Business status prediction method and device for robotic process automation data

Abstract

The application relates to the technical field of robot process automation, and discloses a business state prediction method and device for robot process automation data. The application constructs a business event relationship network, obtains target fusion features of mutual fusion of various target business data, combines target confidence features of the target business data under various preset state data, and then utilizes a pre-trained business state prediction model to obtain business state prediction results of the various target business data, so that the business state of robot process automation for human-machine collaboration can be accurately analyzed.

Description

Technical Field

[0001] This invention relates to the field of robotic process automation technology, and specifically to a method and apparatus for predicting the business status of robotic process automation data. Background Technology

[0002] With the acceleration of enterprise digital transformation, Robotic Process Automation (RPA) technology has been widely used in various industries due to its ability to efficiently execute repetitive and well-defined tasks. RPA technology automates business processes by simulating manual operations, greatly improving work convenience.

[0003] In related technologies, RPA business data from different sources are generally analyzed separately. However, due to the diversity of RPA business data, it is impossible to fully explore the business data of mutual collaboration between robots and humans through separate analysis, which in turn makes it impossible to accurately analyze the business status of human-machine collaborative processing of business data from different sources. Summary of the Invention

[0004] According to a first aspect, the present invention provides a method for predicting the business status of robotic process automation data, the method comprising: Acquire various target business data, which are robotic process automation business data arranged in time series; For each type of target business data, determine the target confidence characteristics of the target business data under multiple preset state data conditions; Based on multiple target business data, a business event relationship network is constructed to obtain the target fusion characteristics of the mutual integration of multiple target business data; The target fusion features and target confidence features are input into the business status prediction model to obtain business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

[0005] In some specific implementations, the various target business data include: business log data, business trajectory data, and business interaction data. Among them, business interaction data includes: human-computer intervention type data, human-computer interaction display data, human-computer interaction time data, and operator identification data. Acquire various target business data, including: Based on the operator's identification data, the human-machine intervention type data, human-machine interaction display data, human-machine interaction time data, business log data, and business trajectory data are associated to obtain the business log data, business trajectory data, and business interaction data, which together form a triplet of data. Arrange the triplet data according to the time series to obtain various target business data.

[0006] In some specific implementations, the various preset state data include: preset business state data, preset system state data, and preset abnormal state data. For each type of target business data, the target confidence feature of the target business data under the various preset business state data is determined, including: For each type of target business data, determine the confidence level parameters of the target business data under preset business state data, preset system state data, and preset abnormal state data, including: For each type of target business data, extract the corresponding business log features, business trajectory features, and business interaction features. For each type of target business data, the business log features, business trajectory features, and business interaction features corresponding to the target business data are input into the confidence coding model for recognition, so as to obtain the confidence parameters of the target business data under preset business state data, preset system state data, and preset abnormal state data; wherein, the confidence coding model is a pre-trained neural network model; For each type of target business data, the target confidence features of the target business data under preset business status data, preset system status data, and preset abnormal status data are determined based on the confidence parameters of the target business data under preset business status data, preset system status data, and preset abnormal status data.

[0007] In some specific implementations, the target confidence characteristics of the target business data under preset business state data, preset system state data, and preset abnormal state data are determined, including: First-level confidence level, second-level confidence level, third-level confidence level, and fourth-level confidence level are predefined; For each type of target business data, based on the first level of confidence, the second level of confidence, the third level of confidence, and the fourth level of confidence, the confidence level of the target business data under the preset business state data, preset system state data, and preset abnormal state data is determined, and the target confidence feature of the target business data under the preset business state data, preset system state data, and preset abnormal state data is obtained.

[0008] In some specific implementations, the target confidence characteristics of the target business data under preset business state data, preset system state data, and preset abnormal state data are determined, including: For each type of target business data, the confidence parameters of the target business data under preset business state data, preset system state data, and preset abnormal state data are normalized to obtain the data state vector corresponding to the target business data. For each type of target business data, the data state vector corresponding to the target business data is used as the target confidence feature of the target business data under preset business state data, preset system state data, and preset abnormal state data.

[0009] In some specific implementations, a business event relationship network is constructed based on multiple target business data to obtain target fusion features that integrate the various target business data, including: Based on various target business data, obtain the target business characteristics of the robotic process automation business entities, as well as the target relationship characteristics between business entities with business relationships; Map the target business characteristics onto the business network topology to determine the nodes corresponding to the business entities; The target relationship features are mapped onto the network topology to determine the edges corresponding to business entities with existing business relationships. Based on the nodes corresponding to business entities and the edges corresponding to business entities with business relationships, a business event relationship network is constructed that integrates multiple target business features. Extract target fusion features from the business event relationship network by integrating multiple target business features.

[0010] In some specific implementations, the target business characteristics include: business name characteristics, business identifier characteristics, and business attribute characteristics; the business type characteristics include: human-machine processing efficiency characteristics, business process rule characteristics, human-machine operation behavior characteristics, and human-machine business collaboration characteristics; the target relationship characteristics include: business relationship name, business relationship occurrence time, business relationship attribute, and business relationship direction; the business relationship name includes: influence, affected, bound, bound, triggered, and triggered; the business relationship occurrence time is the time when the business relationship is generated; and the business relationship direction is the direction of the two business characteristics based on the business relationship.

[0011] In some specific implementations, embodiments of this application also provide a method for predicting the business status of robotic process automation data, further comprising: Based on the business status prediction results of various target business data, generate data processing efficiency suggestions; Send data processing performance recommendations to the target audience for learning.

[0012] In some specific implementations, data processing performance recommendations include: business log modification recommendations, business trajectory optimization recommendations, and human training optimization recommendations. Based on the business status prediction results of various target business data, data processing performance recommendations are generated, including: When anomalies are predicted in business log data, modification suggestions for the business logs are generated based on the attribute information of historical business logs that have not shown anomalies in the historical time period. When abnormalities occur in the predicted business trajectory data, the business entity causing the abnormality is identified using the business event relationship network, and business trajectory optimization suggestions are generated by combining the records in the abnormality handling manual. When a delay in human-computer interaction is predicted in the business interaction data, human-computer operation data with human-computer collaboration capability indicators lower than preset indicators is obtained based on the business interaction data. Combined with the recorded content in the human-computer operation manual, human-computer optimization suggestions are generated.

[0013] According to a second aspect, the present invention also provides a business status prediction device for robotic process automation data, the device comprising: The business data acquisition module is used to acquire various target business data, which are robotic process automation business data arranged in a time series. The first feature determination module is used to determine the target confidence features of the target business data under multiple preset state data for each type of target business data. The second feature determination module is used to construct a business event relationship network based on multiple target business data to obtain target fusion features that integrate multiple target business data. The business status prediction module is used to input the target fusion features and target confidence features into the business status prediction model to obtain business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

[0014] Thirdly, the present invention also provides an electronic device, comprising: The memory and processor are interconnected and communicate with each other. The memory stores computer instructions, and the processor executes the computer instructions to perform the business state prediction method for robotic process automation data in the first aspect or any embodiment of the first aspect.

[0015] Fourthly, the present invention also provides a computer-readable storage medium storing computer instructions for causing a computer to execute the business state prediction method for robotic process automation data in the first aspect or any embodiment of the first aspect.

[0016] The technical solution of this invention has the following advantages: This invention discloses a method and apparatus for predicting the business status of robotic process automation data. The method constructs a business event relationship network to obtain target fusion features of multiple target business data, combines the target confidence features of target business data under multiple preset state data, and then uses a pre-trained business status prediction model to obtain the business status prediction results of multiple target business data. This is beneficial for accurately analyzing the business status of human-machine collaboration in robotic process automation. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0018] Figure 1 This is a flowchart illustrating a method for predicting the business status of robotic process automation data according to an embodiment of the present invention. Figure 2 This is another schematic diagram of the business status prediction method for robotic process automation data according to an embodiment of the present invention; Figure 3 This is a structural block diagram of a business status prediction device for robotic process automation data according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] It is understood that before using the technical solutions disclosed in the various embodiments of the present invention, users should be informed of the types, scope of use, and usage scenarios of the personal information involved in the present invention and their authorization should be obtained in accordance with relevant laws and regulations through appropriate means.

[0021] The terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0022] This application specifically belongs to the field of Robotic Process Automation (RPA) technology. Currently, RPA technology is widely used in enterprise office automation systems to automatically execute repetitive tasks. Related technologies, such as RPA, are mainly used for automated task execution, such as data entry and report generation. In other words, related RPA technologies typically only focus on the execution results of RPA, such as basic indicators like task completion rate and execution time.

[0023] In other words, while current RPA technology can automate tasks, it lacks the ability to deeply analyze the RPA execution process. Consequently, it cannot extract valuable business insights from operation logs, making it difficult for enterprises to quantify the true value of RPA or use RPA data to support business optimization. Furthermore, RPA analysis methods in related technologies often view business logs, business processes, and business interactions in isolation, leading to fragmented and incomplete analysis results. They only show isolated behavioral records, failing to connect them to inter-business relationships, thus hindering enterprises from gaining a comprehensive understanding that spans the entire business and integrates processes and outcomes.

[0024] In summary, related technologies typically analyze RPA business data from different sources separately. However, due to the diversity of RPA business data, individual analysis methods cannot fully uncover the collaborative business data between robots and humans, leading to an inability to accurately analyze the business status of human-machine collaborative processing of business data from different sources. In other words, while the aforementioned technologies can reveal the execution results of RPA, they cannot deeply analyze the technical challenges of broken event causal chains and ambiguous state determinations caused by automation interruptions, human intervention, and asynchronous logging between systems.

[0025] This application provides a method for predicting the business status of robotic process automation data, which can be used in electronic devices such as mobile phones, tablets, desktop computers, laptops, and servers. Figure 1 This is a flowchart of a business status prediction method for robotic process automation data according to an embodiment of the present invention, such as... Figure 1 As shown, the process includes the following steps: Step S101: Obtain various target business data, wherein the target business data is robotic process automation business data arranged in a time series.

[0026] In a specific example, the target business data includes: business log data, business trajectory data, and business interaction data.

[0027] Specifically, business log data records the automated operation data for each step of the RPA process. This business log data includes, but is not limited to, process instance identifiers, process node data, process node actions, operator identifiers, timestamps, status changes, and exception codes in processes such as order approval, risk control review, production flow, and operation and maintenance changes. For example, the process instance identifier is the unique identifier MA for the 1024 order process instance in order approval; the process node data is the current stage of the process instance, which can be supervisor approval or finance payment; the process node actions include, but are not limited to, submit, pass, reject, transfer, cancel, and timeout actions; the operator identifier is the operator's identity information; the timestamp is the precise time when the action occurred; the status change indicates a change from the first state to the second state, such as from pending approval to approved; and the exception codes are encoded information such as process interruption, rejection reasons, retry information, and system errors.

[0028] Business trajectory data records the digital traces left by RPA business processes during execution. The components of this business trajectory data include: process instance, event name, event timestamp, trajectory path, and business attributes. For example, the process instance is "Order 1024"; the event name includes, but is not limited to: "Submit for Approval," "Warehouse Shipment," and "Create Work Order"; the event timestamp is "2024-05-21 09:32:15"; the trajectory path is: Department A accepts → Department B approves → Department C executes; business attributes include, but are not limited to: order amount, approver, cost center, and material code.

[0029] Business interaction data is recorded during the execution of the RPA process, showing the interaction behavior between the operator and the robot or the application that carries the robot's process.

[0030] In a specific example, business interaction data includes: human-computer intervention type data, human-computer interaction display data, human-computer interaction time data, and operator identification data.

[0031] Specifically, human-machine intervention type data includes, but is not limited to, data on manual approval, manual correction, anomaly handling confirmation, and process suspension; human-machine interaction display data includes, but is not limited to, pop-up content, form data, robot input status, and error status displayed on the interactive interface; human-machine interaction time data is the time it takes for the robot to prompt the completion of the manual operation. Operator identification data is the operator's identity information.

[0032] In some specific implementations, step S101 involves acquiring various target service data, including: Step a1: Based on the operator's identification data, associate the human-machine intervention type data, human-machine interaction display data, human-machine interaction time data, business log data, and business trajectory data to obtain the triplet data consisting of business log data, business trajectory data, and business interaction data.

[0033] Step a2: Arrange the triplet data according to the time series to obtain various target business data.

[0034] For example, human-computer interaction data categorized as "manually corrected data" is linked to the "data validation failure" operation record in the RPA business log data and the "B department approval process instance of order 1024" in the business trajectory data, thus achieving the core of "human-machine-business" three-dimensional data integration. Further, based on the linked data, a complete hybrid business process instance containing both automated and manual stages is reconstructed in chronological order, resulting in data with a time-series arrangement. Multiple target business data are represented by W. .

[0035] in, Indicates the first i Triple data corresponding to each time point This represents business log data. This represents business trajectory data. This represents business interaction data.

[0036] Step S102: For each type of target business data, determine the target confidence characteristics of the target business data under multiple preset state data.

[0037] In a specific example, the preset state data includes: preset business state data, preset system state data, and preset abnormal state data.

[0038] Specifically, the preset business status data includes, but is not limited to, data such as operation success, operation failure, network timeout, business approval in progress, and manual correction; the preset system status data includes, but is not limited to, data such as system normal, high latency, database connection pool exhausted, and memory overflow; the preset abnormal status data includes, but is not limited to, data such as network abnormality, permission abnormality, data verification abnormality, and third-party service abnormality.

[0039] In some specific implementations, step S102 above, for each type of target service data, determines the target confidence characteristics of the target service data under multiple preset service state data, including: Step a1: For each type of target business data, determine the confidence level parameter of the target business data under preset business status data, preset system status data, and business abnormal status data.

[0040] Step a2: For each type of target business data, extract the corresponding business log features, business trajectory features, and business interaction features.

[0041] Specifically, each type of target business data is typically presented in the form of structured, semi-structured, and unstructured data. For example, business log data is usually semi-structured data, while business trajectory data and business interaction data are usually presented in both structured and unstructured formats. For unstructured data, Optical Character Recognition (OCR) technology can be used to extract relevant business trajectory features and business interaction features from image-formatted data. Business interaction features include, as in the example above, features such as human-computer intervention type, human-computer interaction display, human-computer interaction time, and operator identification. For structured and semi-structured data, Natural Language Processing (NLP) can be used to extract relevant business log features from text-formatted data. Business log features include, as in the example above, features such as process instance identifiers, node data, node actions, operator identifiers, timestamps, status changes, and exception codes in processes such as order approval, risk control review, production flow, and maintenance changes.

