Intelligent process task allocation method and system based on combination of Activiti and machine learning

Abstract

The invention provides an intelligent process task allocation method and system based on combination of Activiti and machine learning, and relates to the technical field of machine learning, and the method comprises the steps: collecting multi-dimensional road conference process data, and carrying out the data cleaning; processing and calculating the multi-dimensional road conference flow data, and extracting road conference type features, task emergency degree features and compliance approver task load features; a task allocation model is constructed, the task allocation model is trained based on the extracted features, and the model is used for predicting task approval time of different approvers; integrating the trained task allocation model to an Activiti process engine, and when a process task reaches an approval node, calling the model to calculate predicted approval time of candidate approvers; according to the predicted approval time dynamic allocation task, selecting the compliance approver with the shortest predicted approval time to execute the current approval task; newly added approval data are collected regularly, and the task distribution model is retrained to iteratively optimize the prediction effect.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to an intelligent process task allocation method and system based on the combination of Activiti and machine learning. Background Art

[0002] With the increasing adoption of informatization, more and more companies are requiring online electronic approval processes for various business processes. Activiti, a Java-based open source workflow and business process management engine, adheres to the Apache 2.0 open source license and complies with the BPMN 2.0 standard. It's highly sought after for its lightweight nature and ease of integration with mainstream frameworks like Spring, facilitating the visual design, automated execution, and monitoring of business processes.

[0003] However, the native Activiti engine primarily relies on fixed allocation when assigning approval tasks, meaning that administrators assign tasks to fixed approvers based on their experience and judgment, or according to simple rules. This allocation method is inefficient, making it difficult to quickly respond to large or complex approval tasks. It also suffers from imbalanced allocation, failing to effectively assess the actual workload and processing capacity of approvers, leading to inadequate resource utilization. Finally, it lacks foresight, making it difficult to analyze and predict based on historical data, making it impossible to optimize the allocation of approval process tasks. Consequently, approval efficiency and quality are low. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent process task allocation method and system based on the combination of Activiti and machine learning, so as to solve the problem mentioned in the above background technology that the native Activiti engine mainly relies on fixed allocation or allocation according to simple rules when allocating approval tasks, which is not only inefficient but also has unbalanced allocation and lack of predictability.

[0005] To achieve the above object, the present application provides the following technical scheme: a kind of intelligent process task allocation method based on Activiti and machine learning combination, steps include: collecting multi-dimensional roadshow meeting process data and carrying out data cleaning;The multi-dimensional roadshow meeting process data is processed and calculated, and the roadshow meeting type feature, task urgency feature and compliance approval personnel task load feature are extracted;Task allocation model is constructed, and the task allocation model is trained based on the extracted features, the model is used to predict the task approval time of different approval personnel;The trained task allocation model is integrated into Activiti process engine, when the process task reaches the approval node, the model is called to calculate the predicted approval time of candidate approval personnel;According to the predicted approval time, dynamically allocate tasks, select the compliance approval personnel with the shortest predicted approval time to execute the current approval task;Periodically collect new approval data, retrain the task allocation model to iteratively optimize the prediction effect.

[0006] Optionally, the preprocessing step specifically includes: recording data sources and processing logic during the multi-dimensional roadshow meeting process data integration process, desensitizing sensitive information, retaining business-related features, and excluding data missing values;Category data is converted into unique binary vectors using one-hot encoding, and numerical data is normalized to fall within the [0,1] interval.

[0007] Optionally, the multi-dimensional roadshow meeting process data includes: roadshow meeting type, task submission time, meeting start time, approval personnel information, professional field matching degree, online status, approval duration and number of approval personnel pending approval tasks;The roadshow meeting type is used to determine the task workload according to different types;The task submission time and the meeting start time are used to calculate the task urgency;The professional field matching degree is used to evaluate the adaptability of the approval personnel to the task type, and the online status is used to judge the real-time availability of the approval personnel, combined with the professional field matching degree and the online status to judge whether the approval personnel is compliant;The approval duration and the number of approval personnel pending tasks are used to calculate the approval personnel load.

[0008] Optionally, the step of judging whether the approval personnel is compliant based on the professional field matching degree and the online status specifically includes: the type of the professional field corresponds to the type of the meeting one by one, judging whether the type of the professional field matches the type of the meeting, if yes, record as 1, otherwise record as 0;Judge whether the approval personnel is online, if online, record as 1, otherwise record as 0;If the type of the professional field matches the type of the meeting and the approval personnel is online, it is determined that the approval personnel is compliant.

[0009] Optionally, the task emergency degree feature calculation step specifically comprises: obtaining a time difference between the task submission time and the meeting start time; converting the time difference into an emergency degree value using a time decay function, the smaller the time difference, the higher the emergency degree value; introducing normal distribution consistent random noise in the calculation process to simulate the uncertainty in the actual approval scene.

[0010] Optionally, the compliance approver task load feature calculation method comprises: counting the number of pending approval tasks of the current approver; calculating the emergency degree of each pending approval task and sorting them from high to low according to the emergency degree; calculating the normalized task load value according to the current task position in the sorting and the historical maximum number of tasks.

[0011] Optionally, the step of constructing the task allocation model specifically comprises: constructing a linear regression model based on the task type feature, the task emergency degree feature, and the compliance approver task load feature; performing logarithmic transformation on the approval time to ensure that the predicted time is positive, and introducing cross-term features to optimize the prediction; solving the model parameters by least squares method to obtain the optimal parameters.

[0012] Optionally, the step of integrating the trained task allocation model into the Activiti process engine specifically comprises: configuring a custom listener in the BPMN file, which triggers the listener when the process task reaches the approval node.

[0013] Optionally, the step of configuring the custom listener specifically comprises: inheriting the TaskListener interface of the Activiti engine and rewriting the notify method; obtaining the context information of the current task in the notify method; calling the API of the task allocation model, inputting the task type feature, the task emergency degree feature, and the compliance approver task load feature, obtaining the predicted approval time of the candidate approver, and determining the optimal approver; dynamically setting the approver variable through the runtime service interface of the Activiti engine, and updating the task allocation through the RuntimeService interface.

[0014] In another aspect, the present application also provides an intelligent process task allocation system based on the combination of Activiti and machine learning, comprising: a data collection module for collecting multi-dimensional roadshow conference process data and performing data cleaning; a feature extraction module for processing and calculating the multi-dimensional roadshow conference process data, and extracting roadshow conference type features, task urgency features and compliance approval personnel task load features; a model construction module for constructing a task allocation model, and training the task allocation model based on the extracted features, the model being used for predicting the task approval time of different approval personnel; an integration module for integrating the trained task allocation model into an Activiti process engine, and calling the model to calculate the predicted approval time of candidate approval personnel when a process task reaches an approval node; an allocation module for dynamically allocating tasks according to the predicted approval time, and selecting the compliance approval personnel with the shortest predicted approval time to perform the current approval task; and an optimization module for periodically collecting new approval data, and retraining the task allocation model to iteratively optimize the prediction effect.

[0015] Compared with the prior art, the present application has the following beneficial effects: The present application dynamically allocates tasks in combination with the Activiti workflow engine and the machine learning model, reduces the approval waiting time and the approval time, and significantly improves the efficiency of the approval process. According to the historical performance and the current working state of the approval personnel, the tasks are intelligently allocated to achieve balanced distribution of workload and avoid waste of resources. By using machine learning technology, the system can continuously learn and optimize the allocation strategy to adapt to changes in approval requirements and realize intelligent process management. This intelligent task allocation method combining the Activiti process engine and machine learning not only improves the efficiency and quality of the approval process, but also improves the intelligent level of the system through dynamic optimization of resource allocation, providing a more efficient and accurate process task allocation scheme for enterprises. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The figure is a schematic diagram of the method steps of the present application.

[0017] Figure 2 The figure is a schematic diagram of the system structure of the present application.

[0018] In the figure: 10 - data collection module, 20 - feature extraction module, 30 - model construction module, 40 - integration module, 50 - allocation module, 60 - optimization module. DETAILED DESCRIPTION

[0019] The scheme of the present application will be clearly and completely described below in combination with 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 the embodiments.

[0020] It should be noted that the terms "first", "second", and the like in the description and in the claims of the present application and the above drawings merely mean different instances of similar objects, and do not necessarily imply a specific order or sequence. It should be understood that the data thus used in the specification enabes the embodiments of the present application described herein to be combined with each other, where appropriate, unless explicitly stated otherwise. Moreover, the terms "comprising" and "having" and any variations thereof in the specification and in the claims are intended to cover a non-exclusive inclusion, for example, a process, method, system, product or apparatus that comprises a list of steps or units not necessarily limited to those specifically listed, but can include other steps or units not expressly listed or inherent to such process, method, product or apparatus.

[0021] Those skilled in the art can understand that, unless specifically stated otherwise, the singular forms "a", "an", and "the" used herein include plural referents unless expressly stated otherwise. It should be further understood that the use of the term "include", "comprise" or "have" in the specification of the present application indicates the presence of the features, integers, steps, operations, elements, and / or components referred to, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof. It should be understood that when we say an element is "connected" or "coupled" to another element, it can be directly connected or coupled to the other element, or there can be an intermediate element. In addition, "connected" or "coupled" used herein can include wireless connection or wireless coupling. The phrase "and / or" used herein includes all or any one of the associated listed items and all combinations thereof.

[0022] Those skilled in the art can understand that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as that understood by a person of ordinary skill in the art to which the present application belongs. It should also be understood that terms such as those defined in a general dictionary should be understood to have meanings consistent with those in the context of the prior art, and should not be interpreted in an idealized or overly formal sense unless specifically defined as such.

[0023] It should be understood that the sequence numbers and sizes of the steps in the embodiments do not mean the order of execution, and the execution order of the processes is determined by their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0024] It should be noted that the embodiments in the present application and the features in the embodiments can be combined with each other without conflict. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0025] Please refer to Figure 1The application discloses an intelligent process task allocation method based on combination of Activiti and machine learning, and steps include: collecting multi-dimensional roadshow conference process data and performing data cleaning; processing and calculating the multi-dimensional roadshow conference process data, and extracting roadshow conference type features, task urgency features and compliance approval personnel task load features; constructing a task allocation model, and training the task allocation model based on the extracted features, the model being used for predicting task approval time of different approval personnel; integrating the trained task allocation model into an Activiti process engine, and when a process task reaches an approval node, calling the model to calculate predicted approval time of candidate approval personnel; dynamically allocating tasks according to the predicted approval time, and selecting compliance approval personnel with the shortest predicted approval time to execute a current approval task; periodically collecting newly added approval data, and retraining the task allocation model to iteratively optimize prediction effect.

[0026] Specifically, by collecting multi-dimensional roadshow conference process data, extracting key features such as roadshow conference type features, task urgency features and compliance approval personnel task load features, and constructing a task allocation model for task allocation, the model is integrated into an Activiti process engine, and when a process task reaches an approval node, the model is called to calculate predicted approval time of candidate approval personnel, tasks are dynamically allocated according to the predicted approval time, and compliance approval personnel with the shortest predicted approval time is selected to execute a current approval task, so that the mechanism of automatically selecting optimal approval personnel significantly shortens approval process time, realizes dynamic intelligent allocation of approval tasks, overcomes defects of low efficiency and uneven resource allocation of traditional fixed allocation methods, and periodically collects newly added approval data to retrain the task allocation model to iteratively optimize prediction effect. The application dynamically allocates tasks by combining an Activiti workflow engine and a machine learning model, reduces approval waiting time and approval time, and significantly improves efficiency of the approval process. According to historical performance and current working state of the approval personnel, tasks are intelligently allocated to realize balanced distribution of workloads and avoid resource waste. By using machine learning technology, the system can continuously learn and optimize allocation strategies, adapt to changes in approval requirements, and realize intelligent process management. Intelligent task allocation combining the Activiti process engine and machine learning not only improves efficiency and quality of the approval process, but also improves intelligent level of the system by dynamically optimizing resource allocation, and provides a more efficient and accurate process task allocation scheme for enterprises.

