CN115081920B - Attendance check-in scheduling management method, system, equipment and storage medium - Google Patents

Abstract

The present invention discloses an attendance check-in scheduling management method, system, equipment and storage medium. The system includes: collecting weather data of the attendance location area in real time, and preprocessing the collected weather data to obtain an initial sample set; building a prediction model based on an improved generative adversarial network; calculating the loss function between the predicted target and the real target image; training the prediction model through the loss function, during which the discriminator D aims to correctly identify the real samples and correctly remove the generated false samples, and the generator G aims to minimize the probability that the generated prediction value is removed by the discriminator D until the trained prediction model is output; the weather data is input into the trained prediction model to obtain the predicted value, and the corresponding attendance operation is performed according to the predicted value. The present invention adopts an improved GAN network to monitor the weather conditions of the attendance location in real time, timely and automatically adjusts the attendance mode and attendance system, improves work efficiency, and avoids data errors.

Description

Attendance check-in scheduling management method, system, device and storage medium

Technical Field

The present invention belongs to the technical field of information processing, and in particular relates to an attendance check-in scheduling management method, system, equipment and storage medium.

Background technique

In the current sign-in method, the setting of holiday breaks is mostly done by manually modifying the attendance rules or pre-setting statutory holidays. However, for emergencies such as typhoons and other force majeure reasons that are not suitable for attendance, manual intervention is often required in advance. In addition, attendance data statistics, attendance rate, salary and other data also need to be manually exported and calculated. These operations are very dependent on manual operations to a certain extent. If manual operations are wrong or delayed, a large amount of erroneous and dirty data may be generated, which will also increase the workload and difficulty of statistical data in the later stage.

Summary of the invention

The main purpose of the present invention is to overcome the shortcomings and deficiencies of the prior art, and to provide an attendance scheduling management method, system, equipment and storage medium, which monitors the weather conditions at the attendance location in real time, and combines relevant information released by the local meteorological bureau such as emergency response notifications, early warning signal data, typhoon data, etc., to predict the weather conditions of the next day and determine whether it is necessary to stop work or classes. If necessary, the attendance of the day will be automatically set to rest and each attendance officer will be notified to avoid the occurrence of personnel safety accidents. Attendance data will be automatically counted according to the attendance parameters set by the management personnel on a regular basis every month, and attendance rate and salary and other related data will be calculated to improve the efficiency of relevant staff and avoid mistakes.

In order to achieve the above object, the present invention adopts the following technical solutions:

The present invention provides an attendance check-in scheduling management method, comprising the following steps:

Collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set;

A prediction model is constructed based on an improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the rectified linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the rectified linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and then the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output a prediction target;

Calculate the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network;

The prediction model is trained through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated fake samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will update the feature information of the data set respectively until the trained prediction model is output;

The weather data is input into a trained prediction model to obtain a prediction value, and corresponding attendance operations are performed according to the prediction value.

As a preferred technical solution, the weather data includes weather conditions, warning signals, typhoon data, S-band radar data and hail data;

The preprocessing includes data cleaning, data conversion and data integration.

As a preferred technical solution, the loss function between the calculated predicted target and the real target image is specifically:

The loss function consists of two parts:

in is the loss optimization function of the discriminator D, Loss optimization function for generator G;

and The calculation formula is as follows:

Wherein, L _bce is the cross entropy loss function; X and Y are the input image sequence and the real image data at the target moment respectively; for multiple scales k of the image, the generated results G _k are obtained respectively; N is the number of scales; _yi is the real value; _yi ′ is the predicted value; n is the number of samples in the cross entropy loss function.

As a preferred technical solution, the objective function of the discriminator D is as follows:

Where P _data (x) is the distribution of real data; x is a real data; P _(z) is the distribution of predicted data.

