Business processing method and system based on process automation robot

Abstract

The invention discloses a business processing method and system based on a process automation robot, and relates to the technical field of business processing, and the method comprises the steps: selecting path-finding factor information 1 and 2 according to target business information, detecting to obtain back-test factor information 1 according to the path-finding factor information 1, and obtaining back-test factor information 2 according to the back-test factor information 1; whether risk nodes exist in the first target sub-step nodes except the prediction risk nodes is judged, second back-detection factor information is obtained through detection according to second path-finding factor information, whether a first to-be-detected node in second auxiliary detection sub-step nodes is a risk node is judged, and the first to-be-detected node in the second auxiliary detection sub-step nodes is a risk node; and if a target weight node in the first target sub-step node is located at the tail ends of all node positions, judging whether the target weight node is a risk node or not according to the back test factor information 2 and a second auxiliary test node in the second auxiliary test sub-step node. According to the business processing method and system based on the process automation robot, the accuracy of risk judgment of the business processing steps can be improved.

Description

Technical Field

[0001] The present application relates to the field of business processing technology, and in particular to a business processing method and system based on a process automation robot. Background Art

[0002] Robotic process automation is a technology that simulates human operation of computers. It can replace humans to complete repetitive process work and realize the automation of business processing.

[0003] In the prior art, there is often a risk of step errors during business processing, and when one of the steps is wrong, the timeliness of the early warning notification is not high, and there is a high probability of misjudgment, which greatly reduces the business processing efficiency of the robot. Summary of the invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present application provides a business processing method and system based on a process automation robot.

[0005] In a first aspect, the present application provides a business processing method based on a process automation robot, the method comprising:

[0006] Obtaining a first target processing step and a second auxiliary test processing step of a business process and target business information to be processed, obtaining a first target sub-step node and a second auxiliary test sub-step node by statistics according to the first target processing step and the second auxiliary test processing step, and predicting a risk node in the first target sub-step node according to the target business information to obtain a risk node prediction result 1;

[0007] According to the target business information, pathfinding factor information one and pathfinding factor information two are selected, and according to the backtest factor information one obtained by detecting the first target sub-step node by the pathfinding factor information one, it is determined whether there are any risk nodes in the first target sub-step node in addition to the risk node prediction result one, and the total risk prediction result of the first type of business processing or the total risk prediction result one of the second type of business processing is output;

[0008] The backtest factor information 2 is obtained by detecting the first target sub-step node and the second auxiliary test sub-step node according to the pathfinding factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node are extracted;

[0009] According to the backtest factor information 2 and the second auxiliary test node, determine whether the first node to be tested is a risk node, and output the first type of business processing risk prediction total result or the second type of business processing risk prediction total result;

[0010] If the target weight node in the first target sub-step node is at the end of all node positions, determine whether the target weight node is a risk node based on the backtest factor information two and the second auxiliary test node, and output the total result of the first category business processing risk prediction or the total result three of the second category business processing risk prediction.

[0011] Preferably, a first target processing step and a second auxiliary test processing step of the business processing are obtained, and the sub-step nodes of each step in the first target processing step are counted to obtain the first target sub-step node, and the sub-step nodes of each step in the second auxiliary test processing step are counted to obtain the second auxiliary test sub-step node;

[0012] Obtain target business information that needs to be processed, and based on the target business information, obtain historical business information of the same category as the target business information, as well as historical risk feature record information of the historical business information at the first target sub-step node and the second auxiliary test sub-step node in the historical period, and train a node risk prediction model based on the historical business information and the historical risk feature record information;

[0013] The target business information, the first target sub-step node and the second auxiliary test sub-step node are input into the node risk prediction model for testing to obtain a risk node prediction result one.

[0014] Preferably, according to the target business information, pathfinding factor information 1 and pathfinding factor information 2 are selected for detecting whether a risk occurs in a sub-step node, the pathfinding factor information 1 and the pathfinding factor information 2 are of the same category but contain different information features, and the pathfinding factor information 2 is released after the pathfinding factor information 1 completes feedback;

[0015] A monitoring time threshold is preset, and backtest factor information one fed back after the pathfinding factor information one has passed the monitoring time threshold is counted, and the backtest factor information one is used to detect the first target sub-step node;

[0016] Preset standard reference factor information, determine whether the backtest factor information 1 is the same as the standard reference factor information, if so, determine that the risk node prediction result 1 is a misjudgment, and output the total result of the first category of business processing risk prediction, if not, determine the specific risk node in the first target sub-step node according to the backtest factor information 1 to obtain the risk node prediction result 2;

[0017] The risk node prediction result 1 and the risk node prediction result 2 are integrated to obtain the second category business processing risk prediction total result 1.

