RPA artificial intelligence control data processing method

Abstract

The invention relates to the technical field of enterprise data processing, in particular to an RPA artificial intelligence control data processing method, which comprises the following steps: acquiring an enterprise data processing task, preprocessing and standardizing multi-source enterprise data, classifying and storing by using a text classification model, and performing quality analysis on the stored data; constructing a linear regression model, analyzing the influence of different factors on business indexes, determining key factors and the influence degree thereof, and evaluating the overall interpretation force of the factors on the sales amount through decision coefficients, namely the adjusted decision coefficients; after an RPA task list is obtained, according to the score of the final priority, resources are distributed to the RPA robot according to the sequence of the priority; and the RPA robot tasks are polled regularly, and the abnormal condition of task operation is judged according to the index range of normal execution of the tasks. According to the method, tasks can be intelligently arranged according to task priorities, time constraints and business value contributions, and errors and delay caused by manual intervention are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of enterprise data processing, and in particular to an RPA artificial intelligence controlled data processing method. Background Art

[0002] In today's era of information explosion, enterprises are facing unprecedented data challenges. With the expansion of business scale and the intensification of market competition, the amount of data that enterprises need to process is growing exponentially. These data include not only traditional structured data, but also a large amount of unstructured data. Effective management and utilization of these data have become one of the key factors for enterprises to enhance their competitiveness and optimize their decision-making processes. In order to meet this challenge, more and more enterprises have begun to adopt technical solutions based on the combination of robotic process automation (RPA) and artificial intelligence (AI) to optimize their data processing processes. This method combines RPA's ability to efficiently execute regular tasks with AI's powerful data analysis and learning capabilities, providing enterprises with a new perspective on how to better manage and utilize the massive information resources generated internally and externally.

[0003] In existing enterprise data processing, there are many challenges in data processing, including low efficiency, high error rate, waste of resources, reduced competitiveness, reduced customer satisfaction, compliance risks, backward technology and data silos. These problems not only affect the operational efficiency of the enterprise, but may also have an adverse impact on the long-term development of the enterprise. Summary of the invention

[0004] In view of the technical problems existing in the prior art, the present invention provides an RPA artificial intelligence control data processing method.

[0005] The technical solution of the present invention to solve the above technical problems is as follows: an RPA artificial intelligence control data processing method, comprising the following steps:

[0006] S101. Obtain enterprise data processing tasks, pre-process and standardize multi-source enterprise data, classify and store using a text classification model, and perform quality analysis on the stored data;

[0007] S102. Construct a linear regression model to analyze the impact of different factors on business indicators, identify key factors and their impact, and evaluate the overall explanatory power of these factors on sales through the coefficient of determination, i.e., the adjusted coefficient of determination;

[0008] S103. After obtaining the RPA task list, allocate resources to the RPA robots in order of priority based on the final priority score, continuously monitor and adjust task execution and resource allocation in a timely manner;

[0009] S104. Periodically poll the RPA robot task, connect to its operating environment to obtain information about the task steps and the amount of data processed, and determine the abnormality of the task operation based on the indicator range of normal task execution.

[0010] In a preferred embodiment, the enterprise data processing task to be executed is obtained in S101, and the obtained enterprise data is preprocessed and standardized, data from different enterprise data sources are uniformly converted, and the preprocessed and standardized enterprise data are subjected to text classification to determine the business type to which the enterprise data belongs, and the text is input into the text classification model to classify the text. When the output layer has n neurons, corresponding to n different business types, the net input of the output layer is recorded as Z L , generate probability distribution for each text in each category, the specific calculation formula is as follows:

[0011]

[0012] Among them, Pi represents the probability of the i-th category, Z i,L Represents the net input of the i-th neuron in the output layer, which is used to calculate the probability Pi corresponding to the text belonging to the i-th category. When the net input vector of the output layer is:

[0013]

[0014] For the first category, the specific calculation formula of its probability value P1 is as follows:

[0015]