As a preferred technical solution, for the discriminator D, weather data samples {x ₁ ,x ₂ ,...,x _m } are obtained from the real data set, and historical weather data samples {z ₁ ,z ₂ ,...,z _m } are obtained from the real data set; for each historical image sample z _k , the prediction result G(z _k ) is obtained through the generator G, and the model parameter θ _d of the discriminator D is updated according to the following formula:

in is the discriminator loss function; D( ^xi ) is the discriminant result of the real weather data set; The discriminator's discrimination result on the image generated by the generator; is the gradient partial derivative; _θd is the model parameter variable. Each time the model is trained, the result is corrected and the training result is back-propagated to the model for correction; η is the learning rate.

As a preferred technical solution, for each iteration of the generator G, historical data samples {z ₁ ,z ₂ ,...,z _m } are obtained from the real data set, and the model parameters θ _g of the discriminator are updated according to the following formula:

in is the generator loss function; G(z ⁱ ) is the result calculated by real data and θ _g ; θ _g is the model parameter variable. The model corrects the result every time it is trained, and the training result is back-propagated to the model for correction.

As a preferred technical solution, during the training process, the first-generation generator generates predicted values, and then these predicted values and true values are put into the first-generation discriminator for learning, so that the first-generation discriminator can truly distinguish between the generated data and the real data; then there is a second-generation generator, and the data generated by the second-generation generator G can deceive the first-generation discriminator D. At this time, the second-generation discriminator D is trained again, and so on.

Another aspect of the present invention provides an attendance check-in scheduling management system, including a data acquisition module, a model building module, a loss function calculation module, a model training module and a prediction module;

The data collection module is used to collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set;

The model building module is used to build a prediction model based on the improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the corrected linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the corrected linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output a prediction target;

The loss function calculation module is used to calculate the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network;

The model training module is used to train the prediction model through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated false samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will respectively update the feature information of the data set until the trained prediction model is output;

The prediction module inputs the weather data into a trained prediction model to obtain a prediction value, and performs corresponding attendance operations according to the prediction value.

Another aspect of the present invention provides an electronic device, the electronic device comprising:

at least one processor; and,

a memory communicatively connected to the at least one processor; wherein,

The memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor to enable the at least one processor to execute the attendance check-in scheduling management and attendance check-in scheduling management method.

On the other hand, the present invention provides a computer-readable storage medium storing a program, characterized in that when the program is executed by a processor, the attendance check-in scheduling management method is implemented.

Compared with the prior art, the present invention has the following advantages and beneficial effects:

1. The present invention adopts the GAN network to monitor the weather conditions at the attendance location in real time, and automatically adjusts the attendance method and attendance system in a timely manner, eliminating complicated manual adjustments, improving the work efficiency of relevant staff, and avoiding data errors.

2. The present invention provides a convenient and fast attendance data statistics method. The attendance management personnel only need to set the attendance rules once, and then they can count the attendance data of the attendance personnel every month with one click, reducing the risk of manual statistics errors.

BRIEF DESCRIPTION OF THE DRAWINGS

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

FIG1 is a flow chart of an attendance check-in scheduling management method according to an embodiment of the present invention;

FIG2 is a schematic diagram of an image prediction model based on an improved GAN method according to an embodiment of the present invention;

FIG3 is a schematic diagram of a weather forecast adversarial learning model based on GAN according to an embodiment of the present invention;

FIG4 is a schematic diagram of attendance management after the model is deployed to the server according to the present invention;

5 is a structural diagram of an attendance check-in scheduling management system according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention.

Detailed ways

In order to enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without making creative work are within the scope of protection of the present application.

Reference to "embodiments" in this application means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described in this application may be combined with other embodiments.

As shown in FIG1 , the attendance check-in scheduling management method of this embodiment includes the following steps:

S1. Collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set.

Furthermore, the weather data includes actual weather conditions, warning signals, typhoon data, S-band radar data and hail data.

Furthermore, the preprocessing includes data cleaning, data conversion and data integration.

S2. A prediction model is established based on the improved Generative Adversarial Networks (GAN), as shown in Figure 2. Specifically:

S21, firstly, the input raw data is corrected and superimposed by a convolution layer and a corrected linear function ReLU to obtain a preliminary feature map, then the initial feature map is upsampled, and the extracted initial feature map is enlarged, so as to perform the subsequent three convolution layers and the corrected linear function ReLU correction with a higher resolution;

S22, then multiplying the initial feature map with the feature map generated by the fourth convolutional layer, and obtaining more specific target feature information through the spatial attention mechanism and the channel attention mechanism;

S23, up-sampling the acquired specific feature information again, and then using a convolution layer and a hyperbolic tangent function Sigmoid correction superposition to obtain a fifth feature map.