[0018] Preferably, if it is impossible to determine whether the backtest factor information 1 is the same as the standard reference factor information, the pathfinder factor information 2 is released, and the backtest factor information 2 fed back by the pathfinder factor information 1 after the monitoring time threshold is counted, and the backtest factor information 2 is used to detect the first target sub-step node and the second auxiliary test sub-step node;

[0019] Performing a weight evaluation on each node in the first target sub-step node to obtain a first type of node weight value, extracting a maximum weight value from the first type of node weight values, and extracting a target weight node from the first target sub-step node according to the maximum weight value;

[0020] If the target weight node is located at the beginning of all node positions in the first target sub-step node, then extract the first node to be tested located at the end of all nodes in the first target sub-step node;

[0021] According to the backtest factor information 2, when it can be directly determined that the target weight node is a risk-free node and it is not possible to directly determine whether the first node to be tested is a risky node, the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node is extracted.

[0022] Preferably, according to the backtest factor information 2, the difference between the backtest factor information of the first node to be tested and the standard reference factor information is counted to obtain a first difference to be tested, the difference between the backtest factor information of the second auxiliary test node and the standard reference factor information is counted to obtain a second auxiliary test difference, and the first difference to be tested and the second auxiliary test difference are summed to obtain a preprocessed backtest difference;

[0023] A backtest information difference interval value is preset. If the pre-processed backtest difference value falls within the backtest information difference interval value, the first risk node prediction result is judged to be a misjudgment, and the total result of the first category of business processing risk prediction is output;

[0024] If the pre-processing backtest difference does not fall within the backtest information difference interval, the first node to be tested is determined to be a risk node to obtain the risk node prediction result three;

[0025] The risk node prediction result one and the risk node prediction result three are integrated to obtain the second category business processing risk prediction total result two.

[0026] Preferably, if the target weight node is located at the end of all node positions in the first target sub-step node, and it is impossible to directly determine whether the target weight node is a risk node, then the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node is extracted;

[0027] According to the backtest factor information 2, it is determined whether the backtest factor information of the second auxiliary test node is the same as the standard reference factor information. If so, the risk node prediction result 1 is determined to be a misjudgment, and the total result of the risk prediction of the first type of business processing is output. If not, the target weight node is determined to be a risk node to obtain the risk node prediction result 4;

[0028] The risk node prediction result one and the risk node prediction result four are integrated to obtain the second category business processing risk prediction total result three.

[0029] In a second aspect, a business processing system based on a process automation robot includes:

[0030] The risk node comprehensive prediction unit is used to obtain the first target processing step and the second auxiliary test processing step of the business processing and the target business information that needs to be processed, and obtain the first target sub-step node and the second auxiliary test sub-step node according to the first target processing step and the second auxiliary test processing step, and predict the risk node in the first target sub-step node according to the target business information to obtain the risk node prediction result 1;

[0031] A risk node verification unit 1 is used to select pathfinding factor information 1 and pathfinding factor information 2 according to target business information, and to determine whether there are risk nodes in the first target sub-step node in addition to the risk node prediction result 1 according to the backtest factor information 1 detected by the pathfinding factor information 1 for the first target sub-step node, and output the total risk prediction result of the first type of business processing or the total risk prediction result 1 of the second type of business processing;

[0032] The main test node screening unit is used to obtain the backtest factor information 2 by detecting the first target sub-step node and the second auxiliary test sub-step node according to the pathfinding factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, then extract the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node;

[0033] The second risk node verification unit is used to determine whether the first node to be tested is a risk node according to the second backtest factor information and the second auxiliary test node, and output the first type of business processing risk prediction total result or the second type of business processing risk prediction total result;

[0034] The risk node verification unit three is used to determine whether the target weight node in the first target sub-step node is a risk node based on the backtest factor information two and the second auxiliary test node if the target weight node in the first target sub-step node is at the end of all node positions, and output the total result of the risk prediction of the first category business processing or the total result three of the risk prediction of the second category business processing.