[0016] By calculating the probability value of the text in each category, we get a probability vector P = [P 1 ,P 2 ,...,P n ], indicating the probability distribution of each text belonging to different business types. Based on the probability distribution, the classification results of the text enterprise data are determined. According to the classification results, the enterprise data is classified and stored in the corresponding enterprise database table. The quality analysis of the stored enterprise data is performed, including calculating the accuracy, completeness and outlier detection of the enterprise data. When the total number of records in the enterprise data is n total When the accuracy verification rule is followed, each field of each record is checked one by one. For records that do not meet the accuracy rules, the number of records is n. error , calculate the accuracy of enterprise data, the specific formula is as follows:

[0017]

[0018] Accuracy indicates accuracy. We traverse the enterprise data and check whether the required fields in each record are empty. We record the number of records where the required fields are missing. missing , the specific calculation formula of missing rate is as follows:

[0019]

[0020] Outliers are detected by calculating the mean of all enterprise data in this field:

[0021]

[0022] Among them, x i represents the i-th enterprise data value of the field, n represents the number of enterprise data in the field, and calculates the standard deviation:

[0023]

[0024] Taking k as the judgment range of outliers, it will exceed Enterprise data outside the specified range is considered as outliers. For detected outliers, corresponding correction strategies are adopted according to the business meaning of the enterprise data.

[0025] In a preferred embodiment, in S102, based on the enterprise data after quality analysis, the enterprise data of factors affecting the enterprise business indicators and the corresponding business indicator enterprise data are obtained, a linear regression model is established, and the dependent variables and independent variables related to the business problem are determined. When the business indicator to be analyzed is sales volume y i When the sales volume is affected, the factors are marketing expenses x i1 、Product price x i2 and the number of customers x i3 , where i represents the i-th sample, and there are n samples, that is, the number of samples in the training set is n, and the number of independent variables B = 3, then the linear regression model can be expressed as:

[0026] y i =β 0 +β 1 x i1 +β 2 x i2 +β 3 x i3 +∈ i

[0027] Among them, β 0 is the intercept term, which represents the value when the marketing cost, product price and number of customers are all 0, β 1 is the regression coefficient corresponding to the marketing expenses, indicating the impact of marketing expenses on sales, β 2is the regression coefficient corresponding to product price, β 3 Represents the regression coefficient corresponding to the number of customers, ∈ i is the error term, which indicates the impact of factors other than marketing expenses, product prices, and customer numbers on sales. When the number of samples in the test set is m, the true sales value of the i-th sample in the test set is The model predicts the sales value of i test samples as Calculate the average difference between the predicted value and the true value. The specific calculation formula is as follows:

[0028]

[0029] A small MSE value indicates that the model has a good prediction effect. 2 and the adjusted R 2 The value of can evaluate the overall explanatory power of marketing expenses, product prices and customer numbers on sales. The coefficient of determination R 2 The specific calculation formula is as follows:

[0030]

[0031] in, represents the mean of the actual sales of the test set, R 2 The value range is between 0 and 1. A value close to 1 indicates that the model fits the enterprise data well. 2 The company can judge the overall explanatory power of the currently selected marketing expenses, product prices and customer numbers on sales by the adjusted R 2 The specific calculation formula is as follows:

[0032]

[0033] Adjusted R 2 When the number of independent variables increases, when the new independent variables do not significantly improve the model fitting effect, the adjusted R 2 The value will decrease.

[0034] In a preferred embodiment, the task list that needs to be processed by RPA is obtained in S103, including the task name, task description, business process and data processing volume information, and the priority of each task is determined. According to the priority of the task and the time requirement of data processing, the workflow of the RPA robot is reasonably arranged. During the peak period of business, limited resources are allocated to key business data processing tasks, and the priority of the task is calculated according to the business value contribution and time constraints. When the business value contribution has m evaluation dimensions, including the impact of revenue growth, customer satisfaction and cost reduction, the weight of the jth dimension is w vj, the score of task i on the jth dimension is s ij , then the business value contribution score of task i is V i The specific calculation formula is as follows:

[0035]

[0036] Determine the time constraint score based on the task deadline, the current time is t now , the deadline of task i is t deadline When , the time constraint score of task i is T i The specific calculation formula is as follows:

[0037]

[0038] The final priority is calculated based on business value contribution and time constraints. The weight of the business value contribution score is w v , the weight of the time constraint score is w t , then the specific calculation formula for the final priority of task i is as follows:

[0039] P i =w v ×V i +w t ×T i

[0040] Sort tasks according to the final priority score, allocate resources to high-priority tasks first, allocate sufficient CPU and memory resources to high-priority order processing tasks, assign tasks to RPA robots according to the established scheduling strategy, continuously monitor the execution progress of tasks and resource usage, and adjust the execution order of tasks and resource allocation in a timely manner when abnormal task execution and resource shortage are found.