S24. By stacking 5 convolutional layers to extract features from the original image, a prediction model is established and the prediction target is output.

Furthermore, the loss function between the predicted target and the real target image is calculated, and then the model is trained through the loss function to optimize the model in the right direction.

The ReLU function formula is as follows:

The Sigmoid function formula is as follows:

S3, calculating the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network;

It is understandable that the GAN network is mainly composed of a generator G and a discriminator D. G is a network that generates images. It accepts a random noise z and then generates images through this noise. The generated data is recorded as G(z). D is a discriminant network that determines whether a picture is "real" (whether it is fabricated). Its input parameter is x, x represents a picture, and the output D(x) represents the probability that x is a real picture. If it is 1, it means it is a real picture, and if the output is 0, it means it cannot be a real picture, as shown in Figure 3.

Combined with the availability of the forecast model, the loss function consists of two parts.

in is the loss optimization function of the discriminator D, The loss optimization function for the generator G.

and The detailed formula is as follows:

Where L _bce is the cross entropy loss function; X and Y are the input image sequence and the real image data at the target moment respectively; for multiple scales k of the image, the generated results G _k are obtained respectively; N is the number of scales; _yi is the real value; _yi ′ is the predicted value; n is the number of samples in the cross entropy loss function.

S4, training the prediction model through the loss function;

During training, the goal of the discriminator D is to identify real samples as accurately as possible (output is "true", or 1), and to remove generated fake samples as accurately as possible (output is "false", or 0). These two goals correspond to the first and second terms of the objective function formula respectively. The goal of the generator G is the opposite of the discriminator, which is to minimize the probability of being removed by the discriminator as much as possible.

The objective function formula is as follows:

Where P _data (x) is the distribution of real data; x is a real data; P _(z) is the distribution of predicted data.

For each iteration, the discriminator D and the generator G will update the feature information of the data set respectively and gradually improve the recognition model.

For the discriminator D, obtain weather data samples {x ₁ ,x ₂ ,...,x _m } from the real data set, and then obtain historical weather data samples {z ₁ ,z ₂ ,...,z _m } from the real data set. For each historical image sample z _k , the prediction result G(z _k ) is obtained through the generator G, and the model parameter θ _d of the discriminator D is updated according to the following formula:

in is the discriminator loss function; D( ^xi ) is the discriminant result of the real weather data set; The discriminator's discrimination result on the image generated by the generator; is the gradient partial derivative; _θd is the model parameter variable. Each time the model is trained, the result is corrected and the training result is back-propagated to the model for correction; η is the learning rate, which is specified based on experience during training and the initial value is set to 4* ^10-5 .

For each iteration of the generator G, historical data samples {z ₁ ,z ₂ ,...,z _m } are obtained from the real data set, and the model parameters θ _g of the discriminator are updated according to the following formula:

in is the generator loss function; G(z ⁱ ) is the result calculated by real data and θ _g ; θ _g is the model parameter variable. The model corrects the result every time it is trained, and the training result is back-propagated to the model for correction.

The first-generation generator G generates predictions, and then puts these predictions and true values into the first-generation discriminator D for learning, so that the first-generation discriminator D can truly distinguish the generated data from the real data. Then there is a second-generation generator G. The data generated by the second-generation generator G can fool the first-generation discriminator D. At this point, the second-generation discriminator D is trained again, and so on.

The generator G and the discriminator D form a min-max game. During the training process, both sides continuously optimize themselves until a balance is reached, that is, neither side can get better, that is, the false samples are completely indistinguishable from the real samples, and the output prediction value is used for the near-term prediction.

S5. Input the weather data into the trained prediction model to obtain a prediction value, and perform corresponding attendance operations according to the prediction value.

