Resource quantity processing method of service platform and electronic equipment

Abstract

The invention provides a resource quantity processing method of a service platform and electronic equipment. A plurality of undetermined function items can be displayed to a to-be-selected target for the to-be-selected target to select, after selection is completed, at least one configuration function item selected by the to-be-selected target is received, and in order to facilitate statistics, the number of the corresponding configuration function items is counted according to a preset statistical period; then counting the participation target quantity for the disposal item; counting the number of system interfaces opened by the system according to a preset counting period; and finally, inputting the configuration function item quantity, the participation target quantity and the system interface quantity of the to-be-selected target into a pre-trained service resource prediction model for resource prediction processing, so that the resource quantity corresponding to the to-be-selected target can be output, and the to-be-selected target executes a corresponding resource task according to the resource quantity. The whole process is automatically completed, manual participation is not needed, manual labor can be saved, and meanwhile the working efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a resource processing method and electronic device for a service platform. Background Art

[0002] The amount of resources required by some service platforms is generally determined and adjusted automatically based on human experience.

[0003] However, this method consumes a lot of human resources and has a high risk of human errors, which may lead to omissions and mistakes. Summary of the invention

[0004] In view of this, the purpose of this application is to propose a resource processing method and electronic device for a service platform to solve or partially solve the above technical problems.

[0005] Based on the above purpose, the present application provides a method for processing resource volume of a service platform, including:

[0006] Display multiple pending functional items to the target for selection;

[0007] receiving at least one configuration function item selected by the candidate from a plurality of pending function items, and counting the number of the configuration function items according to a predetermined statistical period;

[0008] Count the number of participants in the disposal project;

[0009] Determine the various system interfaces provided by the service platform and count the number of system interfaces according to the predetermined statistical period;

[0010] The number of configuration function items, the number of participating targets and the number of system interfaces of the target to be selected are input into the pre-trained service resource prediction model for resource prediction to obtain the resource quantity corresponding to the target to be selected.

[0011] Based on the same inventive concept, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable by the processor, wherein the processor implements the method as described above when executing the computer program.

[0012] From the above, it can be seen that the resource quantity processing method and electronic device of the service platform provided by the present application can display multiple pending function items to the selected target for the selected target to select. After the selection is completed, at least one configuration function item selected by the selected target will be received. For the convenience of statistics, the number of corresponding configuration function items will be counted according to the predetermined statistical period; then the number of participating targets for the disposal project will be counted; then the number of system interfaces opened by the system will be counted according to the predetermined statistical period; finally, the number of configuration function items, the number of participating targets and the number of system interfaces of the selected target will be input into the pre-trained service resource prediction model for resource prediction processing, so that the resource quantity corresponding to the selected target can be output, and then the selected target can perform the corresponding resource task according to the resource quantity. The whole process is completed automatically without human participation, which can save manpower and improve work efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 A flowchart of a method for processing resource volume of a service platform according to an embodiment of the present application;

[0015] Figure 2 This is a structural block diagram of a resource processing device of a service platform in an embodiment of the present application;

[0016] Figure 3 A schematic diagram of the structure of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.

[0018] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be the usual meanings understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but can include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0019] The embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0020] The resource amount processing method of the service platform proposed in the embodiment of the present application is as follows: Figure 1 As shown, it is applied to the service platform, including:

[0021] Step 101: present multiple pending function items to a target for selection.

[0022] In specific implementation, the target to be selected includes at least one of the following: user, enterprise or team. The pending function items specifically refer to the function items of various functional configurations required for various targets to be selected to perform corresponding tasks. For example, each pending function item is specifically displayed through a menu list.

[0023] Step 102: receiving at least one configuration function item selected by a candidate from a plurality of pending function items, and counting the number of configuration function items according to a predetermined statistical period.

[0024] In specific implementation, the predetermined statistical cycle includes: every month, every two months or every three months, which can be set according to actual needs.

[0025] Step 103: Count the target amount of participation in the disposal project.

[0026] During specific implementation, the processing project includes at least one of the following: an engineering processing project, a data processing project, a resource analysis project, a processing project for a specific enterprise, etc.

[0027] The participation target includes at least one of the following: participating users, participating enterprises, participating teams, etc.

[0028] Step 104, determine various types of system interfaces provided by the service platform, and count the number of system interfaces according to a predetermined statistical period.

[0029] During specific implementation, the service platform will provide various standard system interfaces for data transmission, including system interfaces required for various file data transmissions, and each type of file data corresponds to a system interface.

[0030] Step 105, input the number of configuration function items, the number of participating targets and the number of system interfaces of the target to be selected into a pre-trained service resource prediction model to perform resource prediction and obtain the resource quantity corresponding to the target to be selected.

[0031] In specific implementation, the service resource prediction model obtained by pre-training a neural network can analyze and process the number of configured function items, the number of participating targets and the number of system interfaces to predict the resources of the corresponding service and obtain the corresponding resource quantity.

[0032] Through the above scheme, multiple pending function items can be displayed to the target to be selected for selection. After the selection is completed, at least one configuration function item selected by the target to be selected will be received. For the convenience of statistics, the number of corresponding configuration function items will be counted according to the predetermined statistical period; then the number of participating targets for the disposal project will be counted; then the number of system interfaces opened by the system will be counted according to the predetermined statistical period; finally, the number of configuration function items, the number of participating targets and the number of system interfaces of the target to be selected will be input into the pre-trained service resource prediction model for resource prediction processing, so that the resource amount corresponding to the target to be selected can be output, and then the target to be selected can perform the corresponding resource task according to the resource amount. The whole process is completed automatically without human participation, which can save manpower and improve work efficiency.

[0033] In some embodiments, the process of constructing the service resource prediction model includes:

[0034] Step A1, determine the number of historical configuration function items corresponding to the target to be selected in each historical statistical period, as well as the historical participation target quantity for the disposal project, and determine the number of historical system interfaces of each historical statistical period of the service platform.

[0035] Step A2, forming a first training sample with the historical resource quantity corresponding to the historical configuration function item quantity, historical participation target quantity and historical system interface quantity mark corresponding to each historical statistical period, and combining the first training samples corresponding to all historical statistical periods into a first training set.

[0036] Step A3, constructing a first initial neural network including a first input layer, a configuration function hidden layer, a participation target quantity hidden layer, a system interface hidden layer, a statistical hidden layer and a first output layer.

