168 results about "Extreme gradient boosting" patented technology

The invention discloses a water quality time sequence prediction method of an SSA-VMD-LSTM-XGBoost hybrid model, and belongs to the technical field of water quality monitoring and prediction. Comprising the following steps: (1) data preparation and preprocessing; (2) optimizing the water quality time sequence decomposition of the VMD based on SSA: optimizing a penalty factor and a modal number of the VMD by adopting a sparrow search algorithm (SSA), and decomposing the water quality time sequence into a plurality of sub-components with high stability and low complexity by utilizing the optimized VMD; (3) construction and training of an LSTM-XGBoost hybrid prediction model: constructing a hybrid prediction model fusing long-short term memory (LSTM) and extreme gradient boost (XGBoost), inputting a high-frequency component into the LSTM model, inputting a low-frequency component into the XGBoost model, and finally performing superposition and integration on prediction results of the models; and (4) multi-component prediction result integration and performance verification. According to the method, adaptive optimization of VMD parameters is realized through SSA, the feature extraction and time sequence modeling capability is improved by combining the advantages of LSTM and XGBoost, and the prediction precision and stability of the water quality time sequence are effectively improved.

The present invention provides a dual-level thermal runaway warning method of a lithium battery based on a sound signal, comprising: obtaining a battery sound signal sequence; performing outlier identification on the battery sound signal sequence, and providing a level-1 thermal runaway warning when an abnormal data point exists; extracting a time-frequency domain feature of the abnormal data point, and identifying a presence of a thermal runaway expansion sound through a sparrow search algorithm-optimized eXtreme Gradient Boosting (SSA-XGBoost) algorithm, to providing a level-2 thermal runaway warning. In the SSA-XGBoost algorithm, optimal parameter adjustment is performed on a number of iterations, a learning rate, and a decision tree depth of the XGBoost algorithm through the SSA. A dual-level thermal runaway warning strategy is adopted to perform grading identification on general anomalies or deep anomalies, thereby effectively improving identification accuracy of a weak abnormal sound signal in an early stage.

The invention discloses a wire harness product quality prediction system based on big data. An initial multi-source data set is acquired; core features in the initial multi-source data set are extracted based on the crimping height, the insulation resistance value and the environment temperature and humidity of the wire harness quality, high-weight features in the core features are screened through a PCA principal component analysis method, and wire harness feature data are obtained; processing time series data based on a long-short-term memory network, performing feature selection by using an extreme gradient boosting tree, and establishing a hybrid prediction model; using an improved IWOA whale optimization algorithm to optimize hyper-parameters of the hybrid prediction model; and inputting the wire harness characteristic data into the target hybrid prediction model for prediction, outputting a quality risk grade index, and if the quality risk grade index exceeds a set threshold, triggering an early warning signal. The limitation of traditional single data or simple model prediction is changed, so that quality prediction better fits an actual production scene, and the accuracy and reliability of prediction are remarkably improved.

The invention relates to the technical field of marine ecological environment monitoring and data analysis, and particularly discloses a sea area phytoplankton biodiversity index prediction method and system based on multi-model integration and feature engineering. After preprocessing, constructing four groups of nonlinear interaction characteristics of a temperature-salt relationship, oxygen-salt balance, chlorophyll chemical oxygen demand coupling and a nitrogen-phosphorus ratio based on environmental factors, combining station characteristics with basic environment and interaction characteristics, carrying out variance threshold screening, inputting a characteristic set into a multi-model integration framework containing models such as linear regression and gradient lifting, and carrying out multi-model integration; training and tuning according to a time sequence segmentation strategy, selecting model output according to a decision coefficient, using a result if the decision coefficient of the support vector regression model is within a preset range, and otherwise, taking a gradient lifting and extreme gradient lifting tree model to predict a mean value. And the prediction accuracy and the model generalization, stability and reliability are improved.

The invention discloses an airport flight quantity and throughput prediction method and system, and relates to the technical field of intelligent airport management. The method comprises the following steps: firstly, determining a future target time period to be predicted, and obtaining related multi-source data according to the future target time period; preprocessing the data and extracting time sequence and non-time sequence features; processing time sequence features by using the trained long-short-term memory neural network model, processing non-time sequence features by using the extreme gradient boosting tree model, and respectively obtaining prediction results; and integrating the outputs of the two types of models through a weighted fusion strategy to obtain an accurate prediction value of the flight sortie and the single-machine passenger capacity, and further calculating to obtain a passenger throughput prediction result. According to the method and the system provided by the invention, through the architecture of multi-source data fusion and hybrid intelligent model cooperation, the prediction precision and reliability are effectively improved, and scientific decision support is provided for airport operation management.

The invention relates to the technical field of water quality prediction, in particular to a river water quality prediction method and system based on a machine learning coupling hydrological model, and the method comprises the following steps: obtaining a data set, and dividing the data set into a training set and a test set; constructing a limit gradient lifting model, and performing hyper-parameter optimization on the model; performing correlation analysis on the water quality parameter set and the water quality index WQI value based on the trained limit gradient lifting model, and screening to obtain key water quality parameters influencing the water quality index WQI value; constructing an LSTM model and a soil and water evaluation tool model; simulating a future hydrological water quality process based on the soil and water evaluation tