Process node jump reminding method and device, electronic equipment and storage medium

Abstract

The invention discloses a process node jump reminding method and device, electronic equipment and a storage medium, and relates to the technical field of artificial intelligence, and the method comprises the steps: training a recurrent neural network model through historical process node jump data, and obtaining a trained recurrent neural network model; inputting the current process node parameters into the trained recurrent neural network model, predicting to obtain each candidate node of the next process node and the node weight thereof, and predicting to obtain each candidate person of the next process node and the person weight thereof; screening out a target node weight and a target personnel weight from the node weights and the personnel weights by using a long and short memory neural network; and sorting the corresponding candidate nodes and candidates according to the weights of the target nodes and the weights of the target persons, and outputting a sorting result and reminding information carrying guidance operation. According to the method, the weights of candidate nodes and candidates are predicted and optimized, then sorting is carried out, reminding information is provided, user operation is guided, and then the process skipping efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to a process jump reminder method, device, electronic device and storage medium. Background Art

[0002] There are many uncertainties in the subsequent selection of jump nodes and candidates in the process engine. The specific operation is mainly manual. In the initial operation, free jumps will cause fluctuations and changes in the gradient. During model training, too large a gradient will cause gradient explosion, while too small a gradient will cause gradient disappearance. Due to the limitations of the initial organizational scale and data volume of the process engine, there is less data, resulting in low practicality of the training model. In addition, from the perspective of business operations, the process engine selects jump nodes and candidates based on personal understanding and cognition, which has many uncertainties. In addition, there are too many options, which requires more people to operate, increasing the complexity of the operation and reducing the accuracy of the operation. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a process jump reminder method, device, electronic device and storage medium to provide reminder information for process jumps and improve the accuracy and efficiency of process jumps.

[0004] To achieve the above purpose, one aspect of an embodiment of the present application provides a process node jump reminder method, the method comprising the following steps:

[0005] The recurrent neural network model is trained using the historical process node jump data to obtain a trained recurrent neural network model;

[0006] Input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight;

[0007] Filtering the target node weight and the target personnel weight from the node weights and the personnel weights using a long short-term memory neural network;

[0008] The corresponding candidate nodes and candidates are sorted according to the weights of the target nodes and the target personnel, and the sorting results and reminder information carrying guiding operations are output.

[0009] In some embodiments, the method of training the recurrent neural network model using the historical process node jump data to obtain the trained recurrent neural network model includes the following steps:

[0010] Classifying the historical process node data to obtain classified historical process node data;

[0011] First, the classified historical process node data is cleaned, converted and enhanced, and then divided into a training set and a test set according to a set ratio;

[0012] The number of neurons and the loss function of the recurrent neural network model are set, and the recurrent neural network model is trained on the training set using the Adam optimizer; wherein the training process uses forward propagation calculation;

[0013] Tuning the learning rate and regularization parameters of the recurrent neural network model on the test set, and then evaluating and verifying the recurrent neural network model;

[0014] The verified recurrent neural network model is determined as the trained recurrent neural network model.

[0015] In some embodiments, the classifying the historical process node data to obtain the classified historical process node data includes the following steps:

[0016] Determine the hierarchical proportion of process definitions, non-jump nodes, and jump nodes in the historical process node data;

[0017] The historical process node data is classified into process definition data, non-jump node data and jump node data in turn according to the stratification proportion.

[0018] In some embodiments, the classifying the historical process node data into process definition data, non-jump node data, and jump node data in sequence according to the hierarchical proportions includes the following steps:

[0019] According to the process definition template and the hierarchical proportion of the process definition, the process definition data is obtained by classifying the historical process node data;

[0020] In the historical process node data except the process definition data, the non-jump node data is obtained by classifying the historical process node data according to the conditional node definition and the hierarchical proportion of non-jump nodes;

[0021] The remaining historical process node data is determined as the jump node data.

[0022] In some embodiments, the use of the long short-term memory neural network to select the target node weight and the target personnel weight from each of the node weights and the personnel weights includes the following steps:

[0023] Determine the network layer weights of the long short-term memory neural network according to the classified historical process node data;

[0024] Outputting a target threshold using the long short-term memory neural network that determines the network layer weights;

[0025] The target node weight and the target personnel weight that are greater than the target threshold are screened out from the various node weights and the personnel weights.

[0026] In some embodiments, the step of sorting the candidate nodes and the candidates according to the target node weights and the target personnel weights includes the following steps:

[0027] Calculate the priorities of the corresponding candidate nodes and candidates according to the weights of the target nodes and the weights of the target personnel;

[0028] According to the priority, the candidate nodes and the candidates corresponding to the target node weights and the target personnel weights are sorted from high to low.

[0029] In some embodiments, the method further comprises the following steps:

[0030] The candidate nodes and the candidates selected by the user according to the sorting result and the reminder information are used as training data, and the trained recurrent neural network model is trained using the training data.

[0031] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application provides a process node jump reminder device, the device comprising:

[0032] A model training unit, used for training a recurrent neural network model using historical process node jump data to obtain a trained recurrent neural network model;

[0033] A node prediction unit, used to input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight;

[0034] A weight screening unit, used for screening target node weights and target personnel weights from each of the node weights and the personnel weights by using a long short-term memory neural network;

[0035] The node sorting unit is used to sort the corresponding candidate nodes and the candidates according to the weights of the target nodes and the target personnel, and output the sorting results and reminder information carrying guiding operations.

[0036] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned process node jump reminder method when executing the computer program.

[0037] To achieve the above-mentioned purpose, another aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above-mentioned process node jump reminder method is implemented.

[0038] The embodiments of the present application include at least the following beneficial effects:

[0039] This application can use historical process node jump data to train a recurrent neural network model to obtain a trained recurrent neural network model; input the current process node parameters into the trained recurrent neural network model to predict the candidate nodes and corresponding node weights of the next process node, and predict the candidates and corresponding personnel weights of the next process node; use the long and short memory neural network to filter out the target node weight and target personnel weight from the node weights and personnel weights; sort the corresponding candidate nodes and candidates according to the target node weights and target personnel weights, and output the sorting results and reminder information that guides the operation. By predicting the candidate nodes and candidates of the next process node through the recurrent neural network model, and optimizing the weights of the candidate nodes and candidates using the long and short memory network, accurate candidate nodes and candidates can be obtained, and the candidate nodes and candidates can also be sorted and reminder information can be provided to guide user operations, thereby improving the efficiency of process jumps. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0041] Figure 1 A flowchart of a process jump reminder method provided in an embodiment of the present application;

[0042] Figure 2 Provided for the embodiments of this application Figure 1 The step flow chart of S1;

[0043] Figure 3 Provided for the embodiments of this application Figure 1 The step flow chart of S3 in the figure;

[0044] Figure 4 An example flowchart of a process jump reminder method provided in an embodiment of the present application;

[0045] Figure 5 An example diagram of the structure of a long short-term memory neural network provided in an embodiment of the present application;

[0046] Figure 6 A composition diagram of a hierarchical adjustment model for historical process node data provided in an embodiment of the present application;

[0047] Figure 7 A schematic diagram of the structure of a process jump reminder device provided in an embodiment of the present application;

[0048] Figure 8 A schematic diagram of the structure of a model training unit provided in an embodiment of the present application;

[0049] Fig. 9 A schematic diagram of the structure of a weight screening unit provided in an embodiment of the present application;

[0050] Fig.10 A schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0051] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are only examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the attached claims.

[0052] It is understood that the terms "first", "second", etc. used in this application can be used to describe various concepts in this article, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another concept. For example, without departing from the scope of the embodiment of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein can be interpreted as "at the time of" or "when" or "in response to determination".

[0053] The terms "at least one", "multiple", "each", "any", etc. used in this application, at least one includes one, two or more, multiple includes two or more, each refers to each of the corresponding multiple, and any refers to any one of the multiple.

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0055] Before describing the embodiments of the present application in detail, some nouns and terms involved in the embodiments of the present application are first described. The nouns and terms involved in the embodiments of the present application are subject to the following interpretations:

[0056] A recurrent neural network (RNN) is a type of neural network specifically designed to process sequence data. Unlike traditional feedforward neural networks, RNNs have cyclic connections that allow information to be continuously transmitted within the network, thereby modeling and processing sequence data.