[0035] Compared with the prior art, the present invention has the following characteristics and beneficial effects:

[0036] Based on the historical information, the prediction model is used to predict the specific risk node in the first target sub-step node. This analysis method is relatively traditional, so there are large problems of misjudgment or missed judgment. In order to improve the accuracy of risk node monitoring, two types of pathfinder factor information are selected, namely pathfinder factor information 1 and pathfinder factor information 2. Pathfinder factor information 1 is released first, and only the first target sub-step node is detected and processed. When the specific wind power node in the first target sub-step node can be directly determined by pathfinder factor information 1, there is no need to release pathfinder factor information 2. Otherwise, pathfinder factor information 2 is released in time for further verification detection and processing, and pathfinder factor information 2 is Both the first target sub-step node and the second auxiliary sub-step node are detected, and the differentiated analysis of the target weight node position in the first target sub-step node is involved to expand the diversity and comprehensiveness of the analysis situation. For different positions of the target weight node, the second auxiliary sub-step node is utilized. By using the information analysis of the next processing step, the information unification formed by the information analysis under the same conditions is avoided, and the risk status of the specific node cannot be accurately judged. Through the above-mentioned differentiated processing and verification methods for multiple situations, the rigor of the risk monitoring work of the entire business processing step is greatly enhanced, and the correlation between multiple information is analyzed, thereby reducing the probability of misjudgment of the risk monitoring results. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flowchart of a business processing method based on a process automation robot mainly embodied in this embodiment.

[0038] Figure 2 This is a structural block diagram of a business process based on a process automation robot mainly embodied in this embodiment. DETAILED DESCRIPTION

[0039] The present invention is further described in detail below in conjunction with the following examples.

[0040] Reference Figure 1 , a business processing method based on a process automation robot, the method comprising the following steps:

[0041] S1. Obtain the first target processing step and the second auxiliary test processing step of the business processing and the target business information that needs to be processed. According to the first target processing step and the second auxiliary test processing step, obtain the first target sub-step node and the second auxiliary test sub-step node by statistics. According to the target business information, predict the risk node in the first target sub-step node to obtain the risk node prediction result one.

[0042] S2. According to the target business information, select pathfinder factor information 1 and pathfinder factor information 2, and use the pathfinder factor information 1 to detect the first target sub-step node to obtain the backtest factor information 1, determine whether there are risk nodes in the first target sub-step node except the risk node prediction result 1, and output the total risk prediction result of the first category business processing or the total risk prediction result 1 of the second category business processing.

[0043] S3. Based on the pathfinding factor information 2, the first target sub-step node and the second auxiliary test sub-step node are detected to obtain the backtest factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node are extracted.

[0044] S4. According to the backtest factor information 2 and the second auxiliary test node, determine whether the first node to be tested is a risk node, and output the total result of the risk prediction of the first type of business processing or the total result 2 of the risk prediction of the second type of business processing.

[0045] S5. If the target weight node in the first target sub-step node is at the end of all node positions, determine whether the target weight node is a risk node based on the backtest factor information two and the second auxiliary test node, and output the total result of the first category business processing risk prediction or the total result three of the second category business processing risk prediction.

[0046] Specifically, based on the historical information, the prediction model is used to predict the specific risk node in the first target sub-step node. This analysis method is relatively traditional, so there are large problems of misjudgment or missed judgment. In order to improve the accuracy of risk node monitoring, two types of pathfinder factor information are selected, namely pathfinder factor information 1 and pathfinder factor information 2. Pathfinder factor information 1 is released first, and only the first target sub-step node is detected and processed. When the specific wind power node in the first target sub-step node can be directly determined by pathfinder factor information 1, there is no need to release pathfinder factor information 2. Otherwise, pathfinder factor information 2 is released in time for further verification detection and processing, and pathfinder factor information Secondly, both the first target sub-step node and the second auxiliary sub-step node are detected, and the differentiated analysis of the target weight node position in the first target sub-step node is involved, so as to expand the diversity and comprehensiveness of the analysis situation. For different positions of the target weight node, the second auxiliary sub-step node is used. By using the information analysis of the next processing step, the information unification formed by the information analysis under the same conditions is avoided, and the risk status of the specific node cannot be accurately judged. Through the above-mentioned differentiated processing and verification methods for multiple situations, the rigor of the risk monitoring work of the entire business processing step is greatly enhanced, and the correlation between multiple information is analyzed, which reduces the probability of misjudgment of the risk monitoring results.

