Supervised time sequence water level data generation method and system and storage medium

Abstract

The invention relates to the technical field of machine learning, in particular to a supervised time sequence water level data generation method and system and a storage medium. The method comprises the following steps: acquiring multi-source hydrological monitoring data; performing abnormal value identification and data preprocessing according to the hydrological monitoring data to generate standardized hydrological basic data; constructing a data generation model of the conditional generative adversarial network based on the standardized hydrological basic data; hydrological variable driven condition feature regulation and control are carried out on the data generation model, parameter optimization of the generation network is carried out, and hydrological adaptability enhanced condition input feature data are generated; and injecting the condition input characteristic data into the data generation model to generate initial water level time sequence data. According to the method, the rare extreme hydrological events are oversampled by using a hierarchical importance sampling and self-service method (Bootstrap), so that the performance of the generative model in rare event prediction is remarkably enhanced.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a supervised time series water level data generation method, system and storage medium. Background Art

[0002] "Supervised" refers to a learning method in which the model learns from labeled training data. In supervised learning, each training sample has an output label associated with it, and the goal of the model is to learn the relationship between the input data (features) and the output label (target). Specifically in the "Supervised Time Series Water Level Data Generation Method", "supervised" means: in this case, the time series water level data is labeled, that is, the water level data at each time point has a corresponding label, which may be the category, level, or some other form of label of the water level. The model learns to predict patterns of water level changes by analyzing these labeled water level data. For example, it may learn how the water level will change under certain specific environmental conditions. Once the model is trained, it can be used to predict future water level changes or classify new water level data.

[0003] However, in flood warning systems, the current supervised time-series water level data generation methods often have the following problems: Flood data itself is highly unbalanced, and extreme flood events are relatively rare, resulting in far fewer positive samples (flood events) than negative samples (normal water levels) in the training data. This imbalance seriously affects the performance of the model in identifying high-risk flood events. The occurrence pattern of hydrological extreme events is difficult to predict. Traditional models lack the ability to adapt to such long-term hydrological change trends. Summary of the invention

[0004] Based on this, it is necessary for the present invention to provide a supervised time-series water level data generation method and system to solve at least one of the above technical problems.

[0005] To achieve the above purpose, a supervised time series water level data generation method includes the following steps: Step S1: Acquire multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on the hydrological monitoring data to generate standardized hydrological basic data; construct a data generation model of a conditional generative adversarial network based on the standardized hydrological basic data; Step S2: The data generation model is regulated by the conditional characteristics driven by the hydrological variables, the parameters of the generation network are optimized, and the conditional input characteristic data with enhanced hydrological adaptability are generated; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; Step S3: Perform stratified importance sampling based on preliminary water level time series data, and perform differentiated generation weight setting based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; perform distribution consistency evaluation based on statistical characteristics based on extreme event data and optimized sampling data for hydrological events to obtain water level time series data; Step S4: Perform multi-scale verification on the water level time series data to obtain water level time series verification data, wherein the multi-scale verification includes spectral characteristic verification and mutual information measurement; perform artifact elimination on the water level time series verification data, and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

[0006] The present invention realizes the comprehensive collection of multi-channel hydrological information by acquiring multi-source hydrological monitoring data, which helps to comprehensively cover the data characteristics of different sources; on this basis, outlier recognition and data preprocessing steps are used to eliminate data noise and abnormal points, generate standardized hydrological basic data, and ensure the accuracy and consistency of the data. This provides a high-quality input source for subsequent data modeling and avoids the error propagation of the original data. Furthermore, the data generation model is constructed through the conditional generative adversarial network, so that the model has strong adaptive ability, can simulate complex hydrological characteristics and data distribution characteristics, and improves the flexibility and controllability of data generation. By introducing the conditional feature regulation driven by hydrological variables, the parameters of the data generation model are optimized, which can better capture the impact of hydrological conditions on hydrological changes, thereby generating conditional input feature data with hydrological adaptability. This process improves the generalization ability of the model in hydrological diversity scenarios; by injecting the conditional input feature data into the generation model, the generated preliminary water level time series data has high authenticity and relevance, laying the foundation for subsequent optimization processing. Through stratified importance sampling, the ability to identify key features of preliminary water level time series data is improved. In combination with rare extreme hydrological event data, differentiated generation weight settings are performed, the sampling strategy is optimized, and the focus on extreme events is enhanced. The use of the bootstrap method to oversample rare event data not only expands the number of samples, but also improves the balance of the data. The distribution consistency assessment based on statistical characteristics further ensures the consistency of the generated hydrological data with the real scene, thereby obtaining more reliable water level time series data. Through multi-scale verification, including spectral characteristic test and mutual information measurement, the frequency domain characteristics of water level time series data and the dependency relationship between variables are strictly tested, which enhances the scientificity and comprehensiveness of data verification; subsequently, artifact elimination and extreme value outlier correction based on physical constraints are performed to ensure the physical rationality and consistency of the data. These measures effectively reduce the artifact interference in the generated data and improve the quality and credibility of the final generated supervised time series water level data.

[0007] The present invention also provides a supervised time series water level data generation system, which is used to execute the above-mentioned supervised time series water level data generation method, and the supervised time series water level data generation system comprises: Model building module, used to obtain multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on hydrological monitoring data to generate standardized hydrological basic data; build a data generation model of conditional generative adversarial network based on standardized hydrological basic data; The hydrological condition enhancement module is used to regulate the conditional characteristics driven by hydrological variables in the data generation model, optimize the parameters of the generation network, and generate conditional input characteristic data with enhanced hydrological adaptability; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; The sampling optimization module is used to perform stratified importance sampling based on preliminary water level time series data, and to perform differentiated generation weight settings based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; to oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; to perform distribution consistency evaluation based on statistical characteristics based on extreme event data and hydrological event optimized sampling data to obtain water level time series data; The data verification and correction module is used to perform multi-scale verification on the water level time series data to obtain water level time series verification data, where the multi-scale verification includes spectral characteristic verification and mutual information measurement; the water level time series verification data is used to eliminate artifacts in the data and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

[0008] The model building module in the present invention provides a solid data foundation for the subsequent generation of water level time series data by acquiring multi-source hydrological monitoring data. The module first cleans the original data through outlier identification and data preprocessing, removes noise and errors that do not conform to the normal hydrological change law, and generates standardized hydrological basic data. This standardization process not only improves the consistency and availability of the data, but also ensures that the subsequent modeling will not be disturbed by abnormal data. Based on these standardized hydrological basic data, a data generation model of a conditional generative adversarial network (GAN) is constructed, which can simulate the complex laws of water level changes and generate time series data that conform to actual hydrological characteristics, laying the foundation for the accuracy and reliability of the entire model. The hydrological condition enhancement module plays a key role in generating water level time series data. The module can optimize the parameters of the generation network according to the actual situation of hydrological changes by regulating the conditional characteristics driven by hydrological variables on the data generation model, so that the generated data is enhanced in hydrological adaptability. This characteristic data with enhanced hydrological adaptability ensures that the generated water level time series data can more realistically reflect the hydrological changes under different conditions and provide more accurate prediction results. The sampling optimization module optimizes the data generation process so that the generated water level time series data is more in line with the actual application requirements. By performing stratified importance sampling on the preliminary water level time series data, the representativeness of the generated data can be improved. Especially in the face of rare extreme hydrological events, these events can be sampled preferentially through differentiated generation weight settings, thereby generating more accurate extreme event data. Oversampling by the self-help method can further improve the quality of extreme event data, so that the model's ability to identify these key events is enhanced. Finally, these data are processed by distribution consistency evaluation based on statistical characteristics to ensure the accuracy and consistency of water level time series data in different scenarios. The data verification and correction module comprehensively checks the generated water level time series data through a multi-scale verification method to ensure the reliability of the data. Spectral characteristic verification and mutual information measurement provide two important perspectives for the verification process. Spectral characteristic verification can check the energy distribution characteristics of data in different frequency ranges, and mutual information measurement reveals the correlation and information flow between data, which helps to detect whether the data meets the expected hydrological characteristics. Artifact elimination and extreme value outlier correction further improve the quality of the data and avoid inaccurate results caused by measurement errors or noise. This process ensures that the generated water level time series data not only conforms to the actual hydrological laws, but also has sufficient physical rationality and accuracy, and finally obtains high-quality supervised time series water level data. In summary, these modules cooperate with each other to provide a complete process from data acquisition, generation, optimization to verification and correction, ensuring the accuracy, reliability and applicability of water level time series data, and providing strong data support for applications such as hydrological prediction and risk assessment.