[0037] Step A4, input the first training sample in the first training set into the first input layer, analyze the historical number of configuration function items in the first training sample using the configuration function hidden layer to obtain the configuration function analysis result, analyze the historical participation target quantity using the participation target quantity hidden layer to obtain the participation target quantity analysis result, analyze the historical number of system interfaces using the system interface hidden layer to obtain the system interface analysis result, perform statistical analysis on the configuration function analysis results, the target quantity analysis results and the system interface analysis results using the statistical hidden layer to obtain the training resource quantity and send it to the first output layer for output.

[0038] Step A5: Compare the training resource amount corresponding to each first training sample with the marked historical resource amount, and determine the first loss function.

[0039] In a specific implementation, the first loss function may be a hinge loss function or a cross entropy loss function.

[0040] Step A6, determining whether a first convergence value of the first loss function is less than or equal to a first convergence threshold.

[0041] Step A7, otherwise, according to the first loss value determined by the first loss function, the parameters of the first input layer, the configuration function hidden layer, the participation target quantity hidden layer, the system interface hidden layer, the statistical hidden layer and the first output layer in the first initial neural network are adjusted.

[0042] Step A8: If yes, the final first initial neural network is used as a service resource prediction model.

[0043] Through the above scheme, a service resource prediction model can be obtained that can accurately predict resources according to the number of configuration function items, the number of participating targets and the number of system interfaces.

[0044] In some embodiments, step A8 includes:

[0045] Step A81: Use the first training set as a first test set, wherein each first training sample is used as a first test sample.

[0046] Step A82: Input each first test sample in the first test set into the first initial neural network obtained finally for prediction processing to obtain the test resource amount.

[0047] Step A83, comparing each test resource amount with the marked historical resource amount, and calculating the first test accuracy rate corresponding to the test process.

[0048] Step A84, in response to determining that the first test accuracy is greater than or equal to the first accuracy threshold, using the finally obtained first initial neural network as the service resource prediction model.

[0049] In specific implementation, if the first test accuracy is less than the first accuracy threshold, a new first training set will be reselected to continue training the final first initial neural network according to the process of steps A1 to A8, and the first test set will be re-determined to be tested according to the process of the above steps A81 to A84, and it will be repeated until it is determined that the first test accuracy is greater than or equal to the first accuracy threshold, and the final first initial neural network will be used as the service resource prediction model.

[0050] Through the above scheme, the accuracy of the first initial neural network finally obtained is tested using the first test set, so that the service resource prediction model can perform resource prediction processing and the obtained resource amount is more accurate.

[0051] In some embodiments, the above steps 101 to 105 are repeatedly performed for different candidates, so that after the corresponding resource amounts are determined for the multiple candidates respectively, the following steps are further included:

[0052] Step 106: Send the resource amount to the corresponding target to be selected.

[0053] Step 107, receiving data resource information of multiple candidate targets with completed resource quantities, and retrieving task requirement information, comparing the data resource information of each candidate target with the task requirement information to obtain a comparison result, wherein each candidate target corresponds to a comparison result.

[0054] In specific implementation, the data resource information includes: the name of the target to be selected, the type of task to be performed, the ability value of the task to be performed, etc. The task requirement information includes: the type required for the corresponding task to be performed, the ability value required for the task to be performed, and the materials required for the task to be performed.

[0055] Step 108, obtaining the historical resource processing data after the completion of the historical tasks for each candidate target, inputting the historical resource processing data into the pre-trained task analysis model, performing task analysis processing, and obtaining the task execution index corresponding to each candidate target.

[0056] In specific implementation, each candidate target stores some historical resource processing data of historical tasks completed in historical time. The historical resource processing data is the corresponding completion status after completing the historical tasks, including: completion amount, completion degree value, completed task type, and degree value of meeting the corresponding task requirements after completion.

[0057] The corresponding task analysis model is a model that uses a neural network through learning and training, and can perform task analysis on various resource processing data to determine the corresponding task execution status.

[0058] In this way, the task analysis model can be used to perform task analysis on the historical resource processing data to obtain a task execution index. Each historical resource processing data of the selected target corresponds to a task execution index.

[0059] Step 109, retrieve the exception handling record information of each target to be selected in the predetermined time period.

[0060] In specific implementation, the predetermined time period can be within one year, two years or three years, etc., and the exception handling record information corresponding to various time periods of each candidate target is stored in the database in advance, so that the exception handling record information of the predetermined time period of each candidate target can be directly retrieved from the database.

[0061] The exception handling record is a record of errors or failures that occur when the corresponding candidate target processes the task.

[0062] Step 110, retrieve the material data volume corresponding to the processing task in each candidate target.

[0063] In specific implementation, the material data volume specifically refers to the material name, material quantity, material abnormality, and record information corresponding to the material abnormality of various materials used to process various task configurations in the corresponding selected target.

[0064] Step 111 , obtaining multiple attribute parameter information of each target to be selected, determining the attribute classification of each attribute parameter information, determining the attribute value corresponding to the attribute classification, and integrating all the attribute values ​​to obtain the total attribute value.

[0065] In specific implementation, the attribute parameter information includes the processing status of the corresponding selected target processing various tasks, such as the number of tasks processed, processing time, the number of tasks processed without exception, the number of tasks participated in, etc.

[0066] Each attribute parameter information is preset with a corresponding attribute value, and the specific value of the attribute parameter information is multiplied by the corresponding attribute value and then added to obtain the total attribute value corresponding to the target to be selected. Each target to be selected obtains a total attribute value.

[0067] Step 112, integrating the comparison result, task execution index, exception handling record information, material data volume and attribute total value to obtain information detection result.

[0068] In specific implementation, the numerical values ​​corresponding to the comparison result, task execution index, exception handling record information, material data volume and attribute total value can be multiplied to obtain the information detection result.

[0069] The numerical values ​​corresponding to the comparison result, task execution index, exception handling record information, material data volume and attribute total value may also be weighted and summed to obtain the information detection result.

[0070] Step 113, sorting the candidate targets according to the information detection results, selecting a predetermined number of candidate targets with the highest information detection results from the sorting as selected targets, generating communication information, and sending the communication information to the selected targets.

[0071] Through the above scheme, the data resource information of each candidate target can be compared with the corresponding task requirement information to obtain the corresponding comparison result; the task analysis processing is also performed based on the historical resource processing data corresponding to each candidate target after completing the historical task, using the pre-trained task analysis model to obtain the task execution index of each candidate target; then the exception processing record information of the predetermined time period corresponding to each candidate target and the material data volume corresponding to the processing task are retrieved; and according to the obtained multiple attribute parameter information of each candidate target, the corresponding attribute classification is determined, and the attribute value of the attribute classification is determined, so that these attribute values ​​can be integrated to obtain the total attribute value; finally, the obtained comparison results, task execution index, exception processing record information, material data volume and total attribute value are integrated to obtain the information detection result, so that the information detection result can integrate the various information situations, and the candidate targets with the highest predetermined number of information detection results obtained according to the sorting are used as the selected targets, so that the corresponding selected targets can be automatically determined without manual screening, and the communication information can also be automatically sent to the selected target, thereby prompting the selected target, which is convenient and fast, and improves efficiency.