[0047] The specific step S1 includes the following sub-steps:

[0048] Obtain the first target processing step and the second auxiliary test processing step of the business processing, count the sub-step nodes of each step in the first target processing step to obtain the first target sub-step node, and count the sub-step nodes of each step in the second auxiliary test processing step to obtain the second auxiliary test sub-step node.

[0049] Obtain target business information that requires business processing, and based on the target business information, obtain historical business information of the same category as the target business information, as well as historical risk feature record information of the historical business information at the first target sub-step node and the second auxiliary sub-step node in the historical period; and train a node risk prediction model based on the historical business information and the historical risk feature record information.

[0050] The target business information, the first target sub-step node and the second auxiliary test sub-step node are input into the node risk prediction model for testing to obtain the risk node prediction result one.

[0051] Specifically, such as business processing (including the following stages: demand analysis and goal setting, process selection, feasibility study, project planning, technology selection, prototype development, testing, employee training, deployment, monitoring and optimization, expansion and iteration, document and knowledge management, change management, risk management, stakeholder communication, ROI evaluation, continuous support and maintenance, innovation and strategic planning, such as using some of the stage steps for subsequent examples), the first target processing step and the second auxiliary testing processing step (if they are document and knowledge management and change management, they are marked with letters here, if they are steps A and B respectively), the sub-step nodes of each step (such as data information cleaning processing, information reduction processing, information conversion processing and other information processing nodes such as A includes sub-nodes a1, a2, a3), the first target sub-step node (such as A includes sub-nodes a1, a2, a3), the second auxiliary testing sub-step node (such as B includes sub-nodes b1, b2), target business information (such as a customer's confidential information needs to be encrypted), historical business information (i.e. The category is also the confidential information of the management customer, but the type of confidential information is different, such as h1, h2, h3, h4), historical risk feature record information (such as the risk node with risk in A of the business processing step corresponding to h1 is a1, and no risk occurs in B, and so on. The explanations of h2, h3, and h4 are the same and will not be explained here), node risk prediction model (that is, h1, h2, h3, h4, and the characteristic information of the risk nodes with risk in A or the risk nodes with risk in B corresponding to each historical business information are listed in a comparison table for matching subsequent information, which can realize the judgment of risk nodes, or, it can be based on the above-mentioned statistical historical feature information, such as the neural network and ensemble learning model in the prior art), risk node prediction result one (if the current target business information is h0, then according to h0, the matching table formed in the node risk prediction model or the neural network and ensemble learning are used to predict the risk nodes with risk in a1, a2, a3 or b1, b2. If it is a2, it is the risk node prediction result one).

[0052] The specific step S2 includes the following sub-steps:

[0053] According to the target business information, pathfinder factor information 1 and pathfinder factor information 2 are selected to detect whether there is a risk in the sub-step node. Pathfinder factor information 1 and pathfinder factor information 2 are of the same category but contain different information characteristics, and pathfinder factor information 2 is released after pathfinder factor information 1 completes feedback.

[0054] A monitoring time threshold is preset, and the pathfinding factor information 1 is statistically fed back after the monitoring time threshold, and the backtest factor information 1 is used to detect the first target sub-step node.

[0055] Preset standard reference factor information, determine whether the backtest factor information 1 is the same as the standard reference factor information, if so, determine that the risk node prediction result 1 is a misjudgment, and output the total result of the first category of business processing risk prediction, if not, determine the specific risk node in the first target sub-step node based on the backtest factor information 1 to obtain the risk node prediction result 2.

[0056] The risk node prediction result one and the risk node prediction result two are integrated to obtain the second category business processing risk prediction total result one.