[0009] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program is executed to implement the supervised time-series water level data generation method as described above. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments thereof made with reference to the following drawings: Figure 1 A schematic diagram of the steps of the supervised time series water level data generation method of the present invention; Figure 2 for Figure 1 Detailed step flow diagram of step S1; Figure 3 for Figure 1 Detailed step flow chart of step S2 in FIG. DETAILED DESCRIPTION

[0011] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0012] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0013] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0014] To achieve this, please refer to Figures 1 to 3 The present invention provides a supervised method for generating time series water level data, the method comprising the following steps: Step S1: Acquire multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on the hydrological monitoring data to generate standardized hydrological basic data; construct a data generation model of a conditional generative adversarial network based on the standardized hydrological basic data; Step S2: The data generation model is regulated by the conditional characteristics driven by the hydrological variables, the parameters of the generation network are optimized, and the conditional input characteristic data with enhanced hydrological adaptability are generated; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; Step S3: Perform stratified importance sampling based on preliminary water level time series data, and perform differentiated generation weight setting based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; perform distribution consistency evaluation based on statistical characteristics based on extreme event data and optimized sampling data for hydrological events to obtain water level time series data; Step S4: Perform multi-scale verification on the water level time series data to obtain water level time series verification data, wherein the multi-scale verification includes spectral characteristic verification and mutual information measurement; perform artifact elimination on the water level time series verification data, and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

[0015] In the embodiment of the present invention, reference Figure 1 FIG. 1 is a flow chart of the steps of a supervised time series water level data generation method of the present invention. In this example, the supervised time series water level data generation method includes the following steps: Step S1: Acquire multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on the hydrological monitoring data to generate standardized hydrological basic data; construct a data generation model of a conditional generative adversarial network based on the standardized hydrological basic data; The embodiment of the present invention obtains hydrological monitoring data of a certain river basin, including water level, rainfall, evaporation and flow information recorded by 5 monitoring stations. The monitoring period is from January 2020 to December 2023, and the collection frequency is once an hour. The original hydrological monitoring data are format converted and timestamp calibrated to ensure that all data are consistent in time. The outlier recognition technology based on statistical methods is used to remove abnormal points in the water level data that exceed 3 times the standard deviation, and the Lagrange interpolation method is used to fill the missing values. Subsequently, the data is further interpolated and repaired based on the time series regression model of adjacent data points to obtain continuous hydrological monitoring cleaned data. For the cleaned data, the mutual information method is used to screen the rainfall and evaporation characteristics with high correlation with the temporal changes of water level height, and generate derived variables such as daily rainfall intensity and cumulative evaporation. Finally, the extended feature data is dimensionless and segmented according to weeks to obtain standardized basic hydrological data. With the standardized data as input, a conditional generative adversarial network is constructed, in which the generator adopts a two-layer LSTM and one-dimensional convolutional hybrid network structure, and the discriminator designs a consistency detection module based on the time and frequency domains. It also defines distribution similarity loss, temporal consistency loss and physical constraint loss functions to complete the construction of the data generation model.

[0016] Step S2: The data generation model is regulated by the conditional characteristics driven by the hydrological variables, the parameters of the generation network are optimized, and the conditional input characteristic data with enhanced hydrological adaptability are generated; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; The embodiment of the present invention utilizes historical hydrological variable data provided by meteorological stations, including temperature (range: -5°C to 40°C), humidity (30% to 90%), precipitation (0 to 200 mm) and atmospheric pressure (950 to 1050 hPa), performs principal component analysis (PCA) on these data, and extracts key hydrological features with a variance contribution rate of more than 95%; utilizes a hydrological variable encoding network based on an attention mechanism to embed the extracted hydrological features into a 32-dimensional vector; then, adaptively optimizes the parameters of the generator network through a variational inference method to enhance the adaptability of the model to the hydrological features; finally, injects the generated conditional input feature data into the generator network for adversarial training to generate preliminary water level time series data with a length of 365 days, wherein the time resolution of the data is at the hourly level, and the data distribution simulates the real historical hydrological trend.

[0017] Step S3: Perform stratified importance sampling based on preliminary water level time series data, and perform differentiated generation weight setting based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; perform distribution consistency evaluation based on statistical characteristics based on extreme event data and optimized sampling data for hydrological events to obtain water level time series data; The embodiment of the present invention divides the initially generated water level time series data into regular intervals and rare intervals based on a stratified sampling weight strategy according to the mean and variance of the data distribution, and assigns a higher sampling weight to the rare interval to obtain stratified importance sampling data; using historical extreme hydrological event data (such as a flood event caused by a once-in-a-century rainstorm in 2021) as a reference, a differentiated generation weight strategy is adopted to improve the sampling probability of extreme event data; the rare extreme hydrological event samples are oversampled 10 times through the self-service method to generate high-quality extreme event data; the stratified importance sampling data and the oversampled extreme event data are combined, and the statistical feature consistency is evaluated based on the Kolmogorov-Smirnov test, and the distribution characteristics of the inconsistent data are adjusted using interpolation and smoothing methods to obtain water level time series data with consistent statistical distribution.

[0018] Step S4: Perform multi-scale verification on the water level time series data to obtain water level time series verification data, wherein the multi-scale verification includes spectral characteristic verification and mutual information measurement; perform artifact elimination on the water level time series verification data, and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

[0019] The embodiment of the present invention uses wavelet transform to verify the spectral characteristics of the time series characteristics of the generated water level time series data, focuses on analyzing the response of high-frequency components to rainfall intensity, and the reflection of low-frequency components to long-term water level changes, and calculates the difference in spectral energy distribution between the real water level data and the generated data; adopts the mutual information measurement method to analyze the nonlinear correlation between water level and variables such as rainfall and evaporation, to ensure that the correlation characteristics of the generated data are consistent with those of the real data; after merging the verification results into water level time series verification data, uses an anomaly recognition algorithm based on a deep residual network to eliminate data artifacts, and corrects abnormal extreme values ​​through physical constraint methods (such as the hydrological continuity equation); finally, high-quality supervised time series water level data that meets hydrological laws and statistical characteristics are obtained.

[0020] The present invention realizes the comprehensive collection of multi-channel hydrological information by acquiring multi-source hydrological monitoring data, which helps to comprehensively cover the data characteristics of different sources; on this basis, outlier recognition and data preprocessing steps are used to eliminate data noise and abnormal points, generate standardized hydrological basic data, and ensure the accuracy and consistency of the data. This provides a high-quality input source for subsequent data modeling and avoids the error propagation of the original data. Furthermore, the data generation model is constructed through the conditional generative adversarial network, so that the model has strong adaptive ability, can simulate complex hydrological characteristics and data distribution characteristics, and improves the flexibility and controllability of data generation. By introducing the conditional feature regulation driven by hydrological variables, the parameters of the data generation model are optimized, which can better capture the impact of hydrological conditions on hydrological changes, thereby generating conditional input feature data with hydrological adaptability. This process improves the generalization ability of the model in hydrological diversity scenarios; by injecting the conditional input feature data into the generation model, the generated preliminary water level time series data has high authenticity and relevance, laying the foundation for subsequent optimization processing. Through stratified importance sampling, the ability to identify key features of preliminary water level time series data is improved. In combination with rare extreme hydrological event data, differentiated generation weight settings are performed, the sampling strategy is optimized, and the focus on extreme events is enhanced. The use of the bootstrap method to oversample rare event data not only expands the number of samples, but also improves the balance of the data. The distribution consistency assessment based on statistical characteristics further ensures the consistency of the generated hydrological data with the real scene, thereby obtaining more reliable water level time series data. Through multi-scale verification, including spectral characteristic test and mutual information measurement, the frequency domain characteristics of water level time series data and the dependency relationship between variables are strictly tested, which enhances the scientificity and comprehensiveness of data verification; subsequently, artifact elimination and extreme value outlier correction based on physical constraints are performed to ensure the physical rationality and consistency of the data. These measures effectively reduce the artifact interference in the generated data and improve the quality and credibility of the final generated supervised time series water level data.

[0021] Preferably, step S1 comprises the following steps: Step S11: collecting water level, rainfall, evaporation and flow information to obtain original hydrological monitoring data; Step S12: performing format conversion and time stamp calibration processing on the original hydrological monitoring data, thereby obtaining time-aligned multi-source hydrological monitoring data; Step S13: performing outlier identification based on statistical methods on the multi-source hydrological monitoring data, and performing interpolation repair processing and outlier removal processing to obtain hydrological monitoring filtered data; Step S14: performing data filling processing on the hydrological monitoring filtered data based on the time series regression model of adjacent data points, thereby obtaining hydrological monitoring cleansing data; Step S15: screening the hydrological monitoring cleaning data for characteristic variables highly correlated with the water level time series changes through the mutual information method, and generating derived variables, thereby obtaining hydrological characteristic extended data; Step S16: performing dimensionless processing on the hydrological characteristic extended data and performing data segmentation based on different time windows, thereby obtaining standardized hydrological basic data; Step S17: Construct a data generation model of a conditional generative adversarial network based on standardized hydrological basic data.

[0022] As an embodiment of the present invention, refer to Figure 2 As shown, Figure 1 Detailed step flow diagram of step S1 in the embodiment of the present invention, step S1 includes the following steps: Step S11: collecting water level, rainfall, evaporation and flow information to obtain original hydrological monitoring data; The embodiment of the present invention takes a typical river as the research object, and deploys 5 hydrological monitoring stations to record water level (range: 0.5-5m), rainfall (0-300mm), evaporation (0-100mm) and flow (0-500m³ / s) respectively. The data collection period is from January 2020 to December 2023, once every hour; data is collected through automated monitoring equipment, and the data is transmitted to the cloud data center through the Internet of Things gateway to form original hydrological monitoring data; the storage format of the monitoring data is a CSV file, which contains the date, time, site number and the corresponding monitoring value. Some files have inconsistent formats or missing fields.

[0023] Step S12: performing format conversion and time stamp calibration processing on the original hydrological monitoring data, thereby obtaining time-aligned multi-source hydrological monitoring data; For the collected original hydrological monitoring data, the embodiment of the present invention first uses Python's pandas library to read the CSV file, standardizes the time field into the "YYYY-MM-DD HH:MM:SS" format, and unifies it into a 24-hour timestamp; for records with missing time, the time difference between the two previous and next data is used to automatically generate a completed timestamp; the multi-source data is classified and sorted according to the site number, and the data of each site is converted into a unified multi-dimensional table format, where each column corresponds to water level, rainfall, evaporation and flow, thereby generating time-aligned multi-source hydrological monitoring data.

[0024] Step S13: performing outlier identification based on statistical methods on the multi-source hydrological monitoring data, and performing interpolation repair processing and outlier removal processing to obtain hydrological monitoring filtered data; The embodiment of the present invention uses a statistical outlier identification method based on a box plot method for time-aligned multi-source hydrological monitoring data to first remove outliers in the water level data that exceed 1.5 times the upper and lower quartile range; for the interpolation and repair processing of outliers, a Lagrange interpolation method is used to estimate a single outlier point, and the interpolation results of continuous outliers are smoothed using a local average method; completely erroneous data rows detected (such as all variables are outliers at the same time) are directly removed; after the repair is completed, a histogram statistical analysis is performed on the distribution of the repaired data to verify the rationality of the data range, thereby obtaining hydrological monitoring filtered data.