[0057] Long Short-Term Memory (LSTM) is a time-recurrent neural network that can solve the long-term dependency problem of general RNN.

[0058] The embodiment of the present application provides a process jump reminder method, which relates to the field of artificial intelligence technology. The process jump reminder method provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or it can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured to provide cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms and other basic cloud computing services. The cloud server, the server can also be a node server in the blockchain network; the software can be an application that implements the process jump reminder method, etc., but is not limited to the above forms.

[0059] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0060] Reference Figure 1 , the embodiment of the present application provides a process node jump reminder method, which may include but is not limited to S1 to S4, as follows:

[0061] S1: Use the historical process node jump data to train the recurrent neural network model to obtain the trained recurrent neural network model.

[0062] Reference Figure 2 , S1 may further include S11 to S15:

[0063] S11: Classify the historical process node data to obtain classified historical process node data.

[0064] Optionally, S11 may include S111-S112:

[0065] S111: Determine the hierarchical proportions of process definitions, non-jump nodes, and jump nodes in the historical process node data.

[0066] S112: Classifying the historical process node data into process definition data, non-jump node data and jump node data in sequence according to the stratification proportions.

[0067] As a further implementation, S112 may include S1121 to S1123:

[0068] S1121: obtaining the process definition data from the historical process node data according to the process definition template and the hierarchical proportion of the process definition;

[0069] S1122: In the historical process node data except the process definition data, classify the historical process node data according to the conditional node definition and the hierarchical proportion of non-jump nodes to obtain the non-jump node data;

[0070] S1123: Determine the remaining historical process node data as the jump node data.

[0071] S12: First, the classified historical process node data is cleaned, converted and enhanced, and then divided into a training set and a test set according to a set ratio.

[0072] S13: Setting the number of neurons and loss function of the recurrent neural network model, and using the Adam optimizer to train the recurrent neural network model on the training set; wherein the training process uses forward propagation calculation.

[0073] S14: Tuning the learning rate and regularization parameters of the recurrent neural network model on the test set, and then evaluating and verifying the recurrent neural network model.

[0074] S15: Determine the verified recurrent neural network model as the trained recurrent neural network model.

[0075] S2: Input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight.

[0076] S3: Utilize the long short-term memory neural network to select the target node weight and the target personnel weight from the various node weights and the personnel weights.

[0077] Reference Figure 3 , S3 may further include S31 to S33:

[0078] S31: Determine the network layer weights of the long short-term memory neural network according to the classified historical process node data;

[0079] S32: Outputting a target threshold using the long short-term memory neural network that determines the network layer weights;

[0080] S33: Filter out the target node weight and the target personnel weight that are greater than the target threshold from the various node weights and the personnel weights.

[0081] S4: Sort the corresponding candidate nodes and candidates according to the target node weights and target personnel weights, and output the sorting results and reminder information carrying guidance operations.

[0082] Furthermore, S4 may further include S41-S42:

[0083] S41: Calculating the priorities of the corresponding candidate nodes and candidates according to the weights of the target nodes and the weights of the target personnel;

[0084] S42: Sort the candidate nodes and the candidates corresponding to the target node weights and target personnel weights from high to low according to the priorities.

[0085] In order to maintain the timeliness of the recurrent neural network, the embodiment of the present application may further include the following steps:

[0086] S5: Using the candidate nodes and the candidates selected by the user according to the ranking result and the reminder information as training data, and using the training data to train the trained recurrent neural network model.

[0087] Next, the solution of the embodiment of the present application will be described in detail and explained in conjunction with specific application examples:

[0088] In order to simplify the selection of nodes and users when the process jumps, improve the accuracy and efficiency of operations, provide accurate reminder content, and avoid the problem of inaccurate sorting priority caused by gradient explosion or disappearance, this embodiment improves the accuracy of training by auxiliary prediction and optimization sorting, as well as gradient clipping. It solves the problem of accurate guidance of process candidate nodes and personnel when the amount of data is small.