[0041] In a preferred embodiment, in S104, the RPA robot performs task execution by periodic polling, monitors the RPA status in real time, establishes a connection with the operating environment of the RPA robot, obtains key information feedback during task execution, including the current step of the task and the amount of processed data, establishes a status indicator range for normal task execution, and monitors the time T from the start of task execution to the current moment. current , the average rate of normal data processing of the task is V normal , set the threshold time for a task to have no progress for a long time as T threshold , the current CPU usage U CPU , set the high CPU usage threshold to U threshold The duration that exceeds the threshold is T. duration , define the function F(T current, V normal , T threshold , U threshold , T duration ) is used to comprehensively judge the task status, which can be specifically expressed as:

[0042]

[0043] When F=0, it means that the task is executed normally. When F=1, it means that an abnormality occurs in the task. When an abnormality is detected in the task, the retry process is automatically started.

[0044] The beneficial effects of the present invention are: the present invention can intelligently arrange tasks according to task priority, time constraints and business value contribution, accurately allocate resources to key business data processing tasks, such as high-priority order processing, to ensure that tasks are completed in an orderly and efficient manner, reduce errors and delays caused by manual intervention, improve the accuracy and timeliness of data processing, monitor in real time during task execution, and detect anomalies in time. Whether the task has no progress for a long time or the resource usage is too high, it can respond quickly, automatically retry or notify the administrator in time, ensure the stability and reliability of system operation, reduce the risk of data loss due to failures, and fully record the task execution log, covering rich information such as start time, end time, execution results and resource usage. These data are helpful for subsequent in-depth auditing and analysis, so that enterprises can clearly understand the bottlenecks and optimization points in the data processing process, and provide strong data support for strategic decision-making. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 is a flow chart of the present invention; DETAILED DESCRIPTION

[0046] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of this application.

[0047] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this application, "plurality" means two or more, unless otherwise clearly and specifically defined.

[0048] In the description of the present application, the term "for example" is used to mean "used as an example, illustration or description". Any embodiment described as "for example" in the present application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is given to enable any technician in the field to implement and use the present invention. In the following description, details are listed for the purpose of explanation. It should be understood that a person of ordinary skill in the art can recognize that the present invention can be implemented without using these specific details. In other examples, well-known structures and processes will not be elaborated in detail to avoid unnecessary details to obscure the description of the present invention. Therefore, the present invention is not intended to be limited to the embodiments shown, but is consistent with the widest scope consistent with the principles and features disclosed in the present application.

[0049] like Figure 1 , this embodiment provides: an RPA artificial intelligence control data processing method, which specifically includes the following steps:

[0050] S101. Obtain enterprise data processing tasks, pre-process and standardize multi-source enterprise data, classify and store using a text classification model, and perform quality analysis on the stored data;

[0051] Furthermore, the enterprise data processing tasks to be performed are obtained, and the obtained enterprise data are preprocessed and standardized, and the data from different enterprise data sources are uniformly converted, and the preprocessed and standardized enterprise data are subjected to text classification to determine the business type to which the enterprise data belongs. The text is input into the text classification model to classify the text. When the output layer has n neurons, corresponding to n different business types, the net input of the output layer is recorded as Z L , generate probability distribution for each text in each category, the specific calculation formula is as follows:

[0052]

[0053] Among them, Pi represents the probability of the i-th category, Z i,L Represents the net input of the i-th neuron in the output layer, which is used to calculate the probability Pi corresponding to the text belonging to the i-th category. When the net input vector of the output layer is:

[0054]