[0025] Step S14: performing data filling processing on the hydrological monitoring filtered data based on the time series regression model of adjacent data points, thereby obtaining hydrological monitoring cleansing data; The embodiment of the present invention uses a time series regression model based on historical time point data to fill missing data in hydrological monitoring filtered data; the specific method is to use water level, rainfall, evaporation and flow as multivariate inputs, and use the hourly time interval as the prediction step of the regression model to train a multi-layer linear regression model to predict missing values; for data intervals that are continuously missing for more than 12 hours, the average value based on historical similar days is used to fill in the data; after filling, the data is reordered and an error analysis is performed to ensure that the interpolation error does not exceed 5%, thereby generating complete hydrological monitoring cleaning data.

[0026] Step S15: screening the hydrological monitoring cleaning data for characteristic variables highly correlated with the water level time series changes through the mutual information method, and generating derived variables, thereby obtaining hydrological characteristic extended data; The embodiment of the present invention uses the mutual information method to analyze the correlation between the water level time series change and other variables (rainfall, evaporation, flow) for the hydrological monitoring cleaning data; selects variables with mutual information values ​​higher than 0.8 as relevant characteristic variables; generates derived variables according to the selected characteristic variables, such as rainfall intensity (rainfall / hour), cumulative evaporation (accumulated value of evaporation) and flow change rate (differential value of flow); adds the derived variables to the original data to form an extended data set, and performs a correlation test on the extended variables to ensure that they have a reasonable physical explanation with the water level time series change, and finally obtains the hydrological characteristic extended data.

[0027] Step S16: performing dimensionless processing on the hydrological characteristic extended data and performing data segmentation based on different time windows, thereby obtaining standardized hydrological basic data; In the embodiment of the present invention, all variables in the hydrological characteristic extended data are normalized according to the formula Dimensionless processing is performed to ensure that the value range of all variables is between [0,1]. For time series data, the data is segmented by week and divided into multiple 7-day time windows, each window containing 168 hours of observation data. For data segments spanning different months, the boundary data of the corresponding months are supplemented through a sliding window method, thereby generating standardized hydrological basic data covering the entire year.

[0028] Step S17: Construct a data generation model of a conditional generative adversarial network based on standardized hydrological basic data.

[0029] The embodiment of the present invention uses the standardized basic hydrological data generated in step S16 as input to construct a conditional generative adversarial network (CGAN); in a specific implementation, the generator adopts a network composed of two layers of LSTM units and a one-dimensional convolutional layer, and the input is standardized hydrological data and conditional feature variables (such as rainfall and evaporation), and the output is predicted water level time series data; the discriminator network extracts time series features through the convolutional layer, and combines the conditional features to discriminate the distribution difference between the generated data and the real data; the loss function of the generation network is defined as the time series distribution difference loss, the conditional constraint loss and the physical consistency loss, and after 1000 rounds of adversarial training, a conditional generative adversarial network model with stable performance is finally generated.

[0030] The present invention obtains the original hydrological monitoring data by collecting key hydrological information such as water level, rainfall, evaporation and flow. This step ensures the comprehensiveness and diversity of the data and provides a reliable basis for subsequent data analysis and modeling. It covers different hydrological variables and helps to comprehensively evaluate the driving factors of hydrological changes. The original hydrological monitoring data is format converted and timestamp calibrated to ensure the time alignment of multi-source data. This step ensures the timeliness and comparability of the data, avoids errors caused by time inconsistency, and thus provides a unified time frame for subsequent analysis, so that various types of hydrological data can be effectively combined. Through the outlier identification based on statistical methods, combined with the processing methods of interpolation repair and outlier removal, the unreasonable points in the data are effectively removed. This processing greatly improves the quality of the data, removes outliers that may interfere with the analysis results, and ensures the accuracy and reliability of subsequent data. On the basis of the hydrological monitoring filtered data, the missing data is filled by the time series regression model, which further optimizes the continuity and integrity of the data. The time series regression model can effectively restore the missing values ​​by referring to the values ​​of adjacent data points, thereby ensuring the stability and availability of the data. By using the mutual information method to perform feature screening and generate derived variables for the hydrological monitoring cleaning data, the key characteristic variables closely related to the temporal changes of the water level can be accurately extracted. This step improves the relevance and effectiveness of the data and provides valuable input variables for subsequent model construction. The hydrological characteristic extension data is dimensionless, making the comparison between different features more scientific and reasonable. By segmenting the data based on different time windows, it is ensured that the data can show different hydrological characteristics according to different time scales. This process helps to standardize the data, eliminates the impact of the dimension on the analysis results, and improves the effect and accuracy of model training. The data generation model of the conditional generative adversarial network is constructed based on the standardized hydrological basic data, which further improves the model's simulation ability for complex hydrological characteristics. This step can effectively generate data that conforms to the actual hydrological laws by using the adversarial network, providing strong data support for subsequent hydrological prediction and analysis.

[0031] Preferably, step S17 comprises the following steps: Step S171: constructing a generator network structure of a conditional generative adversarial network based on standardized hydrological basic data, designing a hybrid architecture including multi-layer long short-term memory and convolutional neural network, thereby obtaining generator network structure data; Step S172: generating a discriminator network structure of the adversarial network according to the standardized hydrological basic data design conditions, and performing multi-domain consistency verification processing using a multi-scale time series discrimination strategy to obtain discriminator network structure data, wherein the multi-scale time series discrimination strategy includes time domain and frequency domain feature extraction modules; Step S173: defining the loss function of the generative adversarial network according to the discriminator network structure data and the generator network structure data, thereby obtaining loss function data, wherein the loss function includes loss functions of distribution similarity loss, temporal consistency loss, and hydrological physical constraint loss; Step S174: Integrate the discriminator network structure data, the generator network structure data and the loss function data into a data generation model.

[0032] The embodiment of the present invention designs a hybrid architecture of the generator network based on standardized hydrological basic data. The generator network consists of two parts: a long short-term memory (LSTM) network module and a convolutional neural network (CNN) module. First, a two-layer stacked LSTM network is used to extract the time series features of the hydrological data. The number of hidden units of the LSTM is set to 128, and the input data is an hourly data segment (length of 168 hours) in the hydrological feature extended data; then, the high-dimensional features of the LSTM output are processed by a one-dimensional convolution layer, with a convolution kernel size of 3, a step size of 1, and an output channel number of 64; finally, after two layers of fully connected layers, prediction data with the same length as the target water level time series data is generated. In order to improve the stability of the generator network, batch normalization and ReLU activation functions are used, and the sigmoid activation function is used in the last layer to map the output to [0,1]. This structure can capture the long-term dependency and local features of the hydrological data in the generator network, thereby obtaining the generator network structure data. When designing the discriminator network structure, a multi-scale time series discrimination strategy is introduced to enhance the detection capability of data characteristics. The discriminator network consists of two modules: a time domain feature extraction module and a frequency domain feature extraction module. The time domain module consists of three layers of one-dimensional convolutional neural networks, with convolution kernel sizes of 3, 5 and 7, and a step size of 2, to extract time series features of different scales; the frequency domain module converts the time series into frequency spectrum features through fast Fourier transform (FFT), and then extracts the frequency domain features through a layer of convolutional network; the outputs of the two modules are fused through a fully connected layer, and finally a discrimination result is output through a sigmoid activation function, indicating whether the input data is real or generated. The input of the discriminator is a combination of the water level time series data output by the generator and the real hydrological data, and the consistency test of the time domain and the frequency domain is realized through multi-scale feature extraction, so as to obtain the discriminator network structure data. Based on the network structure of the generator and the discriminator, the loss function of the generative adversarial network is defined. The loss function consists of three parts: distribution similarity loss, temporal consistency loss and hydrological physical constraint loss. The distribution similarity loss calculates the difference between the distribution of generated data and real data through cross entropy; the temporal consistency loss uses the dynamic time warping (DTW) method to measure the degree of alignment between the generated data and the real data on the time axis; the hydrological physical constraint loss calculates the physical consistency of the generated data by constraining the physical meaning of the generated data (such as the upper and lower limits of the water level, the correlation between the fluctuation trend and rainfall) using the penalty term. The three parts of the loss are weighted and summed, with weights of 0.5, 0.3 and 0.2 respectively, and the generative adversarial network is optimized with the Adam optimization algorithm to obtain the loss function data. The discriminator network structure data, the generator network structure data and the loss function data are integrated into the conditional generative adversarial network model.In the specific operation, the parameters of the generator and discriminator networks are first initialized. The weights of the generator are initialized using Xavier, and the weights of the discriminator are randomly initialized using a normal distribution. Then, the generator and discriminator are adversarially trained through the loss function definition. In each round of training, the discriminator accepts real data and generated data for true and false discrimination, and the generator adjusts parameters to deceive the discriminator to the greatest extent. During the training process, a learning rate adjustment strategy is adopted, and the initial learning rate is set to 0.001, which decreases by 10% every 100 rounds. After 2000 rounds of training, the distribution error of the generative adversarial network model on the verification data is less than 5%, and the temporal consistency error is less than 3%, thus obtaining a hydrological data generation model with stable performance.