[0057] Specifically, such as pathfinder factor information 1 and pathfinder factor information 2 (if the categories are the same as those of h0, pathfinder factor information 1 is T1 and pathfinder factor information 2 is T2, wherein T1 is a very small sub-information block in the h0 information, and similarly, T2 is a very small sub-information block in the h0 information, and the information block contents of T1 and T2 are different, so that when the result cannot be determined under the detection condition of T1, T2 enhances the information detection effect on the basis of T1), preset monitoring time threshold (if it is t), backtest factor information 1 (that is, the return information of T1 after t time, the sub-information block after being processed by a1, a2, and a3, if it is T11), preset standard reference factor information (if it is T0, that is, the standard information obtained after processing under the risk-free conditions of a1, a2, and a3), judgment Whether the backtest factor information one is the same as the standard reference factor information (if T11 is the same as T0, it means that the sub-nodes a1, a2, and a3 are risk-free, so the risk node a2 in the risk node prediction result one is a misjudgment result), risk node prediction result two (if T11 is different from T0, the information processed by T1 in a1, a2, and a3 are all included in T11. If the information processed by T1 in a2 and a3 is standard information, and the information processed by T1 in a1 is non-standard information, the risk node judgment here is the same as the comparison between T11 and T0 mentioned above, and no further explanation is given here. If T0 contains sub-standard reference factor information, a1 is determined to be a risk node, that is, risk node prediction result two), the second type of business processing risk prediction total result one (that is, the risk nodes are a1 and a2).

[0058] The specific step S3 includes the following sub-steps:

[0059] If it is impossible to determine whether the backtest factor information 1 is the same as the standard reference factor information, the pathfinder factor information 2 is released, and the backtest factor information 2 fed back by the pathfinder factor information 1 after the monitoring time threshold is counted. The backtest factor information 2 is used to detect the first target sub-step node and the second auxiliary test sub-step node.

[0060] A weight evaluation is performed on each node in the first target sub-step node to obtain a first-category node weight value, a maximum weight value is extracted from the first-category node weight value, and a target weight node is extracted from the first target sub-step node according to the maximum weight value.

[0061] If the target weight node is located at the beginning of all node positions in the first target sub-step node, the first node to be tested located at the end of all nodes in the first target sub-step node is extracted.

[0062] According to the backtest factor information 2, when it can be directly determined that the target weight node is a risk-free node, and it is impossible to directly determine whether the first node to be tested is a risky node, the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node is extracted.

[0063] Specifically, such as backtest factor information 2 (if the detected backtest factor information 1 is uncertain, if the backtest factor information of a3 is estimated, it is impossible to directly and accurately determine whether a3 is a risk node or a normal node, and further verification and analysis are required. At this time, it is necessary to monitor and verify through T2, that is, the return information of T2 after t time, the sub-information block after processing by a1, a2, a3, b1, and b2), the weight value of the first type of node (that is, the efficiency of each sub-node in processing information, if it is 60%, 20%, and 20% respectively, it should be noted that the node with a large weight value has a higher fault tolerance), the maximum weight value (60%), the target weight node (if it is a1), the first node to be tested Point (i.e. a3), when it can be directly judged that the target weight node is a risk-free node, and it is impossible to directly judge whether the first node to be tested is a risky node (the same as the processing steps of the above-mentioned backtest factor information one, no further explanation is given here, that is, a1 is judged to be a normal node, and when the judgment result of a3 is the same as the above: if it is impossible to judge whether the backtest factor information one is the same as the standard reference factor information, it is necessary to use the sub-node processing status in step B for auxiliary analysis. Because B is a connecting step after A, the risks between the two are highly correlated), the second auxiliary test node (such as b1, that is, the starting processing node in B. It should be noted that the fault tolerance of b1 here is relatively low).

[0064] The specific step S4 includes the following sub-steps:

[0065] According to the backtest factor information 2, the difference between the backtest factor information of the first node to be tested and the standard reference factor information is counted to obtain the first difference to be tested, the difference between the backtest factor information of the second auxiliary test node and the standard reference factor information is counted to obtain the second auxiliary test difference, and the first difference to be tested and the second auxiliary test difference are summed to obtain the preprocessed backtest difference.

[0066] A backtest information difference interval value is preset. If the pre-processed backtest difference falls within the backtest information difference interval value, the risk node prediction result 1 is judged to be a misjudgment, and the total result of the first category of business processing risk prediction is output.

[0067] If the pre-processed backtest difference does not fall within the backtest information difference interval value, the first node to be tested is judged to be a risk node to obtain the risk node prediction result three.

[0068] The risk node prediction result one and the risk node prediction result three are integrated to obtain the second category business processing risk prediction total result two.