[0055] For the first category, the specific calculation formula of its probability value P1 is as follows:

[0056]

[0057] By calculating the probability value of the text in each category, we get a probability vector P = [P 1 ,P2 ,...,P n ], indicating the probability distribution of each text belonging to different business types. Based on the probability distribution, the classification results of the text enterprise data are determined. According to the classification results, the enterprise data is classified and stored in the corresponding enterprise database table. The quality analysis of the stored enterprise data is performed, including calculating the accuracy, completeness and outlier detection of the enterprise data. When the total number of records in the enterprise data is n total When the accuracy verification rule is followed, each field of each record is checked one by one. For records that do not meet the accuracy rules, the number of records is n. error , calculate the accuracy of enterprise data, the specific formula is as follows:

[0058]

[0059] Accuracy indicates accuracy. We traverse the enterprise data and check whether the required fields in each record are empty. We record the number of records where the required fields are missing. missing , the specific calculation formula of missing rate is as follows:

[0060]

[0061] Outliers are detected by calculating the mean of all enterprise data in this field:

[0062]

[0063] Among them, x i represents the i-th enterprise data value of the field, n represents the number of enterprise data in the field, and calculates the standard deviation:

[0064]

[0065] Taking k as the judgment range of outliers, it will exceed Enterprise data outside the specified range is considered as outliers. For detected outliers, corresponding correction strategies are adopted according to the business meaning of the enterprise data.

[0066] It should be noted that after calculating the probability value corresponding to each category, the category described by the text will be determined based on the maximum probability value. For example, for a text classification problem with three categories, the probability vector P = [P1, P2, P3] is obtained after calculation. Assuming P1 = 0.6, P2 = 0.3, P3 = 0.1, because P1 is the maximum value in the probability vector, the text is classified as the first category.

[0067] Suppose in an enterprise data set containing sales order enterprise data, there are a total of n total= 1000 records. After checking the product number and customer information fields according to the rules, it is found that there are n error = When there are 50 records with errors, the accuracy error ratio

[0068] It should be noted that abnormal values ​​in enterprise data are corrected or marked as suspicious enterprise data. For duplicate values, deduplication is performed based on the unique identifier of the enterprise data. The field used to uniquely identify the record in the enterprise data set is determined, and the enterprise data is sorted according to the unique identifier field. The sorted enterprise data is traversed, and the displacement identifier field values ​​of adjacent records are compared. If the ID value of record i is equal to the ID value of record j (i≠j), record i and record j are determined to be duplicate records, one of the records is retained, the other duplicate records are deleted, and the deleted duplicate records are recorded, and missing values ​​are filled.

[0069] S102. Construct a linear regression model to analyze the impact of different factors on business indicators, identify key factors and their impact, and evaluate the overall explanatory power of these factors on sales through the coefficient of determination, i.e., the adjusted coefficient of determination;

[0070] Furthermore, based on the enterprise data after quality analysis, we obtain the enterprise data of factors affecting the enterprise business indicators and the corresponding business indicator enterprise data, establish a linear regression model, and determine the dependent variables and independent variables related to the business problem. When the business indicator we want to analyze is sales y i When the sales volume is affected, the factors are marketing expenses x i1 、Product price x i2 and the number of customers x i3 , where i represents the i-th sample, and there are n samples, that is, the number of samples in the training set is n, and the number of independent variables B = 3, then the linear regression model can be expressed as:

[0071] y i =β 0 +β 1 x i1 +β 2 x i2 +β 3 x i3 +∈ i

[0072] Among them, β 0 is the intercept term, which represents the value when the marketing cost, product price and number of customers are all 0, β 1 is the regression coefficient corresponding to the marketing expenses, indicating the impact of marketing expenses on sales, β 2 is the regression coefficient corresponding to product price, β 3 Represents the regression coefficient corresponding to the number of customers, ∈i is the error term, which indicates the impact of factors other than marketing expenses, product prices, and customer numbers on sales. When the number of samples in the test set is m, the true sales value of the i-th sample in the test set is The model predicts the sales value of i test samples as Calculate the average difference between the predicted value and the true value. The specific calculation formula is as follows:

[0073]