[0033] The present invention constructs a generator network structure of a conditional generative adversarial network (GAN) based on standardized hydrological basic data, and designs a hybrid architecture including multi-layer long short-term memory (LSTM) and convolutional neural network (CNN). This architecture combines the time series modeling capability of LSTM and the feature extraction advantage of CNN, and can effectively capture the temporal and spatial characteristics of hydrological data, thereby generating more accurate and close to real hydrological time series data. The design of the generator network structure provides a strong foundation for the efficient data generation of the conditional generative adversarial network, ensuring the effectiveness and generalization ability of the network in a complex hydrological change environment. By designing the discriminator network structure of the conditional generative adversarial network and introducing a multi-scale time series discrimination strategy for multi-domain consistency verification, the model's comprehensive analysis of data time series and frequency domain features is enhanced. The discrimination strategy includes time domain and frequency domain feature extraction modules, so that the discriminator can not only identify temporal consistency, but also evaluate the authenticity of generated data from the frequency level. The addition of this strategy enables the network to comprehensively consider the performance of data at different levels, thereby improving the accuracy and robustness of the discrimination, and further improving the data quality of the generator network. The loss function of the generative adversarial network is defined according to the network structure of the generator and the discriminator, including distribution similarity loss, temporal consistency loss and hydrological physical constraint loss. Through the design of these loss functions, the network can optimize the generated data during the training process, which not only ensures the distribution similarity between the generated data and the real data, but also ensures the consistency of the data in the time series, and strengthens the practical applicability of the generated data in combination with the hydrological physical constraints. This multi-dimensional loss function design ensures that the generated data meets the physical laws and temporal characteristics of the hydrological system while meeting the statistical characteristics, and improves the comprehensive effect of the model. The discriminator network structure data, the generator network structure data and the loss function data are integrated into a complete data generation model, completing the construction of the conditional generative adversarial network. The completion of this step means that the model has the ability to generate hydrological data from the input conditions. By optimizing the collaboration between the generator and the discriminator, it can effectively simulate and generate time series data that conforms to actual hydrological changes. This integrated generation model provides strong data support for subsequent hydrological prediction and decision-making, and has high application value and practical operability.

[0034] Preferably, step S2 comprises the following steps: Step S21: acquiring historical hydrological variable data, including temperature, humidity, precipitation, and atmospheric pressure information; performing feature dimensionality reduction processing on the historical hydrological variable data, and extracting key hydrological features, thereby obtaining key hydrological feature data; Step S22: designing a hydrological variable encoding network based on an attention mechanism based on the key hydrological characteristic data, and using the hydrological variable encoding network to perform nonlinear transformation on the key hydrological characteristic data, thereby obtaining conditional input characteristic data; Step S23: According to the conditional input feature data, the generator network of the data generation model is tuned for hydrological adaptability parameters, thereby obtaining hydrological adaptability optimization model parameters, wherein the hydrological adaptability parameter tuning is specifically to design a generation network parameter adaptive optimization strategy based on variational inference; Step S24: optimizing the generator network of the data generation model according to the hydrological adaptability optimization model parameters, thereby obtaining a data generation optimization model; Step S25: Inject the conditional input feature data into the generator network of the data generation optimization model, and generate preliminary water level time series data through adversarial training of the generator and discriminator of the generative adversarial network.

[0035] As an embodiment of the present invention, refer to Figure 3 As shown, Figure 1 Detailed step flow diagram of step S2 in the embodiment of the present invention, step S2 includes the following steps: Step S21: acquiring historical hydrological variable data, including temperature, humidity, precipitation, and atmospheric pressure information; performing feature dimensionality reduction processing on the historical hydrological variable data, and extracting key hydrological features, thereby obtaining key hydrological feature data; The embodiment of the present invention extracts hydrological variable data for the past 30 years from the hydrological observation database, including temperature, humidity, precipitation and atmospheric pressure information recorded daily, and the data resolution is the daily average. First, principal component analysis (PCA) is used to reduce the dimensionality of the hydrological variable data, reducing the original 4-dimensional features to 2 dimensions, where the cumulative variance contribution rate of the principal component reaches 95%, reducing redundant information. Then, the correlation between the reduced-dimensional features and the target water level changes is evaluated by the mutual information method, and key hydrological features are extracted, including the annual average temperature change range and the seasonal variation value of precipitation. These key features are particularly important in drought and flood risk analysis, thereby obtaining key hydrological feature data.

[0036] Step S22: designing a hydrological variable encoding network based on an attention mechanism based on the key hydrological characteristic data, and using the hydrological variable encoding network to perform nonlinear transformation on the key hydrological characteristic data, thereby obtaining conditional input characteristic data; The embodiment of the present invention designs a hydrological variable encoding network based on the attention mechanism based on key hydrological feature data. The encoding network adopts a two-layer Transformer structure, inputs key hydrological feature data, and captures feature dependencies at different time steps through a self-attention mechanism. In the specific implementation, the number of multi-head attention units in the first Transformer encoding layer is set to 8, and the hidden layer dimension is 64; the number of attention units in the second Transformer layer is set to 4, and the hidden layer dimension is 32, and residual connections and layer normalization operations are added to enhance training stability. Finally, the network embeds the nonlinearly transformed hydrological features into the 32-dimensional conditional input feature vector space through a fully connected layer and a ReLU activation function, thereby obtaining conditional input feature data.

[0037] Step S23: According to the conditional input feature data, the generator network of the data generation model is tuned for hydrological adaptability parameters, thereby obtaining hydrological adaptability optimization model parameters, wherein the hydrological adaptability parameter tuning is specifically to design a generation network parameter adaptive optimization strategy based on variational inference; The embodiment of the present invention uses the conditional input feature data generated in step S22 to tune the hydrological adaptability parameters of the generator network. A generative network adaptive optimization strategy based on variational inference is designed. First, the conditional input feature data is defined as the conditional input of the generator network. The conditional vector distribution is converted into an approximate normal distribution through a variational autoencoder (VAE), and the KL divergence loss between distributions is defined. Secondly, the hidden layer weight initialization of the generator is dynamically adjusted during the optimization process, and the Adam optimizer with a learning rate of 0.001 is used. The training batch is set to 64 and 5000 iterations are performed. Through this method, the generator network can more effectively learn the impact of hydrological variables on hydrological changes, thereby obtaining the parameters of the hydrological adaptability optimization model.

[0038] Step S24: optimizing the generator network of the data generation model according to the hydrological adaptability optimization model parameters, thereby obtaining a data generation optimization model; The embodiment of the present invention optimizes the model parameters according to the hydrological adaptability and further optimizes the model of the generator network. In the optimization, the discriminator parameters are first frozen and the generator is trained separately so that the water level time series data generated by it is more in line with the actual hydrological change characteristics. A periodic learning rate strategy is adopted, the initial learning rate is set to 0.0005, and it is adjusted to 0.0001 after 2000 rounds of training, and L2 regularization is added to prevent overfitting. During the optimization process, the distribution differences and hydrological correlations of the generated data are monitored through the validation set until the distribution error between the generated water level data and the historical real data is less than 3%, and finally a data generation model with optimized performance is obtained.

[0039] Step S25: Inject the conditional input feature data into the generator network of the data generation optimization model, and generate preliminary water level time series data through adversarial training of the generator and discriminator of the generative adversarial network.

[0040] The embodiment of the present invention injects the conditional input feature data in step S22 into the optimized generator network to perform adversarial training of the generative adversarial network. In the specific training, the generator generates preliminary water level time series data according to the conditional input features, and inputs the data together with the real water level time series data into the discriminator for true and false discrimination. The goal of generator optimization is to minimize the discrimination accuracy of the discriminator, and the optimization goal of the discriminator is to improve the discrimination accuracy. The small batch stochastic gradient descent method (batch size=32) is used in training. Each time the generator is updated for 1 round of training, the discriminator is trained for 2 rounds, and after 3000 rounds of continuous training, the preliminary water level time series data generated by the generator has a high consistency with the real data in time distribution, and the preliminary water level time series data is finally obtained.

[0041] The present invention effectively extracts key hydrological characteristic data such as temperature, humidity, precipitation and atmospheric pressure by acquiring historical hydrological variable data and performing feature dimension reduction processing. This process can reduce the interference of redundant information while retaining hydrological information that is crucial to the generation of water level time series data. Through dimension reduction and feature extraction, the data becomes more concise and efficient, thereby improving the speed and accuracy of subsequent model training. Key hydrological characteristic data provides the necessary information basis for hydrological adaptability optimization, ensuring that subsequent steps can be more effectively modeled and optimized based on refined hydrological characteristics. Using the hydrological variable encoding network design based on the attention mechanism, nonlinear transformation is performed on key hydrological characteristic data to obtain conditional input feature data. The attention mechanism helps the model focus on hydrological factors that have a greater impact on water level prediction by weighting the influence of different hydrological variables, thereby improving the sensitivity and accuracy of the model to hydrological changes. Nonlinear transformation further enhances the network's modeling ability for complex hydrological patterns, ensures that the generated conditional input feature data can fully reflect the impact of hydrological changes on water level time series, and improves the hydrological adaptability of the data generation model. The hydrological adaptability parameters of the generator network of the data generation model are tuned by conditional input feature data, and an adaptive optimization strategy for the parameters of the generator network based on variational inference is designed. This tuning process enables the generator to better adapt to the data generation needs under different hydrological conditions. The parameters are optimized by the variational inference method, which improves the generalization ability of the network. The hydrological adaptability optimization ensures that the generator can generate water level time series data that meet the actual conditions under various hydrological change scenarios, thereby improving the hydrological adaptability and reliability of the generation model. After obtaining the parameters of the hydrological adaptability optimization model, the generator network of the data generation model is optimized. This optimization further improves the performance of the generator network, enabling it to generate more accurate water level time series data that is highly consistent with the actual hydrological conditions. Model optimization ensures that the generator can stably and efficiently generate expected water level data in the face of various hydrological scenarios, thereby providing reliable data support for water level prediction and management. Preliminary water level time series data are generated by injecting conditional input feature data into the optimized generator network and through adversarial training of the generator and discriminator of the generative adversarial network. This adversarial training process ensures that the generated data not only conforms to the changing trend of hydrological characteristics, but also can be verified by the discriminator for its authenticity and consistency. Through this interaction between the generator and the discriminator, the water level time series data is further optimized and verified, providing a high-quality data foundation for subsequent water level prediction and hydrological change analysis.