Further, as shown in Figure 4, the obtained model is deployed on the server and added to the sign-in system. The system performs the corresponding attendance operation according to the predicted value and sends the confirmation information to the attendance supervisor. After the attendance supervisor confirms, the system will automatically adjust the attendance system and sign-in method. Finally, the attendance information is sent to the attendance personnel.

It should be noted that, for the sake of convenience, the aforementioned method embodiments are all expressed as a series of action combinations, but those skilled in the art should know that the present invention is not limited to the described order of actions, because according to the present invention, certain steps can be performed in other orders or simultaneously.

Based on the same idea as the attendance check-in scheduling management method in the above embodiment, the present invention also provides an attendance check-in scheduling management system, which can be used to execute the above attendance check-in scheduling management method. For ease of explanation, the structural diagram of the attendance check-in scheduling management system embodiment only shows the parts related to the embodiment of the present invention. Those skilled in the art can understand that the illustrated structure does not constitute a limitation on the device, and may include more or fewer components than shown in the diagram, or combine certain components, or arrange the components differently.

Please refer to FIG5 , in another embodiment of the present application, there is provided an attendance check-in scheduling management system 100 , the system comprising a data collection module 101 , a model building module 102 , a loss function calculation module 103 , a model training module 104 and a prediction module 105 ;

The data collection module 101 is used to collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set;

The model building module 102 is used to build a prediction model based on the improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the corrected linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the corrected linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output the prediction target;

The loss function calculation module 103 is used to calculate the loss function between the predicted target and the real target image, and the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network;

The model training module 104 is used to train the prediction model through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated false samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will respectively update the feature information of the data set until the trained prediction model is output;

The prediction module 105 inputs the weather data into a trained prediction model to obtain a prediction value, and performs corresponding attendance operations according to the prediction value.

It should be noted that the attendance check-in scheduling management system of the present invention corresponds one-to-one to the attendance check-in scheduling management method of the present invention. The technical features and beneficial effects described in the embodiment of the above-mentioned attendance check-in scheduling management method are applicable to the embodiment of the attendance check-in scheduling management. For specific contents, please refer to the description in the embodiment of the method of the present invention. It will not be repeated here. This is hereby declared.

In addition, in the implementation of the attendance check-in scheduling management system in the above-mentioned embodiment, the logical division of each program module is only an example. In actual application, the above-mentioned functions can be assigned to different program modules as needed, for example, for the configuration requirements of the corresponding hardware or the convenience of software implementation. That is, the internal structure of the attendance check-in scheduling management system is divided into different program modules to complete all or part of the functions described above.

Please refer to Figure 6. In one embodiment, an electronic device for implementing an attendance check-in scheduling management method is provided. The electronic device 200 may include a first processor 201, a first memory 202 and a bus, and may also include a computer program stored in the first memory 202 and executable on the first processor 201, such as an attendance check-in scheduling management program 203.

The first memory 202 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the first memory 202 can be an internal storage unit of the electronic device 200, such as a mobile hard disk of the electronic device 200. In other embodiments, the first memory 202 can also be an external storage device of the electronic device 200, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. equipped on the electronic device 200. Further, the first memory 202 can also include both an internal storage unit of the electronic device 200 and an external storage device. The first memory 202 can not only be used to store application software and various types of data installed in the electronic device 200, such as the code of the attendance check-in scheduling management program 203, but also can be used to temporarily store data that has been output or is to be output.

In some embodiments, the first processor 201 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips, etc. The first processor 201 is the control core (Control Unit) of the electronic device, and uses various interfaces and lines to connect various components of the entire electronic device, and executes various functions and processes data of the electronic device 200 by running or executing programs or modules stored in the first memory 202, and calling data stored in the first memory 202.

FIG6 merely shows an electronic device with components. Those skilled in the art will appreciate that the structure shown in FIG6 does not limit the electronic device 200 and may include fewer or more components than shown in the figure, or a combination of certain components, or a different arrangement of components.