[0074] A small MSE value indicates that the model has a good prediction effect. 2 and the adjusted R 2 The value of can evaluate the overall explanatory power of marketing expenses, product prices and customer numbers on sales. The coefficient of determination R 2 The specific calculation formula is as follows:

[0075]

[0076] in, represents the mean of the actual sales of the test set, R 2 The value range is between 0 and 1. A value close to 1 indicates that the model fits the enterprise data well. 2 The company can judge the overall explanatory power of the currently selected marketing expenses, product prices and customer numbers on sales by the adjusted R 2 The specific calculation formula is as follows:

[0077]

[0078] Adjusted R 2 When the number of independent variables increases, when the new independent variables do not significantly improve the model fitting effect, the adjusted R 2 The value will decrease.

[0079] It should be noted that the regression coefficient obtained through training directly represents the degree of influence of each feature on the business indicator. For example, if the regression coefficient corresponding to the marketing cost is 0.5, it means that when other features remain unchanged, for every unit increase in the marketing cost, the business indicator increases by 0.5 units on average. For the product price feature, if its regression coefficient is -2, it means that for every unit increase in the product price, the sales volume decreases by 2 units on average. If the company plans to increase the product price, the decline in sales volume can be estimated based on this coefficient, so as to adjust the production plan in advance. If the regression coefficient corresponding to the number of customers is 0.3, that is, for every unit increase in the number of customers, the sales volume increases by 0.3 units on average, the company can know the positive impact of increasing the number of customers on sales, and adjust the production plan and inventory management strategy according to the expected growth in sales.

[0080] S103. After obtaining the RPA task list, allocate resources to the RPA robots in order of priority based on the final priority score, continuously monitor and adjust task execution and resource allocation in a timely manner;

[0081] Furthermore, obtain a list of tasks that need to be processed by RPA, including task name, task description, business process and data processing volume information, determine the priority of each task, and reasonably arrange the workflow of the RPA robot according to the priority of the task and the time requirement for data processing. During business peak periods, allocate limited resources to key business data processing tasks, and calculate the priority of tasks based on business value contribution and time constraints. When business value contribution has m evaluation dimensions, including revenue growth impact, customer satisfaction and cost reduction, the weight of the jth dimension is w. vj , the score of task i on the jth dimension is s ij , then the business value contribution score of task i is V i The specific calculation formula is as follows:

[0082]

[0083] Determine the time constraint score based on the task deadline, the current time is t now , the deadline of task i is t deadline When , the time constraint score of task i is T i The specific calculation formula is as follows:

[0084]

[0085] The final priority is calculated based on business value contribution and time constraints. The weight of the business value contribution score is w v , the weight of the time constraint score is w t , then the specific calculation formula for the final priority of task i is as follows:

[0086] P i =w v ×V i +w t ×T i

[0087] Sort tasks according to the final priority score, allocate resources to high-priority tasks first, allocate sufficient CPU and memory resources to high-priority order processing tasks, assign tasks to RPA robots according to the established scheduling strategy, continuously monitor the execution progress of tasks and resource usage, and adjust the execution order of tasks and resource allocation in a timely manner when abnormal task execution and resource shortage are found.

[0088] S104. Periodically poll the RPA robot task, connect to its operating environment to obtain information about the task steps and the amount of data processed, and determine the abnormality of the task operation based on the indicator range of normal task execution.

[0089] Furthermore, the RPA robot is polled regularly to monitor the status of the RPA robot in real time, establish a connection with the operating environment of the RPA robot, obtain key information feedback during the task execution process, including the current step of the task and the amount of data processed, establish the status indicator range for normal task execution, and monitor the time T from the start of task execution to the current moment. current , the average rate of normal data processing of the task is V normal , set the threshold time for a task to have no progress for a long time as T threshold , the current CPU usage U CPU , set the high CPU usage threshold to U threshold The duration that exceeds the threshold is T. duration , define the function F(T current , V normal , T threshold , U threshold , T duration ) is used to comprehensively judge the task status, which can be specifically expressed as:

[0090]

[0091] When F=0, it means that the task is executed normally. When F=1, it means that an abnormality occurs in the task. When an abnormality is detected in the task, the retry process is automatically started.