[0042] Preferably, step S25 comprises the following steps: Step S251: using the conditional input feature data as conditional vector input data to generate the initial layer of the generator network in the optimization model, thereby constructing conditionally guided generator network input data; Step S252: performing multi-layer nonlinear transformation on the generator network input data through the generator network in the data generation optimization model, thereby generating candidate water level time series data; Step S253: inputting the candidate water level time series data into the discriminator network in the data generation optimization model, and performing authenticity discrimination and adversarial training according to the preset hydrological basic data, thereby obtaining adversarial training feedback data; Step S254: adaptively adjusting the parameters of the generator network in the data generation optimization model according to the adversarial training feedback data of the discriminator network, and optimizing the quality of the candidate water level time series data, thereby obtaining generator iterative optimization data; Step S255: Loop through steps S252 to S254 until the generated generator iterative optimization data meets the preset adversarial training convergence criteria, thereby obtaining preliminary water level time series data.

[0043] The embodiment of the present invention inputs the conditional input feature data generated in step S22 as the conditional vector into the initial layer of the generator network to construct condition-guided generator network input data. In a specific implementation, the conditional vector is a 32-dimensional feature embedding vector, which is expanded to the input dimension 256 of the generator network using a linear mapping layer. The linear mapping layer is followed by a Batch Normalization layer and a ReLU activation function to ensure the stability of the distribution of the input data, and to form the input data tensor of the generator network by splicing a random noise vector (normal distribution with a mean of 0, a standard deviation of 1, and a dimension of 128) with the conditional input feature. The input data tensor generated in step S251 is input into the generator network for multi-layer nonlinear transformation to generate candidate water level time series data. The generator network adopts a hybrid architecture that integrates a long short-term memory network (LSTM) and a fully connected layer. First, the long-term dependency features in the time series are extracted through three layers of LSTM units, and the number of hidden units in each layer of LSTM is 128, 64 and 32 respectively; then, the nonlinear mapping and compression of the data dimension are completed through two layers of fully connected layers, and the water level time series data is output, and the time series length is set to 365 days (daily average water level). During the network training process, the Adam optimizer is used, the learning rate is set to 0.0001, and the initial quality of the candidate water level time series data is dynamically adjusted based on the generator parameters. The candidate water level time series data generated in step S252 is input into the discriminator network and compared with the preset historical hydrological basic data for authenticity. The discriminator network adopts a convolutional neural network (CNN) architecture, which includes three one-dimensional convolutional layers and two fully connected layers, with convolution kernel sizes of 5, 3 and 3 respectively, and the number of channels is 64, 128 and 256 respectively. Finally, the network outputs the true and false discrimination probability through the Softmax activation function, and the frequency domain transformation module is combined to perform additional inspections on the spectral distribution characteristics of the candidate data. The loss function for authenticity discrimination is binary cross entropy, and adversarial training feedback data is generated in combination with the frequency domain consistency discrimination result to guide the improvement of the generator network. According to the adversarial training feedback data obtained in step S253, the parameters of the generator network are adaptively adjusted. The specific adjustment method is to optimize the generator network weights based on the gradient back propagation of adversarial loss (including generator loss and discriminator loss) and time distribution consistency loss. At the same time, the output of the candidate water level time series data is post-processed, high-frequency artifacts are eliminated by a low-pass filter, and linear interpolation correction is performed on the abnormal points. After each adjustment, the distribution difference between the data output by the generator and the real data is gradually reduced, and the optimized data is called the generator iterative optimization data. Repeat steps S252 to S254 to perform multiple rounds of adversarial training and candidate data optimization on the generator network until the preset adversarial training convergence criteria are met.The specific convergence criteria are set as follows: 1) the distribution difference between the output data of the generator and the real hydrological data is less than 2%; 2) the adversarial loss of the generator changes by less than 0.01 in 100 consecutive iterations; 3) the statistical characteristics of the generated water level time series data in the time domain and frequency domain are consistent with the historical data. Finally, when the above criteria are met, the iterative optimization data output by the generator is the high-quality preliminary water level time series data.

[0044] The present invention inputs conditional input feature data as a conditional vector into the initial layer of the generator network of the data generation optimization model, thereby constructing condition-guided generator network input data. This process ensures that the generator network can guide the generation of water level time series data based on hydrological characteristics and other relevant conditions. The conditional input feature data provides the generator with important information about external environmental conditions, so that the generated water level time series data can accurately reflect the changing trend of hydrological variables, increasing the accuracy and pertinence of data generation. The input data is subjected to multi-layer nonlinear transformation by the generator network, thereby generating candidate water level time series data. Multi-layer nonlinear transformation can capture the complex relationships and patterns in the input data, thereby generating more accurate and rich candidate water level time series data. This transformation process makes the generated data not only meet the hydrological conditions, but also have a high degree of detail and diversity, which helps to better simulate the dynamic changes of water level time series and improve the quality and availability of generated data. The candidate water level time series data is input into the discriminator network, and authenticity discrimination and adversarial training are performed based on the preset hydrological basic data, thereby obtaining adversarial training feedback data. Through the evaluation and adversarial training of the discriminator, the generated data has been verified for authenticity to ensure that it is consistent with the actual hydrological conditions. The adversarial process between the discriminator and the generator prompts the generator to continuously improve, the quality of the generated data is gradually improved, and the output of unreal or inappropriate data is reduced, thereby enhancing the robustness and credibility of the model. According to the adversarial training feedback data of the discriminator network, the parameters of the generator network are adaptively adjusted, and the quality of the candidate water level time series data is optimized to obtain the generator iterative optimization data. This feedback mechanism allows the generator network to continuously adjust its parameters according to the opinions of the discriminator and optimize the data generation process. Through repeated iterative optimization, the generator can generate water level time series data that is more in line with the actual situation and has higher quality. This process helps to improve the accuracy and consistency of the generated data, and finally obtains a more expected output. By cyclically executing steps S252 to S254, until the generated generator iterative optimization data meets the preset adversarial training convergence criteria, preliminary water level time series data is obtained. Through repeated iterations and optimization, the generator continuously improves its output, and the water level time series data finally generated meets the set convergence standard and has a high degree of accuracy and reliability. This process ensures that the generated water level time series data can provide effective support in practical applications, accurately reflect the changing trends of hydrological conditions, and ultimately meet the needs and expectations of the model.

[0045] Preferably, step S3 comprises the following steps: Step S31: Designing a stratified sampling weight strategy according to the distribution characteristics and statistical attributes of the preliminary water level time series data, thereby obtaining sampling weight strategy data; Step S32: performing stratified importance sampling processing on the preliminary water level time series data according to the sampling weight strategy data, thereby obtaining importance stratified sampling data; Step S33: setting the differentiated generation weights of the sampling probabilities of the prominent extreme hydrological events based on the preset rare extreme hydrological event characteristics, thereby obtaining differentiated generation weight data; Step S34: performing weighted sampling processing of different types of hydrological events on the extreme hydrological event data according to the differentiated generated weight data, thereby obtaining optimized sampling data of hydrological events; Step S35: oversampling rare extreme hydrological event data by a bootstrap method, thereby obtaining high-quality extreme event data; Step S36: Perform distribution consistency evaluation based on statistical characteristics according to importance stratified sampling data, extreme event data and hydrological event optimized sampling data to obtain water level time series data.

[0046] The embodiment of the present invention analyzes the distribution characteristics and statistical properties of the preliminary water level time series data. First, the overall distribution form of the data is described by calculating statistical indicators such as mean, standard deviation, skewness, and kurtosis. Then, combined with the time series properties of the data, the change rate and frequency characteristics under different time windows are calculated, with special attention paid to the sudden increase and decrease areas of the water level. According to these characteristics, the preliminary water level time series data is divided into three layers: low water level interval, medium water level interval, and high water level interval. The sample size weight of each layer is preliminarily allocated according to its proportion in the overall data, and the sampling weight of the high water level interval is increased to capture potential extreme events. The designed stratified sampling weight strategy is output in the form of a weighted coefficient matrix, in which each row corresponds to the weight of an interval. According to the sampling weight strategy data of step S31, stratified importance sampling is implemented for the preliminary water level time series data. The specific method is to resample the sample distribution of each layer interval according to the weight matrix, adopt a sampling strategy with replacement during the resampling process, and set different sampling probabilities according to the importance weight of each layer. For example, the weight of the low water level interval is set to 0.3, the medium water level interval is 0.5, and the high water level interval is 0.7. The sampling process is implemented through Python, and the random.choice method of the Numpy library is called to generate stratified samples. Finally, the importance stratified sampling data is output. The sample size is consistent with the original data, but the distribution pays more attention to the characteristics of the high water level interval. Combined with the preset rare extreme hydrological event characteristics, such as high water level events with peak water levels exceeding the 90% quantile or duration exceeding 7 days, the sampling probability is adjusted differentially. By setting the threshold range of extreme event characteristics, such as water level ≥5 meters and duration ≥3 days, the generation weight of qualified event samples is increased to 1.5 times the original weight, and the adjacent time windows are expanded to increase the sampling probability of their associated samples. The differentiated generation weight is realized through a weighting function, and the weight function form is: ,in is the adjustment factor, is the distance attenuation coefficient, is the time distance between the sample and the extreme event, and the generated differentiated weight is output in the form of a vector. The differentiated weight data generated in step S33 is applied to the sampling process of extreme hydrological event data, and the sampling probability of different types of events is optimized separately by setting the weight coefficients of different types of hydrological events (such as short-term floods, long-term waterlogging, etc.). For example, the weight of short-term flood events is set to 1.2, the weight of long-term waterlogging events is set to 1.5, and the extreme events are weighted sampled by random weighted sampling method to ensure that the distribution ratio of each type of event in the optimized sampling data meets the set target. The sampling process is completed based on the resample function of Sklearn, and finally the optimized sampling data of hydrological events covering multiple types of events and focusing on extreme events are obtained. The rare extreme hydrological event data is oversampled by the self-help method. First, the extreme event sample set is repeatedly sampled. The specific operation of the self-help method is: randomly extract samples from the original data set to form a new data set, and the samples are put back after each sampling to ensure that each sample may be selected multiple times. The oversampling ratio is set to 2 times, that is, the final number of samples is twice the original rare event data set. The random.choices method of Python is used for sampling, and the oversampled data is deduplicated and time-sorted to ensure that the expanded data set maintains temporal consistency, and finally obtains high-quality extreme event data. After merging the importance stratified sampling data of step S32, the hydrological event optimization sampling data of step S34, and the extreme event data of step S35, the distribution consistency is evaluated based on statistical characteristics. First, the distribution fitting analysis is performed on the merged data, and the overlap rate between its frequency distribution histogram and the original water level time series data is calculated; then the distribution consistency of the two is evaluated by the Kolmogorov-Smirnov (KS) test, and the distribution proportion of rare events is independently counted to ensure that the proportion of rare events in the optimized data is increased by at least 15%. During the evaluation process, Matplotlib is used to draw a distribution comparison chart, and the water level time series data that finally meets the consistency requirements is output for subsequent processing.