[0027] In some embodiments, the preprocessing step specifically includes: recording data sources and processing logic, desensitizing sensitive information, retaining business-related features and eliminating missing data during multi-dimensional roadshow conference process data integration; using one-hot encoding to convert category data into unique binary vectors, and normalizing numerical data to fall within the [0, 1] interval.

[0028] Specifically, the collected multi-dimensional roadshow meeting process data is standardized and cleaned, and the data quality is improved through desensitization processing and missing value elimination to provide reliable input for model training. The data traceability mechanism enhances the traceability of feature engineering and reduces the impact of data anomalies on allocation results. The collected multi-dimensional roadshow meeting process data is processed, specifically including: professional field matching is recorded as 1, otherwise 0; approval personnel online is recorded as 1, otherwise 0; approval personnel passing approval is recorded as 1, otherwise 0; roadshow meeting type includes industry view report, company view report and company performance exchange, recorded as: , in the formula, is the roadshow meeting type, and different meeting types require different information to be checked by compliance approval personnel, so the time consumption is different; the category type data is converted into a unique binary vector by using one-hot encoding, and the numerical value type data is normalized to fall within the interval [0, 1].

[0029] In some embodiments, the multi-dimensional roadshow meeting process data includes: roadshow meeting type, task submission time, meeting start time, approval personnel information, professional field matching degree, online state, approval passing time and approval personnel pending task quantity; the roadshow meeting type is used to determine the task workload according to different types; the task submission time and the meeting start time are used to calculate the task urgency; the professional field matching degree is used to evaluate the adaptability of the approval personnel to the task type, and the online state is used to judge the real-time availability of the approval personnel, combined with the professional field matching degree and the online state to judge whether the approval personnel is compliant; the approval passing time and the approval personnel pending task quantity are used to calculate the approval personnel load.

[0030] Specifically, the multi-dimensional data such as roadshow meeting type, task submission time, meeting start time, approval personnel information, professional field matching degree, online state, approval passing time and approval personnel pending task quantity are evaluated to construct a comprehensive task allocation index system. Through the joint judgment mechanism of professional field matching degree, online state, task urgency and approval personnel load, the efficiency and quality of the approval process are improved while ensuring compliance.

[0031] In some embodiments, the step of judging whether the approval personnel is compliant in combination with the professional field matching degree and the online state specifically includes: the type of the professional field corresponds to the meeting type one-to-one, judging whether the type of the professional field matches the meeting type, if yes, recording as 1, otherwise 0; judging whether the approval personnel is online, if online, recording as 1, otherwise 0; if the type of the professional field matches the meeting type and the approval personnel is online, it is determined that the approval personnel is compliant.

[0032] Specifically, the binary quantization standard is adopted to realize efficient compliance screening, and a candidate list is accurately generated through professional matching and online state double verification, thereby effectively reducing invalid task allocation.

[0033] In some embodiments, the calculation of the task urgency feature specifically includes: obtaining a time difference between the task submission time and the meeting start time; converting the time difference into an urgency value by using a time decay function, the smaller the time difference, the higher the urgency value; and introducing random noise conforming to a normal distribution in the calculation process to simulate uncertainty in an actual approval scenario.

[0034] Specifically, a self-defined time decay function is used to convert the time interval into the task urgency, and the calculation formula is: , wherein, is the task urgency, is the roadshow meeting start time, is the process task submission time, is a standard overall distribution, and The urgency quantification model based on the time decay function in the application, combined with random noise simulation of the actual scenario, can more accurately identify the priority difference of the task and ensure that the urgent task is processed in time.