[0042] Step a3: For each type of target business data, input the business log features, business trajectory features, and business interaction features corresponding to the target business data into the confidence coding model for recognition, and obtain the confidence parameters of the target business data under preset business state data, preset system state data, and preset abnormal state data; wherein, the confidence coding model is a pre-trained neural network model.

[0043] Step a4: For each type of target business data, determine the target confidence feature of the target business data under the preset business status data, preset system status data, and preset abnormal status data based on the confidence parameters of the target business data under the preset business status data, preset system status data, and preset abnormal status data.

[0044] Specifically, confidence encoding models include, but are not limited to, logistic regression models, random forests, or multi-classifier models. For example, the confidence parameters of business log data for the order process instance named 1024 are 0.15, 0.80, and 0.05 under preset business status data, preset system status data, and preset abnormal status data, respectively.

[0045] In some specific implementations, step a3 above, determining the target confidence characteristics of the target business data under preset business state data, preset system state data, and preset abnormal state data, includes: Step a31A: Predefine the first level confidence level, the second level confidence level, the third level confidence level, and the fourth level confidence level.

[0046] For example, the confidence level range for the first level is (0, 0.25]; the confidence level range for the second level is (0.25, 0.5]; the confidence level range for the third level is (0.5, 0.75]; and the confidence level range for the fourth level is (0.75, 1).

[0047] Step a32A: For each type of target business data, based on the first level of confidence, the second level of confidence, the third level of confidence, and the fourth level of confidence, determine the confidence level of the confidence parameter of the target business data under the preset business state data, preset system state data, and preset abnormal state data, and obtain the target confidence feature of the target business data under the preset business state data, preset system state data, and preset abnormal state data.

[0048] For example, if the target business data is the business log data in the example above, when the preset business status is the status data of business approval, the confidence parameter corresponding to the preset business status data is 0.15, which belongs to the confidence range of the first level of confidence in the example above; the confidence parameter corresponding to the preset system status is 0.80, which belongs to the confidence range of the fourth level of confidence in the example above; and the confidence parameter corresponding to the preset abnormal status data is 0.05, which belongs to the confidence range of the first level of confidence in the example above.

[0049] In some other specific implementations, step a3 above, determining the target confidence characteristics of the target business data under preset business state data, preset system state data, and preset abnormal state data, includes: Step a31B: For each type of target business data, normalize the confidence parameters of the target business data under preset business state data, preset system state data, and preset abnormal state data to obtain the data state vector corresponding to the target business data.

[0050] Step a32B: For each type of target business data, the data state vector corresponding to the target business data is used as the target confidence feature of the target business data under the preset business state data, preset system state data, and preset abnormal state data.

[0051] For example, the confidence parameters of the business interaction data of the target business data, which is the business process instance named 1024, are 0.1, 0.9 and 0.2 under the preset business state data, preset system state data and preset abnormal state data, respectively. After normalization, we get 0.1 / (0.1+0.9+0.2)≈0.083; 0.9 / (0.1+0.9+0.2)=0.750; 0.2 / (0.1+0.9+0.2)≈0.167. The data state vector corresponding to this target business data is [0.083,0.750,0.167].

[0052] Through the above implementation methods, the target confidence characteristics of the target business data under preset business state data, preset system state data, and preset abnormal state data are determined, which facilitates the acquisition of the initial state characteristics of the target business data.

[0053] Step S103: Based on multiple target business data, construct a business event relationship network to obtain the target fusion features of the mutual integration of multiple target business data.

[0054] In some specific implementations, step S103 above involves constructing a business event relationship network based on multiple target business data to obtain target fusion features that integrate the various target business data, including: Step b1: Based on various target business data, obtain the target business characteristics of the robotic process automation business entities, as well as the target relationship characteristics between business entities with business relationships.

[0055] In a specific example, the target business features include: business type features, business identification features, and business attribute features. The business type features include: human-machine processing efficiency features, business process rule features, human-machine operation behavior features, and human-machine business collaboration features. The target relationship features include: business relationship name, business relationship occurrence time, business relationship attribute, and business relationship direction. The business relationship name includes: influence, affected, bound, bound, triggered, and triggered. The business relationship occurrence time is the time when the business relationship is generated, and the business relationship direction is the direction of the two business features based on the business relationship.

[0056] For example, human-machine processing efficiency characteristics affect business process rules, and business process rules are affected by human-machine processing efficiency characteristics.

[0057] For example, human-machine operation behavior characteristics trigger human-machine business collaboration characteristics, and human-machine business collaboration characteristics are triggered by human-machine operation behavior characteristics.

[0058] For example, human-machine operation behavior characteristics are bound to business process rules, and business process rules are bound to human-machine operation behavior characteristics.