[0072] In some embodiments, the process of constructing the task analysis model includes:

[0073] Step B1, obtaining a first predetermined number of sample resource processing data corresponding to each candidate target as a second training set, wherein each sample resource processing data in the second training set corresponds to a correctly marked task execution index.

[0074] Step B2, constructing a second initial neural network including a second input layer, a plurality of second hidden layers and a second output layer.

[0075] Step B3, determining the type of sample resource processing data, and setting the input port of the second input layer according to the type of sample resource processing data, wherein one input port corresponds to one type of sample resource processing data.

[0076] Step B4, determining the number of second hidden layers according to the type of sample resource processing data, connecting each second hidden layer to the input port of the type of sample resource processing data, and connecting each second hidden layer to the second output layer.

[0077] Step B5, input the sample resource processing data in the second training set from the corresponding input port of the second input layer of the second initial neural network according to the type of each sample resource processing data, and transmit it to the second hidden layer corresponding to the type of sample resource processing data for task analysis. Each second hidden layer sends the analysis result to the second output layer for integration to obtain the sample task execution index.

[0078] Step B6: compare the sample task execution index corresponding to each sample resource processing data with the marked correct task execution index, and determine the second loss function.

[0079] In a specific implementation, the second loss function may be a hinge loss function or a cross entropy loss function.

[0080] Step B7, determining whether a second convergence value of the second loss function is less than or equal to a second convergence threshold.

[0081] Step B8: Otherwise, according to the second loss value determined by the second loss function, the parameters of the second input layer, each second hidden layer and the second output layer in the second initial neural network are adjusted.

[0082] Step B9: if yes, use the final second initial neural network as the task analysis model.

[0083] Through the above scheme, a task analysis model can be obtained that can accurately perform task analysis processing on various resource processing data.

[0084] In some embodiments, step B9 includes:

[0085] Step B91, obtaining a second predetermined number of test resource processing data corresponding to each target to be selected as a second test set, wherein each test resource processing data corresponds to a correctly marked task execution index.

[0086] Step B92, inputting the processing data of each test resource in the second test set into the final second initial neural network for task analysis and processing to obtain the test task execution index.

[0087] Step B93, comparing the test task execution index of each test resource processing data with the marked correct task execution index, and statistically analyzing the second test accuracy rate corresponding to the test process.

[0088] Step B94, in response to determining that the second test accuracy is greater than or equal to the second accuracy threshold, using the finally obtained second initial neural network as the task analysis model.

[0089] In specific implementation, if the second test accuracy is less than the second accuracy threshold, the second test set will be used as the second training set to continue training the final second initial neural network according to the process of steps B1 to B9, and the second test set will be re-determined to be tested according to the process of the above steps B91 to B94, and it will be repeated until it is determined that the second test accuracy is greater than or equal to the second accuracy threshold, and the final second initial neural network will be used as the task analysis model.

[0090] Through the above scheme, the accuracy of the second initial neural network finally obtained is tested using the second test set, so that the obtained task analysis model can predict the task execution index more accurately.

[0091] In some embodiments, it also includes:

[0092] Step 114, determining a plurality of targets to be detected from a predetermined number of selected targets.

[0093] Step 115, obtaining material resource data corresponding to the target task, inputting the material resource data into a pre-constructed material fluctuation prediction model to perform a fluctuation analysis of the material resource quantity, and obtaining a fluctuation analysis result, wherein the material fluctuation prediction model is a model pre-constructed based on a neural network model, and the material fluctuation prediction model can analyze the fluctuation of the material resource quantity corresponding to the material resource data.

[0094] In a specific implementation, the material resource data is at least one of the name, type, price, quantity, and source information of the raw materials required for processing tasks related to the power equipment.

[0095] Step 116, for each target to be detected: determine the position information corresponding to the target to be detected, retrieve the condition information of the position information, obtain the adaptation parameter information of the target to be detected, match the adaptation parameter information with the condition information, and obtain the matching result of the target to be detected.

[0096] In specific implementation, since the conditions for the target to be detected are different at different locations, the adaptation parameter information of the target to be detected will be matched with the condition information of its location, and then the matching results of each target to be detected will be obtained. The matching results represent the degree to which the target to be detected meets the condition information of its location.

[0097] Step 117, obtaining abnormal execution data of each target to be detected for the target task.

[0098] During specific implementation, if there are records of device processing for each target to be detected in history, the records will be retrieved to see whether there are any abnormal situations that cannot be executed or whether the execution can be carried out smoothly when the processing is carried out on the device. These are all abnormal execution data.

[0099] Step 118, for each target to be detected: collect index parameter information of the target to be detected, input the index parameter information into a pre-built support vector machine model to perform abnormality analysis, and obtain an abnormality index.

[0100] In specific implementation, the support vector machine (SVM) model is a pre-built abnormality index that can predict whether interruptions, errors, abnormal stops, and other abnormal situations will occur when executing the current power equipment's tasks based on some indicator parameter information of each target to be detected.

[0101] The index parameter information includes: the resource volume of the target to be detected, the size of the area, the type of tools, the number of each type of tools, the type of materials and the number of each type of materials, etc.

[0102] Step 119, obtaining the target resource value of each target to be detected for the target task.

[0103] In a specific implementation, the target resource value is the resource demand provided by the target to be detected to execute the task of the power equipment.

[0104] Step 120 is performed for each target to be detected: combining the fluctuation analysis result, matching result, abnormal execution data, abnormality index and target resource value of the target to be detected to obtain a resource detection result.

[0105] Step 121, sorting the various targets to be detected according to the resource detection results, selecting the target to be detected with the highest resource detection result from the sorting as the resource determination target, and outputting and displaying the resource determination target.