[0092] It should be noted that when the average rate of data processing of the task is 100 pieces / minute, and the threshold time is set to 30 minutes, the current T current= 20 minutes, CPU usage = 95%, the set CPU high threshold is 90%, and the duration is 5 minutes, first calculate the data volume related V normal ×T current =100×20=2000, V normal ×T threshold =100×30=3000. Since 2000<3000, there is an anomaly in the data volume.

[0093] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0094] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0095] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable enterprise data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable enterprise data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable enterprise data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0097] These computer program instructions may also be loaded onto a computer or other programmable enterprise data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for executing on the computer or other programmable device to implement the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0098] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0099] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A RPA artificial intelligence control data processing method, characterized in that: The following steps are involved: S101. Obtain enterprise data processing tasks, pre-process and standardize multi-source enterprise data, classify and store using a text classification model, and perform quality analysis on the stored data; S102. Construct a linear regression model to analyze the impact of different factors on business indicators, identify key factors and their impact, and evaluate the overall explanatory power of these factors on sales through the coefficient of determination, i.e., the adjusted coefficient of determination; S103. After obtaining the RPA task list, allocate resources to the RPA robots in order of priority based on the final priority score, continuously monitor and adjust task execution and resource allocation in a timely manner; S104. Periodically poll the RPA robot task, connect to its operating environment to obtain information about the task steps and the amount of data processed, and determine the abnormality of the task operation based on the indicator range of normal task execution.

2. According to claim 1, a RPA artificial intelligence control data processing method is characterized in that: In S101, the enterprise data processing task to be executed is obtained, and the obtained enterprise data is preprocessed and standardized, data from different enterprise data sources are uniformly converted, and the preprocessed and standardized enterprise data are subjected to text classification to determine the business type to which the enterprise data belongs. The text is input into the text classification model to classify the text. When the output layer has n neurons, corresponding to n different business types, the net input of the output layer is recorded as Z L , generate probability distribution for each text in each category, the specific calculation formula is as follows: Among them, Pi represents the probability of the i-th category, Z i,L Represents the net input of the i-th neuron in the output layer, which is used to calculate the probability Pi corresponding to the text belonging to the i-th category. When the net input vector of the output layer is: For the first category, the specific calculation formula of its probability value P1 is as follows: By calculating the probability value of the text in each category, we get a probability vector P = [P1, P2, ..., P n ], which indicates the probability distribution of each text belonging to different business types. Based on the probability distribution, the classification result of the text enterprise data is determined. According to the classification result, the enterprise data is classified and stored in the corresponding enterprise database table. The quality analysis of the stored enterprise data is performed, including calculating the accuracy and completeness of the enterprise data and detecting outliers.

3. The RPA artificial intelligence control data processing method according to claim 2 is characterized in that: The accuracy calculation is based on the accuracy verification rules, and each field of each record is checked one by one. When the total number of records in the enterprise data is n total When the number of records that do not meet the accuracy rule is n error , calculate the accuracy of enterprise data, the specific formula is as follows: Among them, accuracy represents accuracy.

4. The RPA artificial intelligence control data processing method according to claim 2 is characterized in that: The missing rate is calculated by traversing the enterprise data, checking whether the required fields in each record are empty, and recording the number of records where the required fields are missing. missing , the specific calculation formula of missing rate is as follows:

5. The RPA artificial intelligence control data processing method according to claim 2 is characterized in that: The detection of outliers is performed by calculating the mean of all enterprise data in this field: Among them, x i represents the i-th enterprise data value of the field, n represents the number of enterprise data in the field, and calculates the standard deviation: Taking k as the judgment range of outliers, Enterprise data outside the specified range is considered as outliers. For detected outliers, corresponding correction strategies are adopted according to the business meaning of the enterprise data.