[0089] This embodiment provides a process node jump reminder method, including: constructing a recurrent neural network model, collecting historical process node jump data, first standardizing the data according to the process specifications and definitions, and then training the recurrent neural network (RNN) model to form a prediction model. Select the cross entropy loss function, determine the LSTM network layer, and optimize the configuration content of the standard process so that the output value of the loss function is less than a predetermined threshold. According to the fallback habit of the operation, the threshold and pre-jump of the post-jump are distinguished.

[0090] Optionally, for the normal operation process, the relevant data is preprocessed, a data matrix is ​​constructed according to the process nodes and candidates, and the data relationship between the previous and next nodes and the time series is established according to the predetermined direction of the process, and converted into the data format of the recurrent neural network model.

[0091] Optionally, for non-predetermined process definitions, their thresholds are set as dynamic parameters during coding to facilitate optimization and adjustment. In addition, data versions are constructed to support differential comparison of data in time series to achieve data timeliness. Optionally, a subclass function of the framework Module class is created, and the process nodes and candidates are brought in to initialize the network model, and then a forward function is written to define the processing flow of the neural network model.

[0092] Furthermore, when executing an operation, the user operation parameters are brought in, and the model calculation is performed on all nodes and candidates other than the current node, and their thresholds are determined and converted into priorities. Candidates with higher priorities are displayed first, and the associated information and case features of each node link are displayed for users to make reference and selection. For the predicted priority and preset metadata, the associated ratio is extracted to supplement the reminder content.

[0093] Furthermore, the operation metadata is obtained according to the process definition nodes and candidates, the business and engine configuration data input by the user is received, and a hierarchical model is established for the nodes and candidates. The nodes and candidates are layered according to the logic and organizational structure of the business process, and the priority is calculated for each node and candidate, considering factors such as task urgency, candidate availability, historical performance, etc. According to the calculated priority, the corresponding priority data is marked for each node and candidate. Then the priority is brought in as a weighted method to optimize the weight of the LSTM network layer.

[0094] Furthermore, the nodes and personnel candidates are displayed in order of priority, so that the user can see the data with higher priority first, and prompt information is provided to guide the user's operation and remind the user to perform the operation.

[0095] Furthermore, we collect users’ operation data, such as the time to complete tasks, candidates selected, decisions made, and other related data, and use a recurrent neural network (RNN) model to iteratively train the collected operation data, and continuously optimize the model to adapt to users’ operation habits and changes in business processes.

[0096] When selecting nodes and personnel in the workflow, this embodiment uses the trained recurrent neural network model to provide optional content for the nodes and candidates of the next operation link, provide a basis for related decision-making, facilitate users to operate quickly, make decisions accurately, and improve work efficiency.

[0097] Furthermore, this embodiment provides a more specific process node jump reminder method, and its step flow chart can refer to Figure 4 .

[0098] Specifically, this embodiment may include:

[0099] S110: Arrange the process definition and form the first layer of rule data according to the template definition.

[0100] Specifically, data is obtained from the process definition and organized according to the process template to form the first layer of rule data. The subsequent two layers of data are governed based on the first layer, and the jump data is distinguished and marked according to the rules.

[0101] S120: According to the process definition and process execution, the irregular jump data is distinguished, and the execution supervision data is organized into the second layer data.

[0102] Specifically, according to the specific execution data of the process definition, the irregular jump data is distinguished, and each process link and executor is compared with the definition, and organized into the second-level execution supervision data.

[0103] S130: Classify normal data and candidate process node execution as third-level data.

[0104] Specifically, according to the process definition, the normally executed nodes and candidate data are sorted and classified into the third level data.

[0105] S140: The marked three-layer data is distinguished and the recurrent neural network is trained, and the network layer weights of the long short-term memory neural network are confirmed.

[0106] Specifically, the three-layer data is labeled, refined according to the operation sequence and the previous and next nodes of the process, and brought into the recurrent neural network for training. At the same time, operations are performed according to the process nodes and candidates, the cross entropy loss function is selected, and the forget gate mechanism is adopted. The sigmoid activation function is used to output a value between 0 and 1 to determine the threshold of the next node when the node jumps.

[0107] S150: Use hierarchical priorities, adjust the loss function threshold, and output optional nodes and candidates according to weights.