[0069] Specifically, such as the first difference to be tested (if the backtest factor information of the first node to be tested is c1, the first difference to be tested is the absolute value of T0-c1, if it is k1), the second auxiliary test difference (if the backtest factor information of the second auxiliary test node is c2, the second auxiliary test difference is the absolute value of T0-c1, if it is k2), the pre-processed backtest difference (is (k1+k2) / 2, if it is P1), the preset backtest information difference interval value (if it is P0-P2), the total result of the risk prediction of the first category of business processing (P1 is within P0-P2, then a2 is judged to be a misjudgment, a2 is a normal node, that is, A operates normally), the risk node prediction result three (if the pre-processed backtest difference is P3, then a3 is a risk node), the second category of business processing risk prediction total result two (that is, the risk nodes are a2 and a3).

[0070] The specific step S5 includes the following sub-steps:

[0071] If the target weight node is located at the end of all node positions in the first target sub-step node, and it is impossible to directly determine whether the target weight node is a risk node, then the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node is extracted.

[0072] Based on the backtest factor information 2, determine whether the backtest factor information of the second auxiliary test node is the same as the standard reference factor information. If so, the risk node prediction result 1 is judged to be a misjudgment, and the total result of the first category of business processing risk prediction is output. If not, the target weight node is judged to be a risk node to obtain the risk node prediction result 4.

[0073] The risk node prediction result one and the risk node prediction result four are integrated to obtain the second category business processing risk prediction total result three.

[0074] Specifically, if the target weight node is located at the end of all node positions in the first target sub-step node (that is, a1 is located at the end positions of a2 and a3, that is, connected to b1), the second auxiliary test node (that is, b1, it should be noted that the fault tolerance of b1 here is relatively low), the backtest factor information of the second auxiliary test node (if it is T20), the total result of the risk prediction of the first type of business processing (if the information of T20 is the same as that of T2, it means that b1 is a normal node and there is no risk), the risk node prediction result four (if the information of T20 is different from that of T2, it can be directly judged that b1 is a risk node, because b1 does not have a high fault tolerance, which means that the previous node processing of b1 is abnormal, that is, a3 is a risk node), the total result of the risk prediction of the second type of business processing is three (that is, the risk nodes are a2 and a3).

[0075] A business processing system based on a process automation robot, by applying a business processing method based on a process automation robot as described above, includes a risk node comprehensive prediction unit, a risk node verification unit 1, a main test node screening unit, a risk node verification unit 2 and a risk node verification unit 3, with reference to Figure 2 , obtain the first target processing step and the second auxiliary test processing step of the business processing and the target business information that needs to be processed through the risk node comprehensive prediction unit, and obtain the first target sub-step node and the second auxiliary test sub-step node according to the first target processing step and the second auxiliary test processing step, and predict the risk node in the first target sub-step node according to the target business information to obtain the risk node prediction result one; select the pathfinder factor information one and the pathfinder factor information two according to the target business information through the risk node verification unit one, and obtain the backtest factor information one by detecting the first target sub-step node according to the pathfinder factor information one, and judge whether there are risk nodes in the first target sub-step node except the risk node prediction result one, and output the total risk prediction result of the first type of business processing or the total risk prediction result of the second type of business processing; output the total risk prediction result of the first type of business processing or the total risk prediction result of the second type of business processing through the main test node screening unit according to the pathfinder factor information The backtest factor information 2 is obtained by detecting the first target sub-step node and the second auxiliary test sub-step node, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node are extracted; through the risk node verification unit 2, according to the backtest factor information 2 and the second auxiliary test node, whether the first node to be tested is a risk node is judged, and the total result of the risk prediction of the first category of business processing or the total result of the risk prediction of the second category of business processing is output; through the risk node verification unit 3, if the target weight node in the first target sub-step node is located at the end of all node positions, according to the backtest factor information 2 and the second auxiliary test node, whether the target weight node is a risk node is judged, and the total result of the risk prediction of the first category of business processing or the total result of the risk prediction of the second category of business processing is output.