[0106] Through the above scheme, the material fluctuation prediction model can be used to perform fluctuation analysis on the corresponding material resource data and obtain the fluctuation analysis results; after matching the corresponding adaptation parameter information of each target to be detected with the condition information of its corresponding position, the matching results of each target to be detected are obtained; and the abnormal execution data of each target to be detected are obtained; and after the abnormal degree analysis is performed using the support vector machine model, the abnormal degree index of each target to be detected is obtained; and the target resource value of each target to be detected is obtained; in this way, the fluctuation analysis result, matching result, abnormal execution data, abnormal degree index and target resource value of the target to be detected can be combined to determine the resource detection result, so that the resource detection result is more in line with various detection-related data, and the resource corresponding to the highest resource detection result screened based on the resource detection result is determined with higher accuracy and more in line with the requirements of the task.

[0107] In some embodiments, the process of constructing the material fluctuation prediction model includes:

[0108] Step C1, obtaining a plurality of historical material resource data corresponding to target tasks arranged in chronological order in a first time period in history, and graphing the plurality of historical material resource data to obtain a first curve graph.

[0109] In a specific implementation, each historical material resource data is classified according to raw materials, for example, classified into: raw material (1), raw material (2), raw material (3), raw material (4).

[0110] Then, for each category of raw materials corresponding to the historical material resource data, a graph is drawn, and each category of raw materials corresponds to a first curve graph. The first curve graph is a curve graph formed by taking time as the horizontal axis and the value of the corresponding historical material resource data as the vertical axis, and the trend of the historical material resources of the raw materials of this category over time can be learned from the first curve graph.

[0111] Step C2, determining the fluctuation index corresponding to each historical material resource data according to the first curve chart, and using the fluctuation index to mark each historical material resource data, combining multiple marked historical material resource data together to form a third training set, wherein the third training set includes multiple third training samples, and each third training sample is a marked historical material resource data.

[0112] In a specific implementation, the curvature corresponding to each time point is calculated according to the first curve graph, and the curvature is used as a fluctuation index. A positive value of the fluctuation index indicates an increasing fluctuation, and a negative value of the fluctuation index indicates a decreasing fluctuation.

[0113] In this way, each historical material resource data can be marked with a corresponding fluctuation index in chronological order, and these marked historical material resource data can be combined as third training samples into a third training set.

[0114] Step C3, constructing a third initial neural network model including a third input layer, multiple third hidden layers and a third output layer.

[0115] In a specific implementation, the third input layer includes a plurality of third input ports, each of which corresponds to a type of raw material and is used to receive historical material resource data of the raw material of the type. One historical material resource data may correspond to multiple raw materials.

[0116] The third hidden layer includes multiple layers, each of which corresponds to a category of raw materials and is connected to the corresponding third input port, and is used to perform fluctuation analysis on the historical material resource data of the raw materials of the category received by the third input port.

[0117] The third output layer is connected to each third hidden layer, and is used for receiving the fluctuation analysis results obtained by each third hidden layer, and integrating these fluctuation analysis results.

[0118] Step C4, input the third training samples in the third training set from the third input layer to the neural network model in chronological order, after the third input layer preprocesses the third training samples, the preprocessed third training samples are input to each third hidden layer for fluctuation analysis, each third hidden layer sends the analysis results to the output layer, and the third output layer is used to organize and output the fluctuation analysis training results.

[0119] In the specific implementation, the third training sample that needs to be input is determined in chronological order. After the third input layer determines the third input port corresponding to the third training sample, it reaches the corresponding third hidden layer, and is sent to the third output layer after the fluctuation analysis is performed using the corresponding third hidden layer. The third output layer is used to integrate the fluctuation analysis results of each third hidden layer to obtain the final fluctuation analysis training result output.

[0120] Step C5, determine the corresponding loss value according to the difference between the fluctuation analysis training result and the marked fluctuation index, determine the adjustment amount of each layer parameter according to the loss value, adjust the parameters of each layer of the third initial neural network model according to the adjustment amount of each layer parameter, and complete the training process for the third training sample.

[0121] In specific implementation, each time a third training sample is trained, the difference between its fluctuation analysis training result and the marked fluctuation index is determined. For example, the absolute value of the difference between the two is calculated as the loss value. If the loss value is larger, the corresponding parameter adjustment amount is also larger, which has a positive correlation.

[0122] In this way, the third initial neural network model is adjusted once each time a third training sample is trained, and when the next third training sample is trained, the last adjusted third initial neural network model is used to continue training and adjusting.

[0123] Step C6, after determining that all the third training samples in the third training set have been trained, the finally obtained third initial neural network model is used as the material fluctuation prediction model.

[0124] Through the above scheme, a material fluctuation prediction model is obtained which can accurately analyze the fluctuation of material resource data.

[0125] In some embodiments, step C6 comprises:

[0126] Step C61, after determining that all third training samples in the third training set have been trained, obtain a plurality of historical material resource data corresponding to the target tasks arranged in chronological order in the second time period in the history, and graph the plurality of historical material resource data to obtain a second curve graph;

[0127] The process of drawing the second curve graph is the same as that of drawing the first curve graph, and will not be described again here.

[0128] Step C62, determining the fluctuation index corresponding to each historical material resource data according to the second curve graph, and marking each historical material resource data in the second time period using the fluctuation index, combining a plurality of marked historical material resource data together to form a third test set, wherein the third test set includes a plurality of third test samples, each third test sample being the marked historical material resource data in the third time period;

[0129] Step C63, inputting each third test sample in the third test set into the third initial neural network model obtained finally in chronological order to perform fluctuation analysis test processing, and obtaining fluctuation analysis test results;

[0130] Step C64, comparing the fluctuation analysis test results corresponding to each third test sample with the marked fluctuation index, and calculating the third test accuracy rate corresponding to the test process;

[0131] Step C65, in response to determining that the third test accuracy is greater than or equal to the third accuracy threshold, using the finally obtained third initial neural network model as the material fluctuation prediction model.

[0132] In specific implementation, if the third test accuracy is less than the third accuracy threshold, the third test set will be used as the third training set to continue training the finally obtained third initial neural network model according to the process of steps C1 to C6, and the third test set will be re-determined to be tested according to the process of the above steps C61 to C64, and it will be repeated until it is determined that the third test accuracy is greater than or equal to the third accuracy threshold, and the finally obtained third initial neural network model will be used as the material fluctuation prediction model.

[0133] Through the above scheme, the accuracy of the third initial neural network model finally obtained is tested using the third test set, which can make the final material fluctuation prediction model predict fluctuations more accurately.

[0134] In some embodiments, the support vector machine model construction process includes:

[0135] Step D1, obtaining multiple historical indicator parameter information of various targets to be detected, and marking the corresponding abnormality degree for each historical indicator parameter information.

[0136] In specific implementation, the historical indicator parameter information includes: the resource volume, area size, tool type, quantity of each type of tool, material type and quantity of each type of material of each target to be detected in the historical time period.