[0108] Specifically, the preceding node and candidate of the jump process, the subsequent node and candidate parameters of the current node are brought in, run through the model, and the priority of the jump node is calculated through the hierarchical model. The priority is converted into a weight, and the priority of the jump node and candidate is determined together with the LSTM threshold so that the system can display it on the interface and facilitate user selection.

[0109] Figure 5 This is a diagram showing an example of the structure of a long short-term memory neural network provided for this embodiment.

[0110] The node jump of the process and the final candidate selection have certain internal rules. The priority of the process definition and the final execution result is predicted and calculated, which can remind and guide users to make choices, so as to improve the accuracy of user operations. This embodiment uses the pytorch framework to train process definition data, prediction data and historical data.

[0111] Specifically, this embodiment may include:

[0112] S210 Input data preparation stage: First, relevant data is preprocessed, including process definition data, process prediction data, and historical data, and the preprocessed data is converted into JSON tree format;

[0113] Secondly, in the development environment, the relevant data is reversed into row data of the equivalent path, and the data class subclass of the PyTorch framework is used to load and encapsulate the data. All data is loaded into the array through the initialization function of the subclass.

[0114] Again, encode the target node and candidate data set, build a data matrix through the node length and historical candidate length, and generate neural network data. Use a unified encoding to represent different data. The entire data array is represented by 0-1, and the length is the number of nodes and the number of historical candidates. The encoding method is an array, which is specifically reflected in an array, such as [0,1,0,0,…,0,n]. The len function is rewritten in a dynamic length method to return the node and candidate data once as defined.

[0115] Secondly, a neural network model and an LSTM network layer are constructed, and the linear m jump nodes and n candidates are input as parameters. When two levels of hidden layers are constructed, hierarchical adjustment input is constructed in accordance with the S220 method and superimposed on the input layer S210.

[0116] S220 hierarchical adjustment: Hierarchical adjustment builds an array according to a fixed 10 levels, focusing on prioritizing the 10 nodes and candidates. The priority of each node and candidate in the formed array is arranged from 1 to 5, such as the generated array data [1,2,3,1,1,1,1,1,4,5], which is used to optimize the sorting and is brought into the loop model training as an input parameter.

[0117] S230 hidden layer: After the superposition of the two input parameter data of S210 and S220, the four inputs corresponding to mn1010 are brought in respectively to form four levels of data of the hidden layer, which constitute the output layer data of the recurrent neural network.

[0118] S240 output layer: LSTM output parameters are mn1010, 1010 are superimposed in the output layer, and weight processing is performed after the LSTM network layer. The model structure is rewritten through the forward function, and mapped to specific nodes in a weighted manner. Finally, the network layer data of mn is output according to the matching weights of 1-5, and each group of data of m and n is returned in the form of the second explicit network layer.

[0119] Considering that the recurrent neural network model trains candidate nodes and candidates at the same time, and adds priority processing for 10 nodes and candidates, in order to make the predicted value and the true value closer, it is necessary to control the threshold of the output value of the loss function. In this embodiment, the Adam algorithm can be used as the optimization function to optimize the network layer weights and the weights of the four network layers at the same time. Considering that the early workflow is limited by the large differences in specific institutions, the data generally used for training is relatively small. In order to support relatively high convergence and reliability of the training model, the optimization function weights are included in the data group of the layered weights, and a mechanism for human intervention is provided. In order to support the assignment of relevant weights by adjusting the threshold when the probability value is small. By default, when the probability is greater than 0.45, this value can be considered to be 1.

[0120] This embodiment uses process node jumps and node personnel selection as data classification and sorting processing, realizes prediction at the data layer, and defines potential relationships between uncertain nodes and candidate personnel selection and previous processing logic and opinions. That is, if there are node flow relationships, personnel scheduling and business changes in the previous sequence, they will be accompanied by the appearance of subsequent nodes and personnel, thereby realizing prediction.

[0121] Figure 6 A composition diagram of a hierarchical adjustment model for historical process node data provided in this embodiment.

[0122] This embodiment can calculate the possible candidate nodes and the priorities of the candidates, and determine the weights and priorities of the nodes and the candidates.