[0047] The present invention designs a stratified sampling weight strategy according to the distribution characteristics and statistical attributes of the preliminary water level time series data, thereby obtaining sampling weight strategy data. This strategy design ensures that the sampling weights of different water level time series data can be optimized according to the distribution characteristics of the data, especially for the deviation or imbalance that may exist in the data. This strategy provides a reasonable framework for subsequent sampling processing, so that the overall picture of the water level time series data can be better represented in the subsequent process, and appropriate weights are given to different data points, thereby improving the representativeness and reliability of the data. According to the sampling weight strategy data, the preliminary water level time series data is subjected to stratified importance sampling processing, thereby obtaining importance stratified sampling data. Through stratified importance sampling, it can be ensured that the relative importance of different data points is accurately reflected during the processing process, especially in the water level time series data, the data under certain time periods or conditions are more critical. This step reduces the risk of over-representation or under-representation of certain time period data by optimizing the sampling process, thereby improving the balance and accuracy of the data. Based on the preset rare extreme hydrological event characteristics, the sampling probability of the extreme hydrological event is highlighted, and the differentiated generation weight data is obtained. This differentiated weight setting focuses on extreme hydrological events, ensuring that these rare but important events can be given proper attention during the generation process. In this way, the model's predictive ability in dealing with extreme events can be effectively enhanced, ensuring that the generation and reflection of extreme events have sufficient accuracy and reliability. According to the differentiated generation weight data, the extreme hydrological event data is subjected to weighted sampling of different types of hydrological events, thereby obtaining optimized sampling data for hydrological events. This process further optimizes the sampling of hydrological events, so that the frequency and importance of extreme hydrological events are reasonably enhanced during the generation process. Through weighted sampling, it can be ensured that extreme events are appropriately expressed in the generated data, making the generated data more consistent with the actual occurrence law of hydrological events, thereby improving the generalization ability of the model. Rare extreme hydrological event data are oversampled by the bootstrap method to obtain high-quality extreme event data. The bootstrap method is a commonly used oversampling method that can effectively balance the sample size of rare events in the data, ensuring that these events will not affect the training effect of the model due to insufficient data. Through this method, more reliable extreme hydrological event data can be obtained, thereby improving the accuracy and stability of the generated model when dealing with extreme situations. The water level time series data is obtained by performing a distribution consistency assessment based on statistical characteristics based on importance stratified sampling data, extreme event data, and hydrological event optimization sampling data. The distribution consistency assessment based on statistical characteristics can ensure that the final generated water level time series data not only conforms to the distribution law of actual data in terms of statistical characteristics, but also accurately reflects the processing effects of importance sampling and extreme events.This step helps to verify the quality of the generated data and ensure that the generated data has high consistency and availability, thus providing a solid foundation for subsequent model applications.

[0048] Preferably, step S36 includes the following steps: Step S361: extracting statistical features from importance stratified sampling data, extreme event data, and hydrological event optimized sampling data, and constructing a multidimensional statistical feature vector; Step S362: designing a distribution distance calculation method based on statistical characteristics according to the multidimensional statistical feature vector, and performing a distribution similarity test based on Kolmogorov-Smirnov, thereby obtaining distribution consistency measurement data; Step S363: According to the distribution consistency measurement data, the importance stratified sampling data, the extreme event data and the hydrological event optimized sampling data are subjected to distribution feature reconciliation processing based on feature interpolation and smoothing technology for different data sets, thereby obtaining aligned water level time series data.

[0049] The embodiment of the present invention first extracts statistical features from importance stratified sampling data, extreme event data, and hydrological event optimized sampling data, respectively. The extracted features include mean, variance, kurtosis, skewness, time series autocorrelation, and extreme value ratio. The specific operation is to calculate the features of each data set one by one. For example, the extreme value ratio is obtained by counting the sample ratio of water levels exceeding the 95% quantile; the time series autocorrelation is obtained by calculating the lag correlation coefficient of the autocorrelation function. The extracted statistical features of each dimension are stored in the form of feature vectors, and the data of each sample is represented as a multidimensional vector, which contains the above-mentioned statistical eigenvalues. In order to ensure the consistency of features of different data sets, the feature vectors are dimensionlessly processed by a standardized method, and finally a multidimensional statistical feature vector set consisting of three groups of feature vectors is generated. Based on the multidimensional statistical feature vectors generated in step S361, a distribution distance calculation method is designed. The method is specifically to measure the similarity of the distribution of different data sets through the Kolmogorov-Smirnov (KS) test. The implementation steps of the KS test are as follows: first, construct a cumulative distribution function (CDF) for the characteristic distribution of each set of data, for example, construct a CDF curve in the range of [0,1] for the characteristic of extreme value ratio; then calculate the maximum vertical distance between the corresponding CDF curves of each pair of data sets as the KS statistic. According to a preset similarity threshold, such as 0.05, determine whether the distribution is significantly different, and record the distribution distance of each set of data at the same time, and store the distribution consistency measurement data in the form of a matrix, where the rows and columns of the matrix represent different data sets, and the element values ​​are KS statistics. According to the distribution consistency measurement data obtained in step S362, the importance stratified sampling data, extreme event data, and hydrological event optimized sampling data are reconciled with distribution characteristics. The specific method is: for data sets with significant distribution differences, a feature correction technique based on linear interpolation is used to bring the data characteristics of fewer samples closer to the data characteristics of more samples. For example, for the kurtosis characteristics of extreme event data, the interpolation function is used. ,in and are the characteristic values ​​of the two data sets, is the interpolation weight (value range is 0.2 to 0.8). At the same time, in order to eliminate local fluctuations, cubic spline interpolation is used to smooth the time series features. The feature distribution of each data set is adjusted by the above method so that its KS statistic is within the preset similarity threshold (for example, 0.03), and finally the water level time series data with consistent and aligned features are generated.

[0050] The present invention extracts statistical features from importance stratified sampling data, extreme event data, and hydrological event optimized sampling data, and constructs a multidimensional statistical feature vector. The main purpose of this process is to conduct in-depth analysis of various types of data, extract their respective key statistical characteristics, and convert these characteristics into multidimensional feature vectors. These feature vectors can fully reflect the inherent laws of the data, including the central tendency, discreteness, and distribution characteristics of the data, provide necessary statistical information support for subsequent analysis and processing, and ensure that the multidimensional characteristics of the data are fully captured. According to the multidimensional statistical feature vector, a distribution distance calculation method based on statistical characteristics is designed, and a distribution similarity test based on Kolmogorov-Smirnov is performed to obtain distribution consistency measurement data. This step can effectively measure the distribution consistency between different data sets by constructing an appropriate distribution distance calculation method and combining the Kolmogorov-Smirnov test for distribution similarity evaluation. The Kolmogorov-Smirnov test is a classic statistical method that can objectively evaluate the distribution differences of data sets and provide a quantitative basis for further data adjustment. This step helps ensure the consistency of all sampled data in statistical characteristics, laying a solid foundation for subsequent data fusion and modeling. According to the distribution consistency measurement data, the distribution characteristics of different data sets of importance stratified sampling data, extreme event data and hydrological event optimized sampling data are reconciled based on feature interpolation and smoothing technology to obtain aligned water level time series data. This step solves the differences in distribution characteristics of different data sets through feature interpolation and smoothing technology, so that these data can be fused and aligned under the same framework. By reconciling the distribution characteristics of different data sets, potential biases and noise can be eliminated to ensure that the statistical attributes and distribution characteristics of water level time series data remain consistent. This process greatly improves the quality and stability of the data, so that the generated water level time series data can better reflect the actual hydrological conditions, thereby improving the prediction ability and accuracy of the model.

[0051] Preferably, step S4 comprises the following steps: Step S41: verifying the spectrum characteristics of the water level time series data to generate spectrum characteristics verification data, wherein the spectrum characteristics verification specifically uses wavelet transform to decompose multi-scale time series characteristics, analyzes the energy distribution of different time frequency segments, and evaluates the consistency of the spectrum characteristics of the water level time series data with the pre-acquired real water level data; Step S42: analyzing the correlation and information transfer characteristics between different variables of the water level time series data based on the mutual information measurement method, thereby obtaining mutual information enhanced verification data; Step S43: merging the spectrum characteristic verification data and the mutual information enhancement verification data into water level time series verification data; Step S44: Eliminate artifacts in the water level time series verification data based on an abnormality recognition algorithm, thereby obtaining artifact elimination optimized data; Step S45: Perform extreme value abnormal point correction processing based on physical constraints on the artifact removal optimization data to obtain supervised time series water level data.