[0137] And the degree of abnormality of each historical indicator parameter information will be known in advance (for example, abnormality is 0.5 and normality is 1).

[0138] Step D2, normalizing the multiple marked historical indicator parameter information of each target to be detected to obtain a normalized result.

[0139] During specific implementation, in order to ensure the consistency of each historical indicator parameter information, it is normalized.

[0140] Step D3, taking the normalization result as the fourth training set, wherein the fourth training set includes a plurality of fourth training samples, and each fourth training sample is historical indicator parameter information of the abnormality degree after normalization.

[0141] Step D4, constructing an initial support vector machine model.

[0142] In a specific implementation, the initial support vector machine is: Support Vector Machine, SVM, a generalized linear classifier that performs binary classification on data in a supervised learning manner.

[0143] Step D5, inputting each fourth training sample in the fourth training set into the initial support vector machine model for abnormality analysis processing to obtain the abnormality degree of the sample.

[0144] Step D6, determine the corresponding loss value according to the difference between the abnormality degree of the sample and the abnormality degree of the mark, determine the adjustment amount of each layer parameter of the initial support vector machine model according to the loss value, adjust the initial support vector machine model according to the adjustment amount of each layer parameter, and complete the training process for the fourth training sample.

[0145] In specific implementation, the loss value is calculated by performing hinge loss function calculation based on the abnormality degree of the sample and the abnormality degree of the label to obtain the loss value.

[0146] The hinge loss function has a relatively small amount of calculation and is relatively fast in the training process, which can better meet the needs of real-time prediction. In addition, the hinge loss function can also be applied to different types of data, such as discrete, continuous, and binary. In short, the hinge loss function is a very effective classification loss function that can better simulate data, improve the generalization ability of the model, and effectively suppress overfitting.

[0147] Step D7, after determining that all the fourth training samples in the fourth training set have been trained, the final obtained initial support vector machine model is used as the support vector machine model.

[0148] In specific implementation, according to the process of steps D1 to D3 above, a fourth test set is re-collected, the test set includes multiple fourth test samples, each fourth test sample is the historical indicator parameter information of the abnormal degree after normalization. The fourth test set can be the same as or different from the fourth training set.

[0149] Each fourth test sample in the fourth test set is input into the final initial support vector machine model for processing, the processing result is compared with the abnormal degree of the mark, and the fourth test accuracy corresponding to the fourth test set processing of the support vector machine model is calculated. If the fourth test accuracy is less than the fourth accuracy threshold, the fourth test set is used as the fourth training set to continue training the final initial support vector machine model according to the process of steps D1 to D7, and the fourth test set is re-determined for testing, and the process is repeated until it is determined that the fourth test accuracy is greater than or equal to the fourth accuracy threshold, and the final initial support vector machine is used as the support vector machine model.

[0150] Through the above scheme, a support vector machine model of anomaly program capable of accurately analyzing the index parameter information of the corresponding target to be detected is obtained.

[0151] In some embodiments, step D5 comprises:

[0152] Step D51, adding a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer and an activation processing layer to the initial support vector machine model.

[0153] Step D52, input each fourth training sample in the fourth training set into the initial support vector machine model and process it in sequence using a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer and an activation processing layer.

[0154] In a specific implementation, the linear kernel function processing layer is used to analyze the linear performance of the fourth training sample and determine the linear value corresponding to the abnormality degree of the fourth training sample;

[0155] The polynomial kernel function processing layer is used to analyze the performance of the polynomial function of the fourth training sample and determine the multinomial values ​​corresponding to the fourth training sample and the abnormality degree;

[0156] The Gaussian kernel function processing layer is used to analyze the performance of the Gaussian distribution of the fourth training sample and determine the Gaussian value corresponding to the abnormality degree of the fourth training sample;

[0157] The activation processing layer is used to analyze the activation performance of the fourth training sample and determine the activation value corresponding to the abnormality degree of the fourth training sample.

[0158] Step D53, integrating the processing results of each processing layer, determining the abnormal separation plane according to the integration result, calculating the distance of the fourth training sample from the abnormal separation plane, and determining the degree of sample abnormality according to the distance.

[0159] In specific implementation, the functional relationship between the linear value and the degree of abnormality, the functional relationship between the multinomial value and the degree of abnormality, the functional relationship between the Gaussian value and the degree of abnormality, and the functional relationship between the activation value and the degree of abnormality are integrated to determine a common plane functional relationship, and the plane functional relationship is used as the abnormal separation plane. In this way, the distance of each fourth training sample from the abnormal separation plane can be calculated, and then the degree of sample abnormality can be obtained according to the size of the distance.

[0160] Through the above scheme, the process of the degree of abnormality of the sample can be accurately determined.

[0161] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.

[0162] It should be noted that the above describes some embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0163] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a resource quantity processing device for a service platform.

[0164] refer to Figure 2, the device comprises:

[0165] The function display module 201 is configured to display a plurality of pending function items to a target to be selected;

[0166] The function quantity statistics module 202 is configured to receive at least one configuration function item selected by the candidate from a plurality of pending function items, and to count the number of configuration function items according to a predetermined statistical period;

[0167] A participation target quantity statistics module 203 is configured to count the participation target quantity for the treatment project;

[0168] The interface quantity statistics module 204 is configured to determine various types of system interfaces provided by the service platform and to count the number of system interfaces according to a predetermined statistical period;

[0169] The resource prediction module 205 is configured to input the number of configuration function items, the number of participating targets and the number of system interfaces of the target to be selected into a pre-trained service resource prediction model to perform resource prediction and obtain the resource quantity corresponding to the target to be selected.