[0123] Process jumps are generally closely related to occurrence nodes and candidates. The node selection and candidate selection after the jump are also closely related to the previous and subsequent nodes. When defining this model, it is processed around three levels: process definition, non-jump nodes, and jump nodes. The layered proportions are 20% for process definition, 25% for conditional nodes, and 55% for jump nodes, and the calculation formula based on the layered value L is obtained:

[0124] L=(21-3.4A)×20%+[0.67(10-6.8B2)]×25%+(37C-232.3)×55%.

[0125] Through data analysis, it is found that general data has a lot of duplication in the nodes where jumps occur and the target nodes of jumps. Based on the hierarchical model, the utilization rate of the C layer can be effectively improved, thereby effectively improving the weight and priority of node and candidate matching. Among them, A is the matching value of the node where the process definition is located, and the calculation formula is: A = number of unjumped nodes × number of unselected candidates number of summary points × number of candidate users A = number of summary points × number of candidate users number of unjumped nodes × number of unselected candidates B = number of selected jump nodes × number of defined jump nodes number of jump predecessor nodes + number of jump selection nodes × selection switch ratio B = number of jump predecessor nodes + number of jump selection nodes number of selected jump nodes × number of defined jump nodes × selection switch ratio C = number of jump occurrence nodes + number of jump target nodes total number of process nodes C = total number of process nodes number of jump occurrence nodes + number of jump target nodes.

[0126] S310 Process definition: Define the main process according to the business model, and define the nodes and predefined candidates of the process through the process classification defined by the template. The process definition uses the template to adapt the final abnormal jump and the difference comparison of dynamic personnel selection to achieve the differentiated distinction and classification of the comparison jump, and match the matching value of the defined candidate and the actual candidate.

[0127] S320 Conditional Nodes: The process definition conditional nodes are hierarchically classified to define the jump nodes, distinguish the definition of normal jump nodes, mark around the preceding and succeeding nodes, distinguish according to the selection ratio, and mark in combination with historical data. In addition, the jumps of the conditional nodes are distinguished and coordinated, and the coordination of the defined candidates and the actual handlers is processed. Conditional nodes: Process the nodes of the process definition conditional branches, distinguish and mark their specific directions, and record the proportion of the preceding and succeeding nodes. Support positioning by the actual direction of the process, calculate the optional proportion and data analysis pairing after distinguishing the specific direction. While processing the conditional nodes, mark the conditional node jumps as normal jumps to distinguish the subsequent abnormal jump nodes and candidates. For the manual selection of candidates, it is only marked as manual processing of personnel candidates.

[0128] S330 Jump Node: For abnormal flow process nodes in process operation data, first mark the initiating node and the target node, and record the actual candidates of the target node. For the candidates of the corresponding process definition node, calculate the proportion and obtain the specific executor from the instance. Confirm the clearly defined candidates and freely selected users, calculate the matching rate, and perform hierarchical data calculation and processing around the jump node.

[0129] In summary, the present application provides a process node jump reminder method, which predicts the candidate task nodes and personnel of the process engine by training the recurrent neural network model, provides users with simplified operation functions, and improves the efficiency and accuracy of the operation, including: 1. Collect and organize historical data, standardize data processing according to process definition and specification, select recurrent neural network (RNN) for data training, and select cross entropy loss function to train the prediction model; 2. Bring in parameters to the model for the next task node and pre-selection of personnel, use the recurrent neural network model to predict the reminder nodes and personnel, and arrange the priorities according to the weights; 3. Obtain relevant data according to the process definition nodes and personnel, receive user input business and engine configuration data, use the hierarchical adjustment model to calculate the priority of all nodes and candidate personnel, and determine the priority of the associated nodes and candidates according to the priority according to the adjustment of the network layer weight of the LSTM; 4. Arrange according to the node and personnel candidate priority, and carry prompt information to guide the user operation. 5. Input the current user's operation data into the recurrent neural network model for iterative training. This application can provide advance predictions and prompts for the subsequent candidate nodes and candidates of the process engine, provide operation reminders, and improve the accuracy and efficiency of the operation.