[0035] In some embodiments, the calculation of the compliance approval personnel task load feature includes: counting the number of to-be-approved tasks of the current approval personnel; calculating the urgency of each to-be-approved task and sorting them from high to low according to the urgency; and calculating the normalized task load value according to the position of the current task in the sorting and the historical maximum number of tasks.

[0036] Specifically, when the approval task is submitted, the to-be-approved tasks of the compliance approval personnel are , the number of tasks is n, the task urgency of each task is calculated, and the task urgency is sorted from high to low, and the sorted result is , the calculation formula of the compliance approval personnel task load is: , wherein, is the compliance approval personnel task load, is the position number of the task urgency sorted from high to low, is the maximum value of the historical to-be-approved task number n of all compliance approval personnel. The load evaluation method through dynamic sorting and normalization processing in the application can reflect the work pressure of the approval personnel in real time, and realize intelligent balanced allocation of the task quantity.

[0037] In some embodiments, the step of constructing the task allocation model specifically comprises: constructing a linear regression model based on the task type feature, the task urgency feature, and the compliance approval personnel task load feature; performing logarithmic transformation on the approval time to ensure that the predicted time is positive, and introducing a cross-term feature to optimize the prediction; and solving the model parameters by least squares method to obtain the optimal parameters.

[0038] Specifically, an intelligent task allocation model is constructed, and the calculation formula thereof is: , wherein y is the approval time, logarithmic transformation is adopted to ensure that the predicted time is positive, is a constant term, is the roadshow meeting type , the corresponding matrix coefficient is in the form of , is the task urgency , the corresponding coefficient is is the compliance approval personnel task load , the corresponding coefficient is. Because the task urgency has an impact on the compliance approval personnel task load, a cross-term is introduced, and the corresponding coefficient is , is a standard normal distribution and .

[0039] First, a coefficient vector is constructed; a feature matrix after data processing is constructed, wherein n is the sample data amount; the target vector is , the least squares method is used to solve the intelligent task allocation model, and the calculation formula of the residual sum of squares function of the least squares method is: , wherein is the residual sum of squares function of the least squares method, is the target vector, is the feature matrix after data processing, is the coefficient vector, in order to find the coefficient vector that makes minimum, the derivative of is calculated, and the calculation formula is: , wherein is a partial differential operator, and the derivative is set to 0 to obtain the normal equation: from which can be solved, considering that the data amount is large, the mumpy library is used to complete the construction of the intelligent task allocation model. The logarithmic linear model with the cross-term feature adopted in the present application can improve the prediction accuracy while ensuring the calculation efficiency. The least squares method parameter estimation enhances the stability of the model and meets the real-time response demand of the process engine.

[0040] In some embodiments, the step of integrating the trained task allocation model into the Activiti process engine specifically comprises configuring a custom listener in a BPMN file, which triggers when a process task reaches an approval node.

[0041] Specifically, when the roadshow meeting process reaches the compliance personnel approval node, the present specification takes the node_a node of the process flow_a as an example. The bpmn file of the node_a node of the process flow_a is newly configured as follows: <extensionelements> <activiti:taskListener event="create" class="listener.MyTaskCreateListener" / > <activiti:taskListener event="start" class="listener.MyTaskCreateListener" / > < / extensionelements> When the process flow_a runs to the node_a node, the custom listener MyTaskCreateListener is triggered. The present application realizes non-invasive system integration through standard BPMN extension, and the listener mechanism seamlessly adds intelligent allocation capability on the basis of maintaining the original functions of Activiti.