[0170] In some embodiments, the apparatus includes a first model building module configured to:

[0171] Determine the number of historical configuration function items corresponding to the selected targets in each historical statistical period, as well as the number of historical participation targets for the disposal projects, and determine the number of historical system interfaces for each historical statistical period of the service platform;

[0172] The historical resource quantities corresponding to the historical configuration function item quantity, the historical participation target quantity, and the historical system interface quantity mark corresponding to each historical statistical period form a first training sample, and the first training samples corresponding to all historical statistical periods are combined into a first training set;

[0173] Constructing a first initial neural network including a first input layer, a configuration function hidden layer, a participation target quantity hidden layer, a system interface hidden layer, a statistical hidden layer and a first output layer;

[0174] Input the first training sample in the first training set to the first input layer, analyze the number of historical configuration function items in the first training sample using the configuration function hidden layer to obtain the configuration function analysis result, analyze the historical participation target quantity using the participation target quantity hidden layer to obtain the participation target quantity analysis result, analyze the historical number of system interfaces using the system interface hidden layer to obtain the system interface analysis result, perform statistical analysis on the configuration function analysis result, the target quantity analysis result and the system interface analysis result using the statistical hidden layer to obtain the training resource quantity and send it to the first output layer for output;

[0175] Compare the training resource amount corresponding to each first training sample with the marked historical resource amount, and determine a first loss function;

[0176] Determine whether a first convergence value of the first loss function is less than or equal to a first convergence threshold;

[0177] Otherwise, according to the first loss value determined by the first loss function, the parameters of the first input layer, the configuration function hidden layer, the participation target quantity hidden layer, the system interface hidden layer, the statistical hidden layer and the first output layer in the first initial neural network are adjusted;

[0178] If yes, the final first initial neural network is used as the service resource prediction model.

[0179] In some embodiments, the first model building module is specifically configured to:

[0180] The method of using the final first initial neural network as a service resource prediction model includes:

[0181] Using the first training set as a first test set, wherein each first training sample is used as a first test sample;

[0182] Inputting each first test sample in the first test set into the finally obtained first initial neural network for prediction processing to obtain the test resource amount;

[0183] Compare each test resource amount with the marked historical resource amount, and calculate the first test accuracy rate corresponding to the test process;

[0184] In response to determining that the first test accuracy is greater than or equal to the first accuracy threshold, the finally obtained first initial neural network is used as the service resource prediction model.

[0185] In some embodiments, the apparatus includes an information detection module configured to:

[0186] After respectively determining corresponding resource amounts for a plurality of candidate targets, the resource amounts are sent to the corresponding candidate targets;

[0187] Receiving data resource information of multiple candidate targets with completed resource quantities, retrieving task requirement information, and comparing the data resource information of each candidate target with the task requirement information to obtain a comparison result, wherein each candidate target corresponds to a comparison result;

[0188] Obtain historical resource processing data for each candidate target after the completion of historical tasks, input the historical resource processing data into a pre-trained task analysis model, perform task analysis processing, and obtain a task execution index corresponding to each candidate target;

[0189] Retrieve the exception handling record information for each target to be selected during the scheduled time period;

[0190] Retrieve the material data volume corresponding to the processing task in each candidate target;

[0191] Acquire multiple attribute parameter information of each target to be selected, determine the attribute classification of each attribute parameter information, determine the attribute value corresponding to the attribute classification, and integrate all attribute values ​​to obtain the total attribute value;

[0192] Integrate the comparison results, task execution index, exception handling record information, material data volume and attribute total value to obtain information detection results;

[0193] The candidate targets are sorted according to the information detection results, a predetermined number of candidate targets with the highest information detection results are selected from the sorting as selected targets, communication information is generated, and the communication information is sent to the selected targets.

[0194] In some embodiments, the apparatus further comprises a second model building module configured to:

[0195] Acquire a first predetermined number of sample resource processing data corresponding to each candidate target as a second training set, wherein each sample resource processing data in the second training set corresponds to a correctly marked task execution index;

[0196] constructing a second initial neural network including a second input layer, a plurality of second hidden layers, and a second output layer;

[0197] Determine the type of sample resource processing data, and set an input port of the second input layer according to the type of sample resource processing data, wherein one input port corresponds to one type of sample resource processing data;

[0198] Determine the number of second hidden layers according to the type of sample resource processing data, connect each second hidden layer to an input port of the type of corresponding sample resource processing data, and connect each second hidden layer to the second output layer;

[0199] The sample resource processing data in the second training set are input from the corresponding input port of the second input layer of the second initial neural network according to the type of each sample resource processing data, and transmitted to the second hidden layer corresponding to the type of the sample resource processing data for task analysis. Each second hidden layer sends the analysis result to the second output layer for integration to obtain the sample task execution index;

[0200] Compare the sample task execution index corresponding to each sample resource processing data with the marked correct task execution index, and determine a second loss function;

[0201] Determine whether a second convergence value of the second loss function is less than or equal to a second convergence threshold;

[0202] Otherwise, adjusting the parameters of the second input layer, each second hidden layer, and the second output layer in the second initial neural network according to the second loss value determined by the second loss function;

[0203] If yes, the final second initial neural network is used as the task analysis model.

[0204] In some embodiments, the apparatus further comprises a resource detection module configured to:

[0205] determining a plurality of targets to be detected from a predetermined number of selected targets;

[0206] Obtaining material resource data corresponding to the target task, inputting the material resource data into a pre-built material fluctuation prediction model to perform fluctuation analysis on the material resource quantity, and obtaining a fluctuation analysis result, wherein the material fluctuation prediction model is a model pre-built based on a neural network model, and the material fluctuation prediction model can analyze the fluctuation of the material resource quantity corresponding to the material resource data;

[0207] For each target to be detected: determine the position information corresponding to the target to be detected, retrieve the condition information of the position information, obtain the adaptation parameter information of the target to be detected, match the adaptation parameter information with the condition information, and obtain the matching result of the target to be detected;

[0208] Obtain abnormal execution data of each target to be detected for the target task;

[0209] For each target to be detected: collecting the index parameter information of the target to be detected, inputting the index parameter information into a pre-built support vector machine model to perform abnormality analysis, and obtaining an abnormality index;

[0210] Obtain the target resource value for each target to be detected for the target task;

[0211] For each target to be detected, the following steps are performed: combining the fluctuation analysis result, matching result, abnormal execution data, abnormal degree index and target resource value of the target to be detected to obtain the resource detection result;

[0212] The various targets to be detected are sorted according to the resource detection results, and the target to be detected with the highest resource detection result is selected from the sorting as the resource determination target, and the resource determination target is output and displayed.

[0213] In some embodiments, the apparatus further comprises a third model building module configured to:

[0214] Acquire a plurality of historical material resource data corresponding to the target tasks arranged in chronological order in a first time period in history, and graph the plurality of historical material resource data to obtain a first curve graph;

[0215] Determine the fluctuation index corresponding to each historical material resource data according to the first curve graph, and use the fluctuation index to mark each historical material resource data, and combine multiple marked historical material resource data to form a third training set, wherein the third training set includes multiple third training samples, and each third training sample is a marked historical material resource data;

[0216] Constructing a third initial neural network model including a third input layer, a plurality of third hidden layers and a third output layer;

[0217] Input the third training samples in the third training set from the third input layer to the neural network model in chronological order, after the third input layer preprocesses the third training samples, input the preprocessed third training samples to each third hidden layer for fluctuation analysis, each third hidden layer sends the analysis results to the output layer, and the third output layer is used to organize and output the fluctuation analysis training results;

[0218] Determine a corresponding loss value according to the difference between the fluctuation analysis training result and the marked fluctuation index, determine the adjustment amount of each layer parameter according to the loss value, adjust the parameters of each layer of the third initial neural network model according to the adjustment amount of each layer parameter, and complete the training process for the third training sample;

[0219] After all the third training samples in the third training set are determined to have been trained, the finally obtained third initial neural network model is used as the material fluctuation prediction model.