[0130] Reference Figure 7 The embodiment of the present application further provides a process node jump reminder device, which can implement the above process jump reminder method, and the device includes:

[0131] A model training unit, used for training a recurrent neural network model using historical process node jump data to obtain a trained recurrent neural network model;

[0132] A node prediction unit, used to input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight;

[0133] A weight screening unit, used for screening target node weights and target personnel weights from each of the node weights and the personnel weights by using a long short-term memory neural network;

[0134] The node sorting unit is used to sort the corresponding candidate nodes and the candidates according to the weights of the target nodes and the target personnel, and output the sorting results and reminder information carrying guiding operations.

[0135] Reference Figure 8 In some embodiments, the model training unit includes:

[0136] A data classification unit, used to classify the historical process node data to obtain classified historical process node data;

[0137] A data preprocessing unit, used to first clean, convert and enhance the classified historical process node data, and then divide it into a training set and a test set according to a set ratio;

[0138] A model training subunit, used to set the number of neurons and loss function of the recurrent neural network model, and train the recurrent neural network model on the training set using an Adam optimizer; wherein the training process uses forward propagation calculation;

[0139] A model verification unit, used to tune the learning rate and regularization parameters of the recurrent neural network model on the test set, and then evaluate and verify the recurrent neural network model;

[0140] A model determination unit is used to determine the verified recurrent neural network model as the trained recurrent neural network model.

[0141] In some embodiments, the data classification unit includes:

[0142] A proportion determination unit, used to determine the hierarchical proportion of process definitions, non-jump nodes and jump nodes in the historical process node data;

[0143] The proportion classification unit is used to classify the historical process node data into process definition data, non-jump node data and jump node data in sequence according to the hierarchical proportion.

[0144] In some embodiments, the proportion classification unit includes:

[0145] A first proportion classification subunit is used to classify the historical process node data according to the process definition template and the hierarchical proportion of the process definition to obtain the process definition data;

[0146] A second proportion classification subunit is used to classify the historical process node data in addition to the process definition data according to the conditional node definition and the hierarchical proportion of the non-jump nodes to obtain the non-jump node data;

[0147] The third proportion classification sub-unit is used to determine the remaining historical process node data as the jump node data.

[0148] Reference Fig. 9 In some embodiments, the weight screening unit includes:

[0149] A weight determination unit, used to determine the network layer weights of the long short-term memory neural network according to the classified historical process node data;

[0150] A threshold determination unit, configured to output a target threshold using the long short-term memory neural network that determines the network layer weights;

[0151] The weight screening subunit is used to screen out the target node weight and the target personnel weight that are greater than the target threshold from the various node weights and the personnel weights.

[0152] In some embodiments, the node sorting unit includes:

[0153] A priority calculation unit, used for calculating the priority of the corresponding candidate nodes and the candidates according to the weight of each target node and the weight of the target person;

[0154] The node sorting subunit is used to sort the candidate nodes and the candidates corresponding to the target node weights and the target personnel weights from high to low according to the priority.

[0155] In some embodiments, the apparatus further comprises:

[0156] An iterative training unit is used to use the candidate nodes and the candidates selected by the user according to the sorting result and the reminder information as training data, and use the training data to train the trained recurrent neural network model.

[0157] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0158] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned process jump reminder method when executing the computer program. The electronic device can be any intelligent terminal including a tablet computer, a car computer, etc.

[0159] It can be understood that the contents of the above method embodiments are all applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0160] See also Fig.10 , Fig.10 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0161] The processor 1001 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0162] The memory 1002 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 1002 can store an operating system and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 1002, and the processor 1001 calls and executes the process jump reminder method of the embodiment of this application;

[0163] Input / output interface 1003, used to implement information input and output;

[0164] The communication interface 1004 is used to realize the communication interaction between the device and other devices. The communication can be realized through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.);

[0165] A bus 1005 , which transmits information between various components of the device (e.g., the processor 1001 , the memory 1002 , the input / output interface 1003 , and the communication interface 1004 );

[0166] The processor 1001 , the memory 1002 , the input / output interface 1003 and the communication interface 1004 are connected to each other in communication within the device via the bus 1005 .