[0059] Step b2: Map the target service characteristics onto the network topology to determine the node corresponding to the service entity.

[0060] Specifically, based on network topology graph theory and combined with the target business characteristics of RPA, the definition of the business event relationship network is carried out, including nodes and node datasets.

[0061] In some specific implementations, step b2 above, which maps the target business characteristics onto a business event relationship network to determine the node corresponding to the business entity, includes: Step b21, the node expression for creating the network topology is formed by the following formula: ; in, For nodes, For node identification, For node type, For z key-value pairs in a node attribute, these z key-value pairs are represented as .

[0062] Step b22: Map the business identifier features of the business entity to the node identifier, the business type features of the business entity to the node type, and the business attribute features of the business entity to the node attribute, to obtain the node corresponding to the business entity.

[0063] For example, a node is defined: ; in, The node identifier feature (node ​​number) represents the node. This indicates the node type, for example: node types include "ca_efficiency" (human-machine processing efficiency feature), "cb_rule" (business process rule feature), "cc_behavior" (human-machine operation behavior feature), and "cd_synergy" (human-machine business collaboration feature). This embodiment supports dynamic definition and expansion of node types. This definition is achieved through (…). , To identify the uniqueness of a node; Representing z key-value pairs of a node's attribute set, in the form of: .

[0064] Furthermore, the nodes corresponding to the business entities constitute a node dataset, and the expression for the node dataset is as follows:

[0065] in, For node datasets, For nodes.

[0066] In this embodiment of the application, during the execution of steps b21-b22 above, node mapping is completed based on business identification features and business attribute features. For example, taking a node corresponding to the human-machine processing efficiency feature as an example, the node type feature is the human-machine processing efficiency feature, the node identification feature is 10, and the mapped node also includes relevant node attribute information, which includes: business name, human-machine processing time, operator identification, robot current status, etc.

[0067] Step b3: Map the target relationship features onto the network topology to determine the edges corresponding to business entities with existing business relationships.

[0068] In some specific implementations, step b3 above maps the target relationship features onto a business event relationship network to determine the edges corresponding to business entities with existing business relationships, including: Step b31: Create edge expressions for edges in the network topology. The edge expression includes edge type, first edge node, second edge node, edge attribute, edge direction, and edge timestamp.

[0069] Step b32: Map the business type feature of the target relationship feature to the edge type, map the first business entity corresponding to the target relationship feature to the first edge node, map the second business entity corresponding to the target relationship feature to the second edge node, map the business relationship attribute of the target relationship feature to the edge attribute, map the business relationship direction of the target relationship feature to the edge direction, and map the business relationship occurrence time of the target relationship feature to the edge timestamp, to obtain the edge corresponding to the business entity with the business relationship.

[0070] Specifically, based on network topology graph theory and combined with the target business characteristics of RPA, the definition of the business event relationship network is carried out, including edges and edge datasets.

[0071] In one alternative implementation, the edge expression is formed by the following formula: ; in, As an edge, These are the two nodes corresponding to the edge, namely the first edge node (edge ​​starting node) and the second edge node (edge ​​reaching node). For edge type, For the direction of the edge, For edge timestamps, , For w key-value pairs in the edge attribute, these w key-value pairs are represented as .

[0072] For example, define an edge:

[0073] in Represent the two nodes of an edge. This indicates the edge type, which includes "affect", "affected", "bind", "bound", "trigger", and "triggered". Dynamic definition and extension of edge types are supported. This indicates the direction of the edge, including four types: undirected, positive, opposite, and bidirectional. This indicates the timestamp information of the edge creation. The w key-value pairs representing the edge attribute set are represented in the form of: .

[0074] In one optional implementation, the edges corresponding to nodes with business relationships constitute an edge dataset, and the expression for the edge dataset is: ; in, For edge datasets, For the edge.

[0075] In this embodiment, edge mapping is completed during the execution of steps b31-b32 described above. For example, the business relationship between human-machine operation behavior features and human-machine business collaboration features is mapped to an edge where the human-machine operation behavior feature triggers the human-machine business collaboration feature. The human-machine operation behavior feature is the first edge node, and the human-machine business collaboration feature is the second edge node. The business relationship between the two is positive, meaning the human-machine operation behavior feature points to the human-machine business collaboration feature. The business relationship name, trigger, and the time when the business relationship occurred are May 15, 2025.

[0076] Step b4: Based on the nodes corresponding to business entities and the edges corresponding to business entities with business relationships, construct a business event relationship network that integrates multiple target business features.

[0077] In some specific implementations, step b4 above, based on the nodes corresponding to business entities and the edges corresponding to business entities with business relationships, determines the target fusion features for the mutual fusion of multiple target business data, including: Step b41: Obtain the network identifier and network creation time of the business event relationship network.

[0078] Specifically, for example, network identifiers can be used The network creation time can be represented by TM.