[0052] The embodiment of the present invention verifies the spectral characteristics of water level time series data. The specific method is to use wavelet transform to perform multi-scale decomposition on water level time series data. By selecting an appropriate wavelet basis (such as Morlet wavelet), the water level time series data is decomposed into multiple frequency sub-bands. Then, the energy distribution in each frequency sub-band is calculated, and the spectral characteristics of different time frequency bands are analyzed. For each sub-band, the ratio of its energy to the total energy is calculated, and the consistency between the energy distribution of the water level time series data in each frequency band and the pre-acquired real water level data is analyzed by comparison. The spectral characteristic consistency evaluation uses the mean square error (MSE) or the Pearson Correlation Coefficient (PCC) to quantify the spectral similarity of two data sets, and then generate spectral characteristic verification data. Based on the mutual information measurement method, the correlation and information transfer characteristics between different variables of the water level time series data are analyzed. First, the information transfer relationship between different variables is evaluated by calculating the mutual information between each variable in the water level time series data. For example, the Shannon mutual information formula is used to calculate the mutual information between each pair of variables. By analyzing the mutual information between multiple variables such as water level data, rainfall, and flow, the dependency relationship between different hydrological characteristics can be revealed. The mutual information of each pair of variables is sorted, and the variable pairs with high mutual information are selected as the core variables of the verification data. Finally, the mutual information enhanced verification data is generated, and the potential correlation patterns and characteristics in the water level time series data are further analyzed through this data. The spectrum characteristic verification data and the mutual information enhanced verification data are merged into the water level time series verification data. First, the two sets of data are standardized to ensure that their dimensions are consistent, and then the weighted average method is used to fuse them. For example, the spectrum characteristic verification data can be given a weight of 0.6, and the mutual information enhanced verification data can be given a weight of 0.4, and then the two can be weighted and summed to obtain a comprehensive verification data set. The merged water level time series verification data contains information on spectral consistency and correlation between variables, which can provide a comprehensive verification basis for subsequent artifact removal and outlier correction. The water level time series verification data is artifact-free using an anomaly recognition algorithm. First, anomalies in the water level time series verification data are identified using algorithms such as Isolation Forest or Local Outlier Factor (LOF). These anomalies may be artifacts caused by measurement errors, equipment failures, or other abnormal factors. The algorithm automatically identifies anomalies that deviate from the normal trend by detecting local density deviations in the data. For these anomalies, they are repaired by interpolation or smoothing methods. For example, sample-based interpolation methods (such as spline interpolation) are used to fill in the anomalies, or weighted average methods are used to smooth fluctuations in the data to obtain artifact-removed optimized data. The artifact-removed optimized data is corrected for extreme outliers based on physical constraints.First, by defining physical constraints, such as the maximum and minimum water levels, the reasonable range of flow and precipitation, etc., the data is screened in combination with actual hydrological conditions. For extreme points that exceed these physical constraints, adjustments are made through correction rule-based processing methods. For example, the maximum and minimum constraint method or other correction algorithms are used to adjust the extreme value data to a reasonable range. In addition, for some extreme abnormal points of mutation type, data smoothing techniques (such as weighted moving average method) can be used for further correction to ensure the smoothness and physical rationality of the water level time series data, and finally obtain the supervised time series water level data.

[0053] The present invention verifies the spectrum characteristics of water level time series data, decomposes multi-scale time series characteristics by wavelet transform, analyzes the energy distribution of different time frequency segments, and evaluates the consistency of spectrum characteristics of water level time series data and pre-acquired real water level data. Through spectrum analysis, the change characteristics of water level time series data can be deeply understood from different frequency scales, revealing potential periodic patterns or abnormal fluctuations. Wavelet transform can effectively capture the local characteristics of water level time series data, so that spectrum analysis can not only identify the overall trend, but also handle local anomalies, thereby improving the accuracy and reliability of water level data. The comparison of spectrum characteristic consistency evaluation with real data can ensure that the generated data conforms to the physical law of actual water level changes. Based on the mutual information measurement method, the correlation and information transfer characteristics between different variables of water level time series data are analyzed to obtain mutual information enhanced verification data. Through mutual information measurement analysis, the nonlinear relationship and mutual influence between variables can be revealed, thereby providing a more comprehensive correlation analysis for water level time series data. The advantage of this analysis method is that it can capture the complex nonlinearity and dependency between variables, which is more flexible and accurate than traditional linear methods. Enhanced verification data can effectively improve the accuracy of the model, especially when dealing with complex multivariate relationships, which helps to generate time series data that is more in line with the actual situation. The spectral characteristic verification data and the mutual information enhancement verification data are merged into the water level time series verification data. By integrating these two types of verification data, the analysis results at the frequency domain and information transmission levels can be combined, thereby comprehensively improving the verification validity of the data. Spectral characteristic verification and mutual information enhancement verification provide verification information from different angles. The combination of the two makes the final water level time series data verification more comprehensive and reliable, and can effectively eliminate the blind spots of a single verification method. Based on the anomaly recognition algorithm, the artifacts in the water level time series verification data are eliminated to obtain artifact removal optimization data. Artifacts are data anomalies introduced by factors such as noise and measurement errors, which may have adverse effects on water level time series data analysis and model training. Through the anomaly recognition algorithm, these artifacts can be automatically identified and removed to ensure that the water level data is more real and accurate, and to avoid the misleading influence of artifacts on the results. The artifact removal optimization data is corrected based on physical constraints for extreme value anomaly points, thereby obtaining supervised time series water level data. Extreme value anomalies are usually deviations caused by measurement errors or abnormal events. Correcting these extreme values ​​can significantly improve the reliability of water level time series data. By introducing physical constraints, such as the physical range and change law of water level, it can ensure that the corrected data conforms to the actual hydrological conditions and avoid the adverse effects of extreme values ​​on the model, thereby obtaining more accurate and reliable water level time series data. This correction process improves the quality of the data and ensures that the generated water level time series data can effectively reflect the real hydrological changes.

[0054] The present invention also provides a supervised time series water level data generation system, which is used to execute the above-mentioned supervised time series water level data generation method, and the supervised time series water level data generation system comprises: Model building module, used to obtain multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on hydrological monitoring data to generate standardized hydrological basic data; build a data generation model of conditional generative adversarial network based on standardized hydrological basic data; The hydrological condition enhancement module is used to regulate the conditional characteristics driven by hydrological variables in the data generation model, optimize the parameters of the generation network, and generate conditional input characteristic data with enhanced hydrological adaptability; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; The sampling optimization module is used to perform stratified importance sampling based on preliminary water level time series data, and to perform differentiated generation weight settings based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; to oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; to perform distribution consistency evaluation based on statistical characteristics based on extreme event data and hydrological event optimized sampling data to obtain water level time series data; The data verification and correction module is used to perform multi-scale verification on the water level time series data to obtain water level time series verification data, where the multi-scale verification includes spectral characteristic verification and mutual information measurement; the water level time series verification data is used to eliminate artifacts in the data and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

[0055] The model building module in the present invention provides a solid data foundation for the subsequent generation of water level time series data by acquiring multi-source hydrological monitoring data. The module first cleans the original data through outlier identification and data preprocessing, removes noise and errors that do not conform to the normal hydrological change law, and generates standardized hydrological basic data. This standardization process not only improves the consistency and availability of the data, but also ensures that the subsequent modeling will not be disturbed by abnormal data. Based on these standardized hydrological basic data, a data generation model of a conditional generative adversarial network (GAN) is constructed, which can simulate the complex laws of water level changes and generate time series data that conform to actual hydrological characteristics, laying the foundation for the accuracy and reliability of the entire model. The hydrological condition enhancement module plays a key role in generating water level time series data. The module can optimize the parameters of the generation network according to the actual situation of hydrological changes by regulating the conditional characteristics driven by hydrological variables on the data generation model, so that the generated data is enhanced in hydrological adaptability. This characteristic data with enhanced hydrological adaptability ensures that the generated water level time series data can more realistically reflect the hydrological changes under different hydrological conditions and provide more accurate prediction results. The sampling optimization module optimizes the data generation process so that the generated water level time series data is more in line with the actual application requirements. By performing stratified importance sampling on the preliminary water level time series data, the representativeness of the generated data can be improved. Especially in the face of rare extreme hydrological events, these events can be sampled preferentially through differentiated generation weight settings, thereby generating more accurate extreme event data. Oversampling by the self-help method can further improve the quality of extreme event data, so that the model's ability to identify these key events is enhanced. Finally, these data are processed by distribution consistency evaluation based on statistical characteristics to ensure the accuracy and consistency of water level time series data in different scenarios. The data verification and correction module comprehensively checks the generated water level time series data through a multi-scale verification method to ensure the reliability of the data. Spectral characteristic verification and mutual information measurement provide two important perspectives for the verification process. Spectral characteristic verification can check the energy distribution characteristics of data in different frequency ranges, and mutual information measurement reveals the correlation and information flow between data, which helps to detect whether the data meets the expected hydrological characteristics. Artifact elimination and extreme value outlier correction further improve the quality of the data and avoid inaccurate results caused by measurement errors or noise. This process ensures that the generated water level time series data not only conforms to the actual hydrological laws, but also has sufficient physical rationality and accuracy, and finally obtains high-quality supervised time series water level data. In summary, these modules cooperate with each other to provide a complete process from data acquisition, generation, optimization to verification and correction, ensuring the accuracy, reliability and applicability of water level time series data, and providing strong data support for applications such as hydrological prediction and risk assessment.

[0056] The present invention also provides a computer-readable storage medium storing a computer program, wherein the computer program is executed to implement the supervised time-series water level data generation method as described above.

[0057] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0058] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A supervised method for generating time series water level data, characterized in that: The following steps are involved: Step S1: Acquire multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on the hydrological monitoring data to generate standardized hydrological basic data; construct a data generation model of a conditional generative adversarial network based on the standardized hydrological basic data; Step S2: The data generation model is regulated by the conditional characteristics driven by the hydrological variables, the parameters of the generation network are optimized, and the conditional input characteristic data with enhanced hydrological adaptability are generated; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; Step S3: Perform stratified importance sampling based on the preliminary water level time series data, and perform differentiated generation weight setting based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; The rare extreme hydrological event data are oversampled by the bootstrap method to obtain high-quality extreme event data; the distribution consistency evaluation based on statistical characteristics is performed on the extreme event data and the optimized sampling data of hydrological events to obtain water level time series data; Step S4: Perform multi-scale verification on the water level time series data to obtain water level time series verification data, wherein the multi-scale verification includes spectral characteristic verification and mutual information measurement; perform artifact elimination on the water level time series verification data, and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

2. The supervised time series water level data generation method according to claim 1, characterized in that: Step S1 includes the following steps: Step S11: collecting water level, rainfall, evaporation and flow information to obtain original hydrological monitoring data; Step S12: performing format conversion and time stamp calibration processing on the original hydrological monitoring data, thereby obtaining time-aligned multi-source hydrological monitoring data; Step S13: performing outlier identification based on statistical methods on the multi-source hydrological monitoring data, and performing interpolation repair processing and outlier removal processing to obtain hydrological monitoring filtered data; Step S14: performing data filling processing on the hydrological monitoring filtered data based on the time series regression model of adjacent data points, thereby obtaining hydrological monitoring cleansing data; Step S15: screening the hydrological monitoring cleaning data for characteristic variables highly correlated with the water level time series changes through the mutual information method, and generating derived variables, thereby obtaining hydrological characteristic extended data; Step S16: performing dimensionless processing on the hydrological characteristic extended data and performing data segmentation based on different time windows, thereby obtaining standardized hydrological basic data; Step S17: Construct a data generation model of a conditional generative adversarial network based on standardized hydrological basic data.

3. The supervised time series water level data generation method according to claim 2, characterized in that: Step S17 includes the following steps: Step S171: constructing a generator network structure of a conditional generative adversarial network based on standardized hydrological basic data, designing a hybrid architecture including multi-layer long short-term memory and convolutional neural network, thereby obtaining generator network structure data; Step S172: generating a discriminator network structure of the adversarial network according to the standardized hydrological basic data design conditions, and performing multi-domain consistency verification processing using a multi-scale time series discrimination strategy to obtain discriminator network structure data, wherein the multi-scale time series discrimination strategy includes time domain and frequency domain feature extraction modules; Step S173: defining the loss function of the generative adversarial network according to the discriminator network structure data and the generator network structure data, thereby obtaining loss function data, wherein the loss function includes loss functions of distribution similarity loss, temporal consistency loss, and hydrological physical constraint loss; Step S174: Integrate the discriminator network structure data, the generator network structure data and the loss function data into a data generation model.

4. The supervised time series water level data generation method according to claim 3, characterized in that: Step S2 includes the following steps: Step S21: acquiring historical hydrological variable data, including temperature, humidity, precipitation, and atmospheric pressure information; performing feature dimensionality reduction processing on the historical hydrological variable data, and extracting key hydrological features, thereby obtaining key hydrological feature data; Step S22: designing a hydrological variable encoding network based on an attention mechanism based on the key hydrological characteristic data, and using the hydrological variable encoding network to perform nonlinear transformation on the key hydrological characteristic data, thereby obtaining conditional input characteristic data; Step S23: According to the conditional input feature data, the generator network of the data generation model is tuned for hydrological adaptability parameters, thereby obtaining hydrological adaptability optimization model parameters, wherein the hydrological adaptability parameter tuning is specifically to design a generation network parameter adaptive optimization strategy based on variational inference; Step S24: optimizing the generator network of the data generation model according to the hydrological adaptability optimization model parameters, thereby obtaining a data generation optimization model; Step S25: Inject the conditional input feature data into the generator network of the data generation optimization model, and generate preliminary water level time series data through adversarial training of the generator and discriminator of the generative adversarial network.

5. The supervised time series water level data generation method according to claim 4, characterized in that: Step S25 includes the following steps: Step S251: using the conditional input feature data as conditional vector input data to generate the initial layer of the generator network in the optimization model, thereby constructing conditionally guided generator network input data; Step S252: performing multi-layer nonlinear transformation on the generator network input data through the generator network in the data generation optimization model, thereby generating candidate water level time series data; Step S253: inputting the candidate water level time series data into the discriminator network in the data generation optimization model, and performing authenticity discrimination and adversarial training according to the preset hydrological basic data, thereby obtaining adversarial training feedback data; Step S254: adaptively adjusting the parameters of the generator network in the data generation optimization model according to the adversarial training feedback data of the discriminator network, and optimizing the quality of the candidate water level time series data, thereby obtaining generator iterative optimization data; Step S255: Loop through steps S252 to S254 until the generated generator iterative optimization data meets the preset adversarial training convergence criteria, thereby obtaining preliminary water level time series data.

6. The supervised time series water level data generation method according to claim 5, characterized in that: Step S3 includes the following steps: Step S31: Designing a stratified sampling weight strategy according to the distribution characteristics and statistical attributes of the preliminary water level time series data, thereby obtaining sampling weight strategy data; Step S32: performing stratified importance sampling processing on the preliminary water level time series data according to the sampling weight strategy data, thereby obtaining importance stratified sampling data; Step S33: setting the differentiated generation weights of the sampling probabilities of the prominent extreme hydrological events based on the preset rare extreme hydrological event characteristics, thereby obtaining differentiated generation weight data; Step S34: performing weighted sampling processing of different types of hydrological events on the extreme hydrological event data according to the differentiated generated weight data, thereby obtaining optimized sampling data of hydrological events; Step S35: oversampling rare extreme hydrological event data by a bootstrap method, thereby obtaining high-quality extreme event data; Step S36: Perform distribution consistency assessment based on statistical characteristics according to importance stratified sampling data, extreme event data and hydrological event optimized sampling data to obtain water level time series data.

7. The supervised time series water level data generation method according to claim 6, characterized in that: Step S36 includes the following steps: Step S361: extracting statistical features from importance stratified sampling data, extreme event data, and hydrological event optimized sampling data, and constructing a multidimensional statistical feature vector; Step S362: designing a distribution distance calculation method based on statistical characteristics according to the multidimensional statistical feature vector, and performing a distribution similarity test based on Kolmogorov-Smirnov, thereby obtaining distribution consistency measurement data; Step S363: According to the distribution consistency measurement data, the importance stratified sampling data, the extreme event data and the hydrological event optimized sampling data are subjected to distribution feature reconciliation processing based on feature interpolation and smoothing technology for different data sets, thereby obtaining aligned water level time series data.

8. The supervised time series water level data generation method according to claim 7, characterized in that: Step S4 includes the following steps: Step S41: verifying the spectrum characteristics of the water level time series data to generate spectrum characteristics verification data, wherein the spectrum characteristics verification specifically uses wavelet transform to decompose multi-scale time series characteristics, analyzes the energy distribution of different time frequency segments, and evaluates the consistency of the spectrum characteristics of the water level time series data with the pre-acquired real water level data; Step S42: analyzing the correlation and information transfer characteristics between different variables of the water level time series data based on the mutual information measurement method, thereby obtaining mutual information enhanced verification data; Step S43: merging the spectrum characteristic verification data and the mutual information enhancement verification data into water level time series verification data; Step S44: Eliminate artifacts in the water level time series verification data based on an abnormality recognition algorithm, thereby obtaining artifact elimination optimized data; Step S45: Perform extreme value abnormal point correction processing based on physical constraints on the artifact removal optimization data to obtain supervised time series water level data.

9. A supervised time series water level data generation system, characterized in that: Used to execute the supervised time series water level data generation method according to claim 1, the supervised time series water level data generation system comprises: Model building module, used to obtain multi-source hydrological monitoring data; perform outlier identification and data preprocessing based on hydrological monitoring data to generate standardized hydrological basic data; build a data generation model of conditional generative adversarial network based on standardized hydrological basic data; The hydrological condition enhancement module is used to regulate the conditional characteristics driven by hydrological variables in the data generation model, optimize the parameters of the generation network, and generate conditional input characteristic data with enhanced hydrological adaptability; the conditional input characteristic data is injected into the data generation model to generate preliminary water level time series data; The sampling optimization module is used to perform stratified importance sampling based on preliminary water level time series data, and to perform differentiated generation weight settings based on preset rare extreme hydrological event data to obtain optimized sampling data for hydrological events; to oversample rare extreme hydrological event data through the self-help method to obtain high-quality extreme event data; to perform distribution consistency evaluation based on statistical characteristics based on extreme event data and hydrological event optimized sampling data to obtain water level time series data; The data verification and correction module is used to perform multi-scale verification on the water level time series data to obtain water level time series verification data, where the multi-scale verification includes spectral characteristic verification and mutual information measurement; the water level time series verification data is used to eliminate artifacts in the data and perform extreme value anomaly correction processing based on physical constraints to obtain supervised time series water level data.

10. A computer-readable storage medium storing a computer program, characterized in that: The computer program is executed to implement the supervised time-series water level data generation method according to any one of claims 1 to 8.

Images