[0167] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned process jump reminder method is implemented.

[0168] It can be understood that the contents of the above method embodiments are all applicable to the present storage medium embodiments, the functions specifically implemented by the present storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0169] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0170] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0171] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0172] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0173] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0174] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0175] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0176] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0177] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0178] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0179] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0180] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A process node jump reminder method, characterized in that: The method comprises the following steps: The recurrent neural network model is trained using the historical process node jump data to obtain a trained recurrent neural network model; Input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight; Filtering the target node weight and the target personnel weight from the node weights and the personnel weights using a long short-term memory neural network; The corresponding candidate nodes and candidates are sorted according to the weights of the target nodes and the target personnel, and the sorting results and reminder information carrying guiding operations are output.

2. A process node jump reminder method according to claim 1, characterized in that: The method of training the recurrent neural network model using the historical process node jump data to obtain the trained recurrent neural network model includes the following steps: Classifying the historical process node data to obtain classified historical process node data; First, the classified historical process node data is cleaned, converted and enhanced, and then divided into a training set and a test set according to a set ratio; The number of neurons and the loss function of the recurrent neural network model are set, and the recurrent neural network model is trained on the training set using the Adam optimizer; wherein the training process uses forward propagation calculation; Tuning the learning rate and regularization parameters of the recurrent neural network model on the test set, and then evaluating and verifying the recurrent neural network model; The verified recurrent neural network model is determined as the trained recurrent neural network model.

3. A process node jump reminder method according to claim 2, characterized in that: The process of classifying the historical process node data to obtain the classified historical process node data comprises the following steps: Determine the hierarchical proportion of process definitions, non-jump nodes, and jump nodes in the historical process node data; The historical process node data is classified into process definition data, non-jump node data and jump node data in turn according to the stratification proportion.

4. A process node jump reminder method according to claim 3, characterized in that: The step of classifying the historical process node data into process definition data, non-jump node data, and jump node data in sequence according to the hierarchical proportions includes the following steps: According to the process definition template and the hierarchical proportion of the process definition, the process definition data is obtained by classifying the historical process node data; In the historical process node data except the process definition data, the non-jump node data is obtained by classifying the historical process node data according to the conditional node definition and the hierarchical proportion of non-jump nodes; The remaining historical process node data is determined as the jump node data.

5. A process node jump reminder method according to claim 2, characterized in that: The method of using the long short-term memory neural network to select the target node weight and the target personnel weight from each of the node weights and the personnel weights includes the following steps: Determine the network layer weights of the long short-term memory neural network according to the classified historical process node data; Outputting a target threshold using the long short-term memory neural network that determines the network layer weights; The target node weight and the target personnel weight that are greater than the target threshold are screened out from the various node weights and the personnel weights.

6. A process node jump reminder method according to claim 1, characterized in that: The step of sorting the candidate nodes and the candidates according to the target node weights and the target personnel weights includes the following steps: Calculate the priorities of the corresponding candidate nodes and candidates according to the weights of the target nodes and the weights of the target personnel; According to the priority, the candidate nodes and the candidates corresponding to the target node weights and the target personnel weights are sorted from high to low.

7. A process node jump reminder method according to any one of claims 1 to 6, characterized in that: The method further comprises the following steps: The candidate nodes and the candidates selected by the user according to the sorting result and the reminder information are used as training data, and the trained recurrent neural network model is trained using the training data.

8. A process node jump reminder device, characterized in that: The device comprises: A model training unit, used for training a recurrent neural network model using historical process node jump data to obtain a trained recurrent neural network model; A node prediction unit, used to input the current process node parameters into the trained recurrent neural network model, predict each candidate node of the next process node and the corresponding node weight, and predict each candidate of the next process node and the corresponding personnel weight; A weight screening unit, used for screening target node weights and target personnel weights from each of the node weights and the personnel weights by using a long short-term memory neural network; The node sorting unit is used to sort the corresponding candidate nodes and the candidates according to the weights of the target nodes and the target personnel, and output the sorting results and reminder information carrying guiding operations.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements a process node jump reminder method as described in any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, a process node jump reminder method according to any one of claims 1 to 7 is implemented.

Images