The attendance check-in scheduling management program 203 stored in the first memory 202 in the electronic device 200 is a combination of multiple instructions. When running in the first processor 201, it can achieve:

Collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set;

A prediction model is constructed based on an improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the rectified linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the rectified linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and then the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output a prediction target;

Calculate the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network;

The prediction model is trained through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated fake samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will update the feature information of the data set respectively until the trained prediction model is output;

The weather data is input into a trained prediction model to obtain a prediction value, and corresponding attendance operations are performed according to the prediction value.

Furthermore, if the module/unit integrated in the electronic device 200 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile computer-readable storage medium. The computer-readable medium may include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a mobile hard disk, a magnetic disk, an optical disk, a computer memory, and a read-only memory (ROM).

Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and/or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

The technical features of the above embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

The above embodiments are preferred implementation modes of the present invention, but the implementation modes of the present invention are not limited to the above embodiments. Any other changes, modifications, substitutions, combinations, and simplifications that do not deviate from the spirit and principles of the present invention should be equivalent replacement methods and are included in the protection scope of the present invention.

Claims (6)

1. The attendance check-in scheduling management method comprises the following steps: Collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set; A prediction model is constructed based on an improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the rectified linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the rectified linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and then the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output a prediction target; Calculate the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network; The prediction model is trained through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated fake samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will update the feature information of the data set respectively until the trained prediction model is output; The weather data is input into a trained prediction model to obtain a prediction value, and corresponding attendance operations are performed according to the prediction value. 2. The attendance check-in scheduling management method according to claim 1, characterized in that the weather data includes weather conditions, warning signals, typhoon data, S-band radar data and hail data; The preprocessing includes data cleaning, data conversion and data integration. 3. The attendance check-in scheduling management method according to claim 1 is characterized in that, during the training process, the first-generation generator generates predicted values, and then these predicted values and true values are put into the first-generation discriminator for learning, so that the first-generation discriminator can truly distinguish the generated data from the real data; then there is a second-generation generator, and the data generated by the second-generation generator G can deceive the first-generation discriminator D. At this time, the second-generation discriminator D is trained again, and so on. 4. The attendance check-in scheduling management system is characterized by comprising a data acquisition module, a model building module, a loss function calculation module, a model training module and a prediction module; The data collection module is used to collect weather data of the attendance location area in real time, and pre-process the collected weather data to obtain an initial sample set; The model building module is used to build a prediction model based on the improved generative adversarial network, specifically: the input raw data is corrected and superimposed by the first convolution layer and the corrected linear function to obtain an initial feature map, then the initial feature map is upsampled, the extracted initial feature map is enlarged, and the subsequent three convolution layers and the corrected linear function are corrected with a higher resolution; then the initial feature map is multiplied with the feature map generated by the fourth convolution layer, and then the specific feature information is obtained through the spatial attention mechanism and the channel attention mechanism; finally, the specific feature information is upsampled again, and the fifth convolution layer and the hyperbolic tangent function are used to correct and superimpose to obtain the fifth feature map, and the feature map extracted from the original image by stacking five convolution layers is used to establish a prediction model, and the prediction model is used to output a prediction target; The loss function calculation module is used to calculate the loss function between the predicted target and the real target image, wherein the loss function includes the loss optimization function of the discriminator D of the generative adversarial network and the loss optimization function of the generator G of the generative adversarial network; The model training module is used to train the prediction model through the loss function. During the training, the goal of the discriminator D is to correctly identify the real samples and correctly remove the generated false samples. The goal of the generator G is to minimize the probability that the generated prediction value is removed by the discriminator D. The training is continuously iterated. For each iteration process, the discriminator D and the generator G will respectively update the feature information of the data set until the trained prediction model is output; The prediction module inputs the weather data into a trained prediction model to obtain a prediction value, and performs corresponding attendance operations according to the prediction value. 5. An electronic device, characterized in that the electronic device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores computer program instructions that can be executed by the at least one processor, and the computer program instructions are executed by the at least one processor so that the at least one processor can execute the attendance check-in scheduling management method as described in any one of claims 1-3. 6. A computer-readable storage medium storing a program, characterized in that when the program is executed by a processor, the attendance check-in scheduling management method described in any one of claims 1 to 3 is implemented.