[0079] Step b42: Based on the nodes of business entities, the edges between business entities with business relationships, the network identifier, and the network creation time, construct a business event relationship network in which business relationships are mutually integrated.

[0080] For example, define a business event relationship network. , in, For business event relationship network, For node datasets, For edge datasets, For network identification, The time of network creation.

[0081] Step b43: Extract target fusion features from the business event relationship network, which are the fusion features of multiple target business features.

[0082] Within the established business event relationship network, detailed target fusion features can be quickly extracted for the AA equipment purchase order, which involves human-machine processing efficiency features, business process rule features, human-machine operation behavior features, and human-machine business collaboration features.

[0083] Step S104: Input the target fusion features and target confidence features into the business status prediction model to obtain business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

[0084] Specifically, the target fusion features and target confidence features are simultaneously input into a pre-trained business status prediction model for prediction, resulting in business status predictions for various target business data. These predictions include business anomaly analysis results and business performance analysis results. The business status prediction model is trained by combining historical fusion features and historical confidence features from various historical business data.

[0085] This application embodiment constructs a business event relationship network to obtain target fusion features of multiple target business data, combines the target confidence features of target business data under multiple preset state data, and then uses a pre-trained business state prediction model to obtain business state prediction results of multiple target business data. This is beneficial for accurately analyzing the business state of human-machine collaboration in robotic process automation.

[0086] This application provides a method for predicting the business status of robotic process automation data, which can be used in electronic devices such as mobile phones, tablets, desktop computers, laptops, and servers. Figure 2 This is a flowchart of a business status prediction method for robotic process automation data according to an embodiment of the present invention, such as... Figure 2 As shown, the process includes the following steps: Step S101: Obtain various target business data, wherein the target business data is robotic process automation business data arranged in a time series.

[0087] Step S102: For each type of target business data, determine the target confidence characteristics of the target business data under multiple preset state data.

[0088] Step S103: Based on multiple target business data, construct a business event relationship network to obtain the target fusion features of the mutual integration of multiple target business data.

[0089] Step S104: Input the target fusion features and target confidence features into the business status prediction model to obtain business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

[0090] Steps S101-S104 have been described in the above embodiments and will not be repeated here.

[0091] Step S105: Based on the business status prediction results of various target business data, generate data processing efficiency suggestions.

[0092] In a specific example, data processing performance recommendations include: business log modification recommendations, business trajectory optimization recommendations, and human training optimization recommendations.

[0093] Step S106: Send the data processing performance recommendations to the target learning object.

[0094] Specifically, the target audience may include operators or maintenance personnel who are learning.

[0095] In some specific implementations, data processing performance recommendations include: business log modification recommendations, business trajectory optimization recommendations, and human-machine optimization recommendations. Based on the business status prediction results of various target business data, data processing performance recommendations are generated, including: Step c1: When an anomaly is predicted in the business log data, generate a business log modification suggestion based on the attribute information of historical business logs that have not shown anomalies in the historical time period.

[0096] In this embodiment of the application, when an abnormal situation is predicted in the business log data, the log modification suggestion is automatically generated based on the relevant operator identification, timestamp, and status change data of the historical business logs that have not shown any abnormality in the historical time period. This makes it easy to send the log modification suggestion to the maintenance personnel for learning, thereby improving the business processing efficiency of RPA.

[0097] Step c2: When abnormal conditions occur in the predicted business trajectory data, the business entity that caused the abnormality is identified by using the business event relationship network, and business trajectory optimization suggestions are generated by combining the records in the abnormality handling manual.

[0098] The business event relationship network is constructed through the above steps S105 and steps b1-b4. When the business trajectory data is abnormal, the business event relationship network can be traversed one by one to identify the abnormal business entities. Then, the abnormal business entities are identified. Combined with the content related to the abnormal business entities recorded in the anomaly handling manual, business trajectory optimization suggestions are obtained. This makes it easy to send the business trajectory optimization suggestions to the relevant business engineers for learning, thereby improving the business processing efficiency of RPA.

[0099] Step c3: When the predicted business interaction data shows a delay in human-computer interaction, obtain human-computer operation data where the human-computer collaboration capability index is lower than the preset index based on the business interaction data, and generate human-computer optimization suggestions by combining the recorded content of the human-computer operation manual.

[0100] For example, based on business interaction data, if the human-computer interaction time delay is determined to be 10 minutes, while the normal human-computer interaction time is 5 minutes, this indicates that the human-computer interaction collaboration time is too long. The human-computer operation manual is retrieved and its recorded content is extracted to generate a human-computer operation manual, which facilitates quick learning for operators and robots, thereby improving the optimization efficiency of RPA.

[0101] Therefore, this application embodiment generates data processing efficiency suggestions based on the business status prediction results of various target business data; the data processing efficiency suggestions are sent to the target object for learning, which ultimately helps to improve the efficiency of RPA human-machine collaborative processing of business.