[0220] In some embodiments, the apparatus further comprises a fourth model building module configured to:

[0221] Obtain multiple historical indicator parameter information of various targets to be detected, and mark the corresponding abnormality degree for each historical indicator parameter information;

[0222] Normalizing the multiple marked historical indicator parameter information of each target to be detected to obtain a normalized result;

[0223] The normalized result is used as a fourth training set, wherein the fourth training set includes a plurality of fourth training samples, and each fourth training sample is historical indicator parameter information of a normalized marked abnormality degree;

[0224] Build an initial support vector machine model;

[0225] Inputting each fourth training sample in the fourth training set into the initial support vector machine model for abnormality analysis processing to obtain the abnormality degree of the sample;

[0226] Determine a corresponding loss value according to the difference between the abnormality degree of the sample and the abnormality degree of the mark, determine the adjustment amount of each layer parameter of the initial support vector machine model according to the loss value, adjust the initial support vector machine model according to the adjustment amount of each layer parameter, and complete the training process for the fourth training sample;

[0227] After it is determined that all the fourth training samples in the fourth training set have been trained, the final obtained initial support vector machine model is used as the support vector machine model.

[0228] In some embodiments, the fourth model building module is specifically configured to:

[0229] The step of inputting each fourth training sample in the fourth training set into the initial support vector machine model for abnormality analysis to obtain the abnormality degree of the sample comprises:

[0230] Adding a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer, and an activation processing layer to the initial support vector machine model;

[0231] Input each fourth training sample in the fourth training set into the initial support vector machine model and process them in sequence using a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer and an activation processing layer;

[0232] The processing results of each processing layer are integrated, an abnormal separation plane is determined according to the integrated result, a distance of the fourth training sample deviating from the abnormal separation plane is calculated, and the degree of sample abnormality is determined according to the distance.

[0233] For the convenience of description, the above device is described in terms of functions divided into various modules. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.

[0234] The device of the above embodiment is used to implement the corresponding method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0235] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method described in any of the above embodiments when executing the program.

[0236] Figure 3A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.

[0237] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.

[0238] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.

[0239] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0240] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0241] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).

[0242] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.

[0243] The electronic device of the above embodiment is used to implement the corresponding method in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be described in detail here.

[0244] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the method described in any of the above embodiments.

[0245] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.

[0246] The computer instructions stored in the storage medium of the above embodiments are used to enable the computer to execute the method described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.

[0247] Based on the same concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a computer program product, including computer program instructions. When the computer program instructions are run on a computer, the computer executes the method described in any of the above embodiments, which has the beneficial effects of the corresponding method embodiments and will not be repeated here.

[0248] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.

[0249] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.

[0250] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.

[0251] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.

Claims

1. A method for processing resource volume of a service platform, characterized in that: include: Display multiple pending functional items to the target for selection; receiving at least one configuration function item selected by the candidate from a plurality of pending function items, and counting the number of the configuration function items according to a predetermined statistical period; Count the number of participants in the disposal project; Determine the various system interfaces provided by the service platform and count the number of system interfaces according to the predetermined statistical period; The number of configuration function items, the number of participating targets and the number of system interfaces of the target to be selected are input into the pre-trained service resource prediction model for resource prediction to obtain the resource quantity corresponding to the target to be selected.

2. The method according to claim 1, characterized in that: The construction process of the service resource prediction model includes: Determine the number of historical configuration function items corresponding to the selected targets in each historical statistical period, as well as the number of historical participation targets for the disposal projects, and determine the number of historical system interfaces for each historical statistical period of the service platform; The historical resource quantities corresponding to the historical configuration function item quantity, the historical participation target quantity, and the historical system interface quantity mark corresponding to each historical statistical period form a first training sample, and the first training samples corresponding to all historical statistical periods are combined into a first training set; Constructing a first initial neural network including a first input layer, a configuration function hidden layer, a participation target quantity hidden layer, a system interface hidden layer, a statistical hidden layer and a first output layer; Input the first training sample in the first training set to the first input layer, analyze the number of historical configuration function items in the first training sample using the configuration function hidden layer to obtain the configuration function analysis result, analyze the historical participation target quantity using the participation target quantity hidden layer to obtain the participation target quantity analysis result, analyze the historical number of system interfaces using the system interface hidden layer to obtain the system interface analysis result, perform statistical analysis on the configuration function analysis result, the target quantity analysis result and the system interface analysis result using the statistical hidden layer to obtain the training resource quantity and send it to the first output layer for output; Compare the training resource amount corresponding to each first training sample with the marked historical resource amount, and determine a first loss function; Determine whether a first convergence value of the first loss function is less than or equal to a first convergence threshold; Otherwise, according to the first loss value determined by the first loss function, the parameters of the first input layer, the configuration function hidden layer, the participation target quantity hidden layer, the system interface hidden layer, the statistical hidden layer and the first output layer in the first initial neural network are adjusted; If yes, the final first initial neural network is used as the service resource prediction model.

3. The method according to claim 2, characterized in that The method of using the final first initial neural network as a service resource prediction model includes: Using the first training set as a first test set, wherein each first training sample is used as a first test sample; Inputting each first test sample in the first test set into the finally obtained first initial neural network for prediction processing to obtain the test resource amount; Compare each test resource amount with the marked historical resource amount, and calculate the first test accuracy rate corresponding to the test process; In response to determining that the first test accuracy is greater than or equal to the first accuracy threshold, the finally obtained first initial neural network is used as the service resource prediction model.