[0042] In some embodiments, the step of customizing the listener specifically comprises inheriting the TaskListener interface of the Activiti engine and rewriting the notify method; obtaining the context information of the current task in the notify method; calling the API of the task allocation model, inputting the task type feature, the task urgency feature, and the compliance approval personnel task load feature, obtaining the predicted approval time of the candidate approval personnel, and determining the optimal approval personnel; dynamically setting the approver variable through the runtime service interface of the Activiti engine, and updating the task allocation through the RuntimeService interface.

[0043] Specifically, the process flow_a is customized with the process listener class MyTaskCreateListener, which implements the TaskListener interface of the acitiviti process engine and rewrites the notify method. When the approval task runs to the node_a node, the notify method in the MyTaskCreateListener listener is triggered.

[0044] The logic in the notify method is to call the API of the trained machine learning model, input the current roadshow meeting type , the task urgency , and the compliance approval personnel task load The predicted approval passing time of multiple candidates is obtained. An approval person who can pass the approval task and has the shortest approval time is selected as the current node approval person. The setVariable method of the Activiti process engine RuntimeService class is called to assign the optimal approval person as the approval person of the current approval task in real time. After the assignment is completed, the approval person receives a pending task reminder and performs the approval process. Based on the dynamic variable setting technology of the runtime interface, the application realizes real-time seamless switching of the approval person and significantly improves the task reassignment efficiency in abnormal situations.

[0045] In another aspect, the application also provides an intelligent process task allocation system based on the combination of Activiti and machine learning, comprising: a data collection module 10 for collecting multi-dimensional roadshow conference process data and performing data cleaning; a feature extraction module 20 for processing and calculating the multi-dimensional roadshow conference process data and extracting roadshow conference type features, task urgency features and compliance approval personnel task load features; a model construction module 30 for constructing a task allocation model and training the task allocation model based on the extracted features, the model being used for predicting the task approval time of different approval personnel; an integration module 40 for integrating the trained task allocation model into the Activiti process engine, calling the model to calculate the predicted approval time of the candidate approval personnel when the process task reaches the approval node; an allocation module 50 for dynamically allocating tasks according to the predicted approval time and selecting the compliance approval personnel with the shortest predicted approval time to perform the current approval task; and an optimization module 60 for periodically collecting new approval data, retraining the task allocation model to iteratively optimize the prediction effect.

[0046] The integrated unit, if realized in the form of a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solutions of the application or the entire or part of the technical solutions that essentially contribute to the prior art can be embodied in the form of a software product, which is stored in a storage medium and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the various embodiment methods of the application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program code storage media.

[0047] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer readable storage medium, and when executed, can include the processes of the above-mentioned embodiments of each method. Any reference to memory, storage, database or other medium used in each embodiment of the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0048] The above is only an embodiment of the present application, and does not limit the patent scope of the present application. Any equivalent transformation or direct or indirect application in the related technical field using the content of the present application specification and drawings is also included in the patent protection scope of the present application.

Claims

1. An intelligent process task allocation method based on Activiti and machine learning, characterized in that the steps include: Collect multi-dimensional roadshow meeting process data and perform pre-processing; Processing and calculating the multi-dimensional roadshow meeting process data, and extracting roadshow meeting type characteristics, task urgency characteristics, and compliance reviewer task load characteristics; Building a task allocation model and training the task allocation model based on the extracted features, wherein the model is used to predict task approval time for different approvers; Integrate the trained task allocation model into the Activiti process engine. When the process task reaches the approval node, call the model to calculate the predicted approval time of the candidate approvers. Dynamically assign tasks based on predicted approval time and select the compliance approver with the shortest predicted approval time to perform the current approval task; Regularly collect new approval data and retrain the task allocation model to iteratively optimize the prediction results.

2. The intelligent process task allocation method according to claim 1, characterized in that: The pre-processing steps specifically include: During the multi-dimensional roadshow meeting process data integration, the data source and processing logic were recorded, sensitive information was desensitized, business-related features were retained, and missing data values ​​were eliminated; One-hot encoding is used to convert categorical data into unique binary vectors, and numerical data is normalized to fall into the [0,1] interval.