6. The RPA artificial intelligence control data processing method according to claim 1 is characterized in that: In S102, based on the enterprise data after quality analysis, the enterprise data of factors affecting the enterprise business indicators and the corresponding business indicator enterprise data are obtained, a linear regression model is established, and the dependent variables and independent variables related to the business problem are determined. When the business indicator to be analyzed is sales volume y i When the sales volume is affected, the factors are marketing expenses x i1 、Product price x i2 and the number of customers x i3 , where i represents the i-th sample, and there are n samples, that is, the number of samples in the training set is n, and the number of independent variables B = 3, then the linear regression model can be expressed as: y i =β0+β1x i1 +β2x i2 +β3x i3 +∈ i Among them, β0 is the intercept term, which represents the value when marketing expenses, product price and number of customers are all 0, β1 is the regression coefficient corresponding to marketing expenses, which represents the influence of marketing expenses on sales, β2 is the regression coefficient corresponding to product price, β3 is the regression coefficient corresponding to the number of customers, ∈ i is the error term, which indicates the impact of factors other than marketing expenses, product prices, and customer numbers on sales. When the number of samples in the test set is m, the true sales value of the i-th sample in the test set is The model predicts the sales value of i test samples as Calculate the average difference between the predicted value and the true value. The specific calculation formula is as follows: A small MSE value indicates that the model has a good prediction effect. 2 and the adjusted R 2 The value of can evaluate the overall explanatory power of marketing expenses, product prices and customer numbers on sales. The coefficient of determination R 2 The specific calculation formula is as follows: in, represents the mean of the actual sales of the test set, R 2 The value range is between 0 and 1. A value close to 1 indicates that the model fits the enterprise data well. 2 The value of R can be used to determine the overall explanatory power of the currently selected marketing expenses, product prices, and customer numbers on sales. 2 The specific calculation formula is as follows: Adjusted R 2 When the number of independent variables increases, when the new independent variables do not significantly improve the model fitting effect, the adjusted R 2 The value will decrease.

7. The RPA artificial intelligence control data processing method according to claim 1 is characterized in that: In S103, a list of tasks that need to be processed by RPA is obtained, including task name, task description, business process and data processing volume information, and the priority of each task is determined. According to the priority of the task and the time requirement for data processing, the workflow of the RPA robot is reasonably arranged. During business peak periods, limited resources are allocated to key business data processing tasks, and the priority of the task is calculated based on business value contribution and time constraints.

8. The RPA artificial intelligence control data processing method according to claim 7 is characterized in that: When business value contribution has m evaluation dimensions, including revenue growth impact, customer satisfaction, and cost reduction, the weight of the jth dimension is w vj , the score of task i on the jth dimension is s ij , then the business value contribution score of task i is V i The specific calculation formula is as follows: Determine the time constraint score based on the task deadline, the current time is t now , the deadline of task i is t deadline When , the time constraint score of task i is T i The specific calculation formula is as follows: The final priority is calculated based on business value contribution and time constraints. The weight of the business value contribution score is w v , the weight of the time constraint score is w t , then the specific calculation formula for the final priority of task i is as follows: P i =w v ×V i +w t ×T i Sort tasks according to the final priority score, allocate resources to high-priority tasks first, allocate sufficient CPU and memory resources to high-priority order processing tasks, assign tasks to RPA robots according to the established scheduling strategy, continuously monitor the execution progress of tasks and resource usage, and adjust the execution order of tasks and resource allocation in a timely manner when abnormal task execution and resource shortage are found.

9. The RPA artificial intelligence control data processing method according to claim 1 is characterized in that: In S104, the RPA robot is polled regularly to perform tasks, the RPA status is monitored in real time, a connection is established with the operating environment of the RPA robot, key information feedback is obtained during the task execution process, including the current step of the task and the amount of data processed, a status indicator range for normal task execution is established, and the time T from the start of task execution to the current moment is monitored. current , the average rate of normal data processing of the task is V normal , set the threshold time for a task to have no progress for a long time as T threshold , the current CPU usage U CPU , set the high CPU usage threshold to U threshold The duration that exceeds the threshold is T. duration , define the function F(T current , V normal , T threshold , U threshold , T duration ) is used to comprehensively judge the task status, which can be specifically expressed as: When F=0, it means that the task is executed normally. When F=1, it means that an abnormality occurs in the task. When an abnormality is detected in the task, the retry process is automatically started.

Images