[0102] This embodiment also provides a business status prediction device for robotic process automation data. This device is used to implement the above embodiments and preferred embodiments, and details already described will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0103] This embodiment provides a business status prediction device for robotic process automation data, such as... Figure 3 As shown, it includes: The business data acquisition module 301 is used to acquire various target business data, which are robotic process automation business data arranged in a time series. The first feature determination module 302 is used to determine the target confidence features of the target business data under multiple preset state data for each type of target business data; The second feature determination module 303 is used to construct a business event relationship network based on multiple target business data to obtain target fusion features that integrate multiple target business data. The business status prediction module 304 is used to input the target fusion features and target confidence features into the business status prediction model to obtain business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

[0104] The business state prediction device for robotic process automation data provided in this embodiment of the invention can execute the business state prediction device for robotic process automation data provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the execution method. Further functional descriptions of the above modules and units are the same as those in the corresponding embodiments described above, and will not be repeated here.

[0105] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0106] The following is a detailed reference. Figure 4 , Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention.

[0107] The following is a detailed reference. Figure 4 This diagram illustrates a suitable structural schematic for implementing an electronic device according to embodiments of the present invention. The electronic device may include a processor (e.g., a central processing unit, graphics processor, etc.) 401, which can perform various appropriate actions and processes based on a program stored in read-only memory (ROM) 402 or a program loaded from memory 408 into random access random access memory (RAM) 403. The RAM 403 also stores various programs and data required for the operation of the electronic device. The processor 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0108] Typically, the following devices can be connected to I / O interface 405: input devices 406 including, for example, touchscreens, touchpads, keyboards, mice, cameras, microphones, accelerometers, gyroscopes, etc.; output devices 407 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; memory devices 408 including, for example, magnetic tapes, hard disks, etc.; and communication devices 409. Communication device 409 allows electronic devices to communicate wirelessly or wiredly with other devices to exchange data. Although Figure 4Electronic devices with various devices are shown, but it should be understood that it is not required to implement or have all of the devices shown, and more or fewer devices may be implemented or have instead.

[0109] In particular, according to embodiments of the present invention, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present invention include a computer program product comprising a computer program carried on a non-transitory computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device 409, or installed from a memory 408, or installed from a ROM 402. When the computer program is executed by the processor 401, it performs the functions defined in the business state prediction method for robotic process automation data according to embodiments of the present invention.

[0110] Figure 4 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0111] This invention also provides a computer-readable storage medium. The methods described above according to embodiments of the invention can be implemented in hardware or firmware, or implemented as computer code that can be recorded on a storage medium, or implemented as computer code downloaded via a network and originally stored on a remote storage medium or a non-transitory machine-readable storage medium and then stored on a local storage medium. Thus, the methods described herein can be processed by software stored on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. The storage medium can be a magnetic disk, optical disk, read-only memory, random access memory, flash memory, hard disk, or solid-state drive, etc.; further, the storage medium can also include combinations of the above types of memory. It is understood that computers, processors, microprocessor controllers, or programmable hardware include storage components capable of storing or receiving software or computer code. When the software or computer code is accessed and executed by the computer, processor, or hardware, the business state prediction method for robotic process automation data shown in the above embodiments is implemented.

[0112] A portion of this invention can be applied as a computer program product, such as computer program instructions, which, when executed by a computer, can invoke or provide the methods and / or technical solutions according to the invention through the operation of the computer. Those skilled in the art will understand that the forms in which computer program instructions exist in a computer-readable medium include, but are not limited to, source files, executable files, installation package files, etc. Correspondingly, the ways in which computer program instructions are executed by a computer include, but are not limited to: the computer directly executing the instructions, or the computer compiling the instructions and then executing the corresponding compiled program, or the computer reading and executing the instructions, or the computer reading and installing the instructions and then executing the corresponding installed program. Here, the computer-readable medium can be any available computer-readable storage medium or communication medium accessible to a computer.

[0113] Although embodiments of the invention have been described in conjunction with the accompanying drawings, those skilled in the art can make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations all fall within the scope defined by the appended claims.

Claims

1. A method for predicting the business status of robotic process automation data, characterized in that, The method includes: Acquire multiple target business data, wherein the target business data is robotic process automation business data arranged in a time series; For each type of target business data, determine the target confidence characteristics of the target business data under multiple preset state data; Based on multiple target business data, a business event relationship network is constructed to obtain the target fusion features of the mutual integration of the multiple target business data; The target fusion features and the target confidence features are input into the business status prediction model for prediction, thereby obtaining business status prediction results for various target business data. The business status prediction model is a pre-trained neural network model.

2. The method according to claim 1, characterized in that, The various target business data include: business log data, business trajectory data, and business interaction data, wherein the business interaction data includes: human-computer intervention type data, human-computer interaction display data, human-computer interaction time data, and operator identification data; Acquire various target business data, including: Based on the operator's identification data, the human-computer intervention type data, human-computer interaction display data, human-computer interaction time data, business log data, and business trajectory data are associated to obtain a triplet of data consisting of the business log data, business trajectory data, and business interaction data; The triplet data is arranged according to the time series to obtain the various target business data.

3. The method according to claim 1, characterized in that, The various preset state data include: preset business state data, preset system state data, and preset abnormal state data. For each type of target business data, the target confidence feature of the target business data under the various preset business state data is determined, including: For each type of target service data, determine the confidence parameter of the target service data under the preset service state data, the preset system state data, and the preset abnormal state data, including: For each type of target business data, extract the corresponding business log features, business trajectory features, and business interaction features; For each type of target business data, the business log features, business trajectory features, and business interaction features corresponding to the target business data are input into the confidence coding model for recognition, so as to obtain the confidence parameters of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data; wherein, the confidence coding model is a pre-trained neural network model; For each type of target business data, based on the confidence parameters of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data, the target confidence feature of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data is determined.

4. The method according to claim 3, characterized in that, Determining the target confidence features of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data includes: First-level confidence level, second-level confidence level, third-level confidence level, and fourth-level confidence level are predefined; For each type of target business data, based on the first level of confidence, the second level of confidence, the third level of confidence, and the fourth level of confidence, the confidence level of the target business data under preset business state data, preset system state data, and preset abnormal state data is determined, thereby obtaining the target confidence feature of the target business data under preset business state data, preset system state data, and preset abnormal state data.

5. The method according to claim 3, characterized in that, Determining the target confidence features of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data includes: For each type of target business data, the confidence parameters of the target business data under the preset business state data, the preset system state data, and the preset abnormal state data are normalized to obtain the data state vector corresponding to the target business data; For each type of target business data, the data state vector corresponding to the target business data is used as the target confidence feature of the target business data under preset business state data, preset system state data, and preset abnormal state data.

6. The method according to claim 1, characterized in that, Based on multiple target business data, a business event relationship network is constructed to obtain the target fusion features of the mutual integration of the multiple target business data, including: Based on the various target business data, the target business characteristics of the robotic process automation business entities and the target relationship characteristics between business entities with business relationships are obtained. The target service characteristics are mapped onto the service network topology to determine the node corresponding to the service entity; The target relationship features are mapped onto the network topology to determine the edges corresponding to business entities with existing business relationships. Based on the nodes corresponding to the business entities and the edges corresponding to business entities with business relationships, a business event relationship network is constructed that integrates the various target business features. The target fusion feature is extracted from the business event relationship network by integrating multiple target business features.

7. The method according to claim 6, characterized in that, The target business features include: business name features, business identifier features, and business attribute features. The business type features include: human-machine processing efficiency features, business process rule features, human-machine operation behavior features, and human-machine business collaboration features. The target relationship features include: business relationship name, business relationship occurrence time, business relationship attribute, and business relationship direction. The business relationship name includes: influence, affected, bound, bound, triggered, and triggered. The business relationship occurrence time is the time when the business relationship is generated, and the business relationship direction is the direction of the two business features based on the business relationship.

8. The method according to claim 2, characterized in that, Also includes: Based on the business status prediction results of various target business data, generate data processing efficiency suggestions; The data processing performance recommendations are sent to the target audience for learning.

9. The method according to claim 8, characterized in that, Data processing performance recommendations include: business log modification suggestions, business trajectory optimization suggestions, and human-machine optimization suggestions. Based on the business status prediction results of various target business data, data processing performance recommendations are generated, including: When an anomaly is predicted in the business log data, a business log modification suggestion is generated based on the attribute information of historical business logs that have not shown anomalies in the historical time period. When an anomaly is predicted in the business trajectory data, the business entity causing the anomaly is identified using the business event relationship network, and business trajectory optimization suggestions are generated in conjunction with the records in the anomaly handling manual. When a human-computer interaction delay is predicted in the business interaction data, human-computer operation data with a human-computer collaboration capability index lower than a preset index is obtained based on the business interaction data, and human-computer optimization suggestions are generated in combination with the recorded content in the human-computer operation manual.

10. A business status prediction device for robotic process automation data, characterized in that, The device includes: The business data acquisition module is used to acquire various target business data, wherein the target business data is robotic process automation business data arranged in a time series. The first feature determination module is used to determine the target confidence features of the target business data under multiple preset state data for each type of target business data. The second feature determination module is used to construct a business event relationship network based on multiple target business data to obtain the target fusion feature of the multiple target business data. The business status prediction module is used to input the target fusion features and the target confidence features into the business status prediction model to obtain business status prediction results for various target business data, wherein the business status prediction model is a pre-trained neural network model.

11. An electronic device, characterized in that, include: A memory and a processor are communicatively connected, the memory stores computer instructions, and the processor executes the computer instructions to perform the business state prediction method for robotic process automation data as described in any one of claims 1 to 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions for causing the computer to execute the business state prediction method for robotic process automation data as described in any one of claims 1 to 9.

Images