3. The intelligent process task allocation method according to claim 1, characterized in that: The multi-dimensional roadshow meeting process data includes: Roadshow meeting type, task submission time, meeting start time, approver information, professional field matching, online status, approval time, and the number of tasks pending approval by the approver; The roadshow meeting type is used to determine the task workload according to different types; The task submission time and the meeting start time are used to calculate the task urgency; The professional field matching degree is used to evaluate the suitability of the reviewer and the task type, and the online status is used to determine the real-time availability of the reviewer. The professional field matching degree and the online status are combined to determine whether the reviewer is compliant. The approval time and the number of tasks to be processed by the approver are used to calculate the approver load.

4. The intelligent process task allocation method according to claim 3, characterized in that: The step of judging whether the approval personnel are compliant by combining the professional field matching degree and the online status specifically includes: Determine whether the professional field of the approver matches the roadshow meeting type. If the professional field of the approver includes the roadshow meeting type, the professional field match is recorded as 1, otherwise it is recorded as 0; Determine whether the approver is online. If online, it is recorded as 1, otherwise it is recorded as 0; If the type of professional field matches the meeting type and the approver is online, the approver is deemed compliant.

5. The intelligent process task allocation method according to claim 3, characterized in that: The calculation steps of the task urgency feature specifically include: Get the time difference between the task submission time and the meeting start time; Converting the time difference into an urgency value using a time decay function, wherein the smaller the time difference, the higher the urgency value; Random noise that conforms to the normal distribution is introduced into the calculation process to simulate the uncertainty in actual approval scenarios.

6. The intelligent process task allocation method according to claim 5, characterized in that: The calculation method of the compliance reviewer's task load characteristics includes: Count the number of tasks awaiting approval by the current approver; Calculate the urgency of each pending task and sort them from high to low urgency; Calculate the normalized task load value based on the current task's position in the sorting and the historical maximum number of tasks.

7. The intelligent process task allocation method according to claim 1, characterized in that: The steps of constructing the task allocation model specifically include: A linear regression model was constructed based on the characteristics of task type, task urgency, and task load of compliance reviewers. Perform a logarithmic transformation on the approval time to ensure that the predicted time is positive, and introduce cross-term features to optimize the prediction; The model parameters are solved by the least squares method to obtain the optimal parameters.

8. The intelligent process task allocation method according to claim 1, characterized in that: The steps of integrating the trained task allocation model into the Activiti process engine specifically include: Configure a custom listener in the BPMN file and trigger the listener when the process task reaches the approval node.

9. The intelligent process task allocation method according to claim 8, characterized in that: The steps of customizing the listener specifically include: Inherit the TaskListener interface of the Activiti engine and rewrite the notify method; Get the context information of the current task in the notify method; Call the task allocation model API, input task type characteristics, task urgency characteristics, and compliance reviewer task load characteristics, obtain the predicted approval time of candidate reviewers, and determine the optimal reviewer; Dynamically set the approver variable through the runtime service interface of the Activiti engine, and update the task assignment through the RuntimeService interface.

10. An intelligent process task allocation system based on the combination of Activiti and machine learning, characterized by: include: Data collection module, used to collect multi-dimensional roadshow meeting process data and perform data cleaning; A feature extraction module is used to process and calculate the multi-dimensional roadshow meeting process data and extract roadshow meeting type features, task urgency features, and compliance reviewer task load features; A model building module is used to build a task allocation model and train the task allocation model based on the extracted features, wherein the model is used to predict the task approval time of different approvers; Integration module, used to integrate the trained task allocation model into the Activiti process engine. When the process task reaches the approval node, the model is called to calculate the predicted approval time of the candidate approvers. The allocation module is used to dynamically allocate tasks based on the predicted approval time and select the compliance approver with the shortest predicted approval time to perform the current approval task; The optimization module is used to regularly collect new approval data and retrain the task allocation model to iteratively optimize the prediction effect.

Images