[0076] The above are all preferred embodiments of the present application, and the protection scope of the present application is not limited thereto. Therefore, any equivalent changes made according to the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A business processing method based on a process automation robot, characterized in that: The following steps are involved: Obtaining a first target processing step and a second auxiliary test processing step of a business process and target business information to be processed, obtaining a first target sub-step node and a second auxiliary test sub-step node by statistics according to the first target processing step and the second auxiliary test processing step, and predicting a risk node in the first target sub-step node according to the target business information to obtain a risk node prediction result 1; According to the target business information, pathfinding factor information one and pathfinding factor information two are selected, and according to the backtest factor information one obtained by detecting the first target sub-step node by the pathfinding factor information one, it is determined whether there are any risk nodes in the first target sub-step node in addition to the risk node prediction result one, and the total risk prediction result of the first type of business processing or the total risk prediction result one of the second type of business processing is output; The backtest factor information 2 is obtained by detecting the first target sub-step node and the second auxiliary test sub-step node according to the pathfinding factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node are extracted; According to the backtest factor information 2 and the second auxiliary test node, determine whether the first node to be tested is a risk node, and output the first type of business processing risk prediction total result or the second type of business processing risk prediction total result; If the target weight node in the first target sub-step node is at the end of all node positions, determine whether the target weight node is a risk node based on the backtest factor information two and the second auxiliary test node, and output the total result of the first category business processing risk prediction or the total result three of the second category business processing risk prediction.

2. A business processing method based on a process automation robot according to claim 1, characterized in that: The steps of obtaining the first target processing step and the second auxiliary test processing step of the business processing and the target business information to be processed, obtaining the first target sub-step node and the second auxiliary test sub-step node by statistics according to the first target processing step and the second auxiliary test processing step, and predicting the risk node in the first target sub-step node according to the target business information to obtain the risk node prediction result 1 are specifically as follows: Obtain a first target processing step and a second auxiliary test processing step of a business process, count the sub-step nodes of each step in the first target processing step to obtain a first target sub-step node, and count the sub-step nodes of each step in the second auxiliary test processing step to obtain a second auxiliary test sub-step node; Obtain target business information that needs to be processed, and based on the target business information, obtain historical business information of the same category as the target business information, as well as historical risk feature record information of the historical business information at the first target sub-step node and the second auxiliary test sub-step node during the historical period, and train a node risk prediction model based on the historical business information and the historical risk feature record information; The target business information, the first target sub-step node and the second auxiliary test sub-step node are input into the node risk prediction model for testing to obtain a risk node prediction result one.

3. A business processing method based on a process automation robot according to claim 2, characterized in that: According to the target business information, a pathfinding factor information 1 and a pathfinding factor information 2 are selected, and according to the backtest factor information 1 obtained by detecting the first target sub-step node by the pathfinding factor information 1, it is determined whether there are risk nodes in the first target sub-step node in addition to the risk node prediction result 1, and the steps of outputting the total risk prediction result of the first type of business processing or the total risk prediction result 1 of the second type of business processing are specifically as follows: According to the target business information, pathfinding factor information 1 and pathfinding factor information 2 are selected for detecting whether a risk occurs in a sub-step node. Pathfinding factor information 1 and pathfinding factor information 2 are of the same category but contain different information features, and pathfinding factor information 2 is released after pathfinding factor information 1 completes feedback; A monitoring time threshold is preset, and backtest factor information one fed back after the pathfinding factor information one has passed the monitoring time threshold is counted, and the backtest factor information one is used to detect the first target sub-step node; Preset standard reference factor information, determine whether the backtest factor information 1 is the same as the standard reference factor information, if so, determine that the risk node prediction result 1 is a misjudgment, and output the total result of the first category of business processing risk prediction, if not, determine the specific risk node in the first target sub-step node according to the backtest factor information 1 to obtain the risk node prediction result 2; The risk node prediction result 1 and the risk node prediction result 2 are integrated to obtain the second category business processing risk prediction total result 1.

4. A business processing method based on a process automation robot according to claim 3, characterized in that: The backtest factor information 2 is obtained by detecting the first target sub-step node and the second auxiliary test sub-step node according to the pathfinding factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, the step of extracting the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node is specifically as follows: If it is impossible to determine whether the backtest factor information 1 is the same as the standard reference factor information, the pathfinder factor information 2 is released, and the backtest factor information 2 fed back by the pathfinder factor information 1 after the monitoring time threshold is counted, and the backtest factor information 2 is used to detect the first target sub-step node and the second auxiliary test sub-step node; Performing a weight evaluation on each node in the first target sub-step node to obtain a first type of node weight value, extracting a maximum weight value from the first type of node weight values, and extracting a target weight node from the first target sub-step node according to the maximum weight value; If the target weight node is located at the beginning of all node positions in the first target sub-step node, then extract the first node to be tested located at the end of all nodes in the first target sub-step node; According to the backtest factor information 2, when it can be directly determined that the target weight node is a risk-free node and it is not possible to directly determine whether the first node to be tested is a risky node, the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node is extracted.