4. The method according to claim 1, characterized in that After determining the corresponding resource amounts for multiple candidate targets respectively, the following steps are also included: Send the resource quantity to the corresponding target to be selected; Receiving data resource information of multiple candidate targets with completed resource quantities, retrieving task requirement information, and comparing the data resource information of each candidate target with the task requirement information to obtain a comparison result, wherein each candidate target corresponds to a comparison result; Obtain historical resource processing data for each candidate target after the completion of historical tasks, input the historical resource processing data into a pre-trained task analysis model, perform task analysis processing, and obtain a task execution index corresponding to each candidate target; Retrieve the exception handling record information for each target to be selected during the scheduled time period; Retrieve the material data volume corresponding to the processing task in each candidate target; Acquire multiple attribute parameter information of each target to be selected, determine the attribute classification of each attribute parameter information, determine the attribute value corresponding to the attribute classification, and integrate all attribute values ​​to obtain the total attribute value; Integrate the comparison results, task execution index, exception handling record information, material data volume and attribute total value to obtain information detection results; The candidate targets are sorted according to the information detection results, a predetermined number of candidate targets with the highest information detection results are selected from the sorting as selected targets, communication information is generated, and the communication information is sent to the selected targets.

5. The method according to claim 4, characterized in that The construction process of the task analysis model includes: Acquire a first predetermined number of sample resource processing data corresponding to each candidate target as a second training set, wherein each sample resource processing data in the second training set corresponds to a correctly marked task execution index; constructing a second initial neural network including a second input layer, a plurality of second hidden layers, and a second output layer; Determine the type of sample resource processing data, and set an input port of the second input layer according to the type of sample resource processing data, wherein one input port corresponds to one type of sample resource processing data; Determine the number of second hidden layers according to the type of sample resource processing data, connect each second hidden layer to an input port of the type of corresponding sample resource processing data, and connect each second hidden layer to the second output layer; The sample resource processing data in the second training set are input from the corresponding input port of the second input layer of the second initial neural network according to the type of each sample resource processing data, and transmitted to the second hidden layer corresponding to the type of the sample resource processing data for task analysis. Each second hidden layer sends the analysis result to the second output layer for integration to obtain the sample task execution index; Compare the sample task execution index corresponding to each sample resource processing data with the marked correct task execution index, and determine a second loss function; Determine whether a second convergence value of the second loss function is less than or equal to a second convergence threshold; Otherwise, adjusting the parameters of the second input layer, each second hidden layer, and the second output layer in the second initial neural network according to the second loss value determined by the second loss function; If yes, the final second initial neural network is used as the task analysis model.

6. The method according to claim 4, characterized in that Also includes: determining a plurality of targets to be detected from a predetermined number of selected targets; Obtaining material resource data corresponding to the target task, inputting the material resource data into a pre-built material fluctuation prediction model to perform fluctuation analysis on the material resource quantity, and obtaining a fluctuation analysis result, wherein the material fluctuation prediction model is a model pre-built based on a neural network model, and the material fluctuation prediction model can analyze the fluctuation of the material resource quantity corresponding to the material resource data; For each target to be detected: determine the position information corresponding to the target to be detected, retrieve the condition information of the position information, obtain the adaptation parameter information of the target to be detected, match the adaptation parameter information with the condition information, and obtain the matching result of the target to be detected; Obtain abnormal execution data of each target to be detected for the target task; For each target to be detected: collecting the index parameter information of the target to be detected, inputting the index parameter information into a pre-built support vector machine model to perform abnormality analysis, and obtaining an abnormality index; Obtain the target resource value for each target to be detected for the target task; For each target to be detected, the following steps are performed: combining the fluctuation analysis result, matching result, abnormal execution data, abnormal degree index and target resource value of the target to be detected to obtain the resource detection result; The various targets to be detected are sorted according to the resource detection results, and the target to be detected with the highest resource detection result is selected from the sorting as the resource determination target, and the resource determination target is output and displayed.

7. The method according to claim 6, characterized in that The process of constructing the material fluctuation prediction model includes: Acquire a plurality of historical material resource data corresponding to the target tasks arranged in chronological order in a first time period in history, and graph the plurality of historical material resource data to obtain a first curve graph; Determine the fluctuation index corresponding to each historical material resource data according to the first curve graph, and use the fluctuation index to mark each historical material resource data, and combine multiple marked historical material resource data to form a third training set, wherein the third training set includes multiple third training samples, and each third training sample is a marked historical material resource data; Constructing a third initial neural network model including a third input layer, a plurality of third hidden layers and a third output layer; Input the third training samples in the third training set from the third input layer to the neural network model in chronological order, after the third input layer preprocesses the third training samples, input the preprocessed third training samples to each third hidden layer for fluctuation analysis, each third hidden layer sends the analysis results to the output layer, and the third output layer is used to organize and output the fluctuation analysis training results; Determine a corresponding loss value according to the difference between the fluctuation analysis training result and the marked fluctuation index, determine the adjustment amount of each layer parameter according to the loss value, adjust the parameters of each layer of the third initial neural network model according to the adjustment amount of each layer parameter, and complete the training process for the third training sample; After all the third training samples in the third training set are determined to have been trained, the finally obtained third initial neural network model is used as the material fluctuation prediction model.

8. The method according to claim 6, characterized in that The construction process of the support vector machine model includes: Obtain multiple historical indicator parameter information of various targets to be detected, and mark the corresponding abnormality degree for each historical indicator parameter information; Normalizing the multiple marked historical indicator parameter information of each target to be detected to obtain a normalized result; The normalized result is used as a fourth training set, wherein the fourth training set includes a plurality of fourth training samples, and each fourth training sample is historical indicator parameter information of a normalized marked abnormality degree; Build an initial support vector machine model; Inputting each fourth training sample in the fourth training set into the initial support vector machine model for abnormality analysis processing to obtain the abnormality degree of the sample; Determine a corresponding loss value according to the difference between the abnormality degree of the sample and the abnormality degree of the mark, determine the adjustment amount of each layer parameter of the initial support vector machine model according to the loss value, adjust the initial support vector machine model according to the adjustment amount of each layer parameter, and complete the training process for the fourth training sample; After it is determined that all the fourth training samples in the fourth training set have been trained, the final obtained initial support vector machine model is used as the support vector machine model.

9. The method according to claim 8, characterized in that The step of inputting each fourth training sample in the fourth training set into the initial support vector machine model for abnormality analysis to obtain the abnormality degree of the sample comprises: Adding a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer, and an activation processing layer to the initial support vector machine model; Input each fourth training sample in the fourth training set into the initial support vector machine model and process them in sequence using a linear kernel function processing layer, a polynomial kernel function processing layer, a Gaussian kernel function processing layer and an activation processing layer; The processing results of each processing layer are integrated, an abnormal separation plane is determined according to the integrated result, a distance of the fourth training sample deviating from the abnormal separation plane is calculated, and the degree of sample abnormality is determined according to the distance.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method according to any one of claims 1 to 9 is implemented.

Images