, and x = 4 respectively. The data fitting coefficient matrix corresponding to each stage contains multiple dimensions of coefficients, such as the coefficient matrix of the "case acceptance" stage, which includes "material completeness coefficient", "input accuracy coefficient", "acceptance time coefficient", etc., which are denoted as y1, y2, y3 respectively; the coefficient matrix of the "evidence collection" stage includes "evidence collection standardization coefficient" and "evidence chain integrity coefficient", which are denoted as y4 and y5 respectively, and so on. Collect the corresponding samples of the independent variable x and each dependent variable yi in the historical data of each stage, such as x = 1, y1 = 1.05, y2 = 0.98, y3 = 1.12; x = 2, y4 = 1.10, y5 = 1.03, etc. For each dependent variable yi, a multivariate linear regression equation is established with x. Taking y1 (material completeness coefficient) as an example, assuming that the collected samples are x = 1, y1 = 1.05, the associated value of y1 corresponding to the relevant stage is 1.08 (calculated by cross-stage data association) when x = 2, 1.11 when x = 3, and 1.14 when x = 4, and the multivariate linear regression formula is calculated to obtain y1 = 1.02 + 0.03x. Similarly, regression equations are fitted for y2, y3 and other coefficients, such as y2 = 0.99 - 0.01x, y3 = 1.09 + 0.01x, etc. These equations are integrated to form a complete mapping model, which can output the coefficient values of each dimension in the data fitting coefficient matrix corresponding to the stage when the independent variable x of the stage is input.

[0056] In practical application, the calibration weight table and the mapping model will be continuously updated with new process data. For example, when a new process node "electronic file synchronization rate" is added to the "court preparation" stage, the deviation change samples of this node will be collected and supplemented to the calibration weight table; when new change trends appear in the data fitting coefficients of each business stage, samples will be re-collected, and the regression equations in the mapping model will be updated by the multivariate linear regression method to adapt to the dynamic changes of the process control of the case handling center.

[0057] Taking a certain control of the "evidence collection" stage as an example, assuming that the "evidence storage temperature" node of this stage deviates, the deviation direction is positive, the deviation amplitude is 0.02 (belongs to the 0-0.03 interval), and the deviation trend is continuously increasing. The corresponding calibration weight is 0.5 for the deviation adjustment of this node through the calibration weight table query. At the same time, if it is necessary to determine the data fitting coefficient matrix of this stage, x=2 is substituted into the mapping model to obtain "forensic normative coefficient" y4=1.10 and "evidence chain integrity coefficient" y5=1.03, which are used as the basis for adjusting the data deviation matrix. In this way, the calibration weight table and the mapping model provide a quantitative standard based on historical experience for data adjustment in the process control, making the adjustment process more targeted and operable.

[0058] Whether it is information input at the case acceptance stage or material checking in the court preparation stage, the deviation adjustment of each process node depends on the weight value provided by the calibration weight table, and the data fitting coefficient of each business stage is quickly determined through the mapping model. Both of them together constitute the basis of data optimization in the whole process control of the case handling center, ensuring that the data matrix after control can accurately reflect the actual running state of each stage.

[0059] It should be noted that in this paper, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "contain" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such process, method, article or equipment.

[0060] Although embodiments of the present application have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A big data-based case-handling center whole-process management and control method, characterized in that, The method comprises: controlling the case-handling center in the whole process based on the initial data matrix corresponding to each business stage, to obtain a controlled data matrix of the case-handling center in each business stage; determining a target stage required to be controlled by the case-handling center, performing process control operation based on the controlled data matrix corresponding to the target stage, and measuring an actual data matrix of the case-handling center through a data acquisition system; in the case that the difference between the target stage and the actual data matrix is greater than or equal to a preset difference threshold, determining a data deviation matrix according to the controlled data matrix corresponding to the target stage and an original data matrix corresponding to the actual data matrix; adjusting the controlled data matrix corresponding to the target stage based on the data deviation matrix, to obtain an adjusted data matrix corresponding to the target stage; taking the adjusted data matrix corresponding to the target stage as a new controlled data matrix corresponding to the target stage, and returning to the step of performing process control operation based on the controlled data matrix corresponding to the target stage, until the difference between the target stage and the actual data matrix is less than the preset difference threshold. 2.The big data-based case-handling center whole-process management method according to claim 1, characterized in that, The controlling the case-handling center in the whole process based on the initial data matrix corresponding to each business stage, to obtain a controlled data matrix of the case-handling center in each business stage, comprises: for any business stage, performing process control operation of the case-handling center based on the initial data matrix corresponding to the business stage, and obtaining process data of the case-handling center by using a data acquisition system, and extracting an actual data matrix from the process data; determining a process consistency variance according to the actual data matrix; in the case that the process consistency variance is greater than or equal to a preset variance threshold, determining an actual data matrix corresponding to the actual data matrix based on a pre-constructed stage data relationship model; the stage data relationship model represents the corresponding relationship between the business stage and the data; iterating the actual data matrix by using a data matrix iteration model, to obtain an iterated data matrix corresponding to the actual data matrix; taking the iterated data matrix as the initial data matrix, and returning to the step of performing process control operation of the case-handling center based on the initial data matrix corresponding to the business stage, until the process consistency variance is less than the preset variance threshold; taking the iterated data matrix as the controlled data matrix. 3.The big data-based case-handling center whole-process management method according to claim 2, characterized in that, The iterating the actual data matrix by using a data matrix iteration model, to obtain an iterated data matrix corresponding to the actual data matrix, comprises: determining a deviation matrix of a current iteration round according to the actual data matrix and the initial data matrix; obtaining a calibration weight matrix corresponding to the current iteration round, and adjusting the deviation matrix of the current iteration round by using the calibration weight matrix corresponding to the current iteration round, to obtain adjustment information of the current iteration round for the initial data matrix; Adjust the initial data matrix based on the adjustment information of the current iteration round to obtain the iteration data matrix. 4.The big data-based case-handling center whole-process management method according to claim 3, characterized in that, The calibration weight matrix comprises calibration weights of each process node in the case handling center; and the calibration weight matrix corresponding to the current iteration round is obtained by: obtaining the deviation change direction information, the deviation change amplitude information and the deviation change trend information of each process node in the current iteration round; generating an iteration state vector corresponding to each process node based on the deviation change direction information, the deviation change amplitude information and the deviation change trend information of each process node; taking the iteration state vector corresponding to each process node as an index to query the calibration weight matched with each process node in the pre-constructed calibration weight table; and generating the calibration weight matrix corresponding to the current iteration round based on the calibration weight matched with each process node. 5.The big data-based case-handling center whole-process management method according to claim 1, characterized in that, After the step of controlling the case handling center based on the initial data matrix corresponding to each business stage to obtain the controlled data matrix of the case handling center in each business stage, the method further comprises: determining a data fitting coefficient matrix corresponding to each business stage according to the initial data matrix and the controlled data matrix corresponding to each business stage; constructing a mapping model according to each business stage and the data fitting coefficient matrix corresponding to each business stage; the mapping model represents the mapping relationship between the business stage and the data fitting coefficient matrix. 6.The big data-based case-handling center whole-process management method according to claim 5, characterized in that, The adjustment of the controlled data matrix corresponding to the target stage based on the data deviation matrix to obtain the adjusted data matrix corresponding to the target stage comprises: determining a data fitting coefficient matrix for the target stage based on the mapping model; adjusting the data deviation matrix based on the data fitting coefficient matrix to obtain adjustment information of the controlled data matrix corresponding to the target stage; adjusting the controlled data matrix corresponding to the target stage based on the adjustment information to obtain the adjusted data matrix corresponding to the target stage. 7.The big data-based case-handling center whole-process management method according to claim 2, characterized in that, The actual data matrix extracted from the process data comprises: normalizing the process data to map the attribute values of each process node to the interval of 0-1, and arranging the nodes in topological order to form a two-dimensional matrix structure. 8.The big data-based case-handling center whole-process management method according to claim 2, characterized in that, The determination method of the process consistency variance comprises: calculating the difference between the data value of each process node and the data mean, squaring the difference, and taking the arithmetic mean of all node square values as the process consistency variance. 9.The big data-based case-handling center whole-process management method according to claim 4, characterized in that, The pre-constructed calibration weight table is formed by the following steps: collecting deviation change samples of each process node in historical process control, classifying and labeling the direction, amplitude and trend information in the samples, counting the optimal calibration weight corresponding to each type of state vector, and establishing a mapping relationship table between the state vector and the calibration weight. 10.The big data-based case management center whole-process management method according to claim 5, characterized in that, The step of constructing the mapping model comprises: taking a business stage as an independent variable, taking a corresponding data fitting coefficient matrix as a dependent variable, adopting a multiple linear regression method to fit a functional relationship between the independent variable and the dependent variable, and forming the mapping model.

Images