5. A business processing method based on a process automation robot according to claim 4, characterized in that: The step of determining whether the first node to be tested is a risk node according to the backtest factor information 2 and the second auxiliary test node, and outputting the total result of risk prediction of the first type of business processing or the total result of risk prediction of the second type of business processing is specifically: According to the backtest factor information 2, the difference between the backtest factor information of the first node to be tested and the standard reference factor information is counted to obtain a first difference to be tested, the difference between the backtest factor information of the second auxiliary test node and the standard reference factor information is counted to obtain a second auxiliary test difference, and the first difference to be tested and the second auxiliary test difference are summed to obtain a preprocessed backtest difference; A backtest information difference interval value is preset. If the pre-processed backtest difference value falls within the backtest information difference interval value, the first risk node prediction result is judged to be a misjudgment, and the total result of the first category of business processing risk prediction is output; If the pre-processing backtest difference does not fall within the backtest information difference interval, the first node to be tested is determined to be a risk node to obtain the risk node prediction result three; The risk node prediction result one and the risk node prediction result three are integrated to obtain the second category business processing risk prediction total result two.

6. A business processing method based on a process automation robot according to claim 5, characterized in that: If the target weight node in the first target sub-step node is located at the end of all node positions, the steps of judging whether the target weight node is a risk node according to the backtest factor information 2 and the second auxiliary test node, and outputting the total result of the risk prediction of the first type of business processing or the total result of the risk prediction of the second type of business processing are specifically as follows: If the target weight node is located at the end of all node positions in the first target sub-step node, and it is impossible to directly determine whether the target weight node is a risk node, then extract the second auxiliary test node located at the beginning of all nodes in the second auxiliary test sub-step node; According to the backtest factor information 2, it is determined whether the backtest factor information of the second auxiliary test node is the same as the standard reference factor information. If so, the risk node prediction result 1 is determined to be a misjudgment, and the total result of the risk prediction of the first type of business processing is output. If not, the target weight node is determined to be a risk node to obtain the risk node prediction result 4; The risk node prediction result one and the risk node prediction result four are integrated to obtain the second category business processing risk prediction total result three.

7. A business processing system based on a process automation robot, characterized in that: The system is used to implement a business processing method based on a process automation robot as described in any one of claims 1 to 6, comprising: The risk node comprehensive prediction unit is used to obtain the first target processing step and the second auxiliary test processing step of the business processing and the target business information that needs to be processed, and obtain the first target sub-step node and the second auxiliary test sub-step node according to the first target processing step and the second auxiliary test processing step, and predict the risk node in the first target sub-step node according to the target business information to obtain the risk node prediction result 1; A risk node verification unit 1 is used to select pathfinding factor information 1 and pathfinding factor information 2 according to target business information, and to determine whether there are risk nodes in the first target sub-step node in addition to the risk node prediction result 1 according to the backtest factor information 1 detected by the pathfinding factor information 1 for the first target sub-step node, and output the total risk prediction result of the first type of business processing or the total risk prediction result 1 of the second type of business processing; The main test node screening unit is used to obtain the backtest factor information 2 by detecting the first target sub-step node and the second auxiliary test sub-step node according to the pathfinding factor information 2, and if the target weight node in the first target sub-step node is located at the beginning of all node positions, then extract the first node to be tested in the first target sub-step node and the second auxiliary test node in the second auxiliary test sub-step node; The second risk node verification unit is used to determine whether the first node to be tested is a risk node according to the second backtest factor information and the second auxiliary test node, and output the first type of business processing risk prediction total result or the second type of business processing risk prediction total result; The risk node verification unit three is used to determine whether the target weight node in the first target sub-step node is a risk node based on the backtest factor information two and the second auxiliary test node if the target weight node in the first target sub-step node is at the end of all node positions, and output the total result of the risk prediction of the first category business processing or the total result three of the risk prediction of the second category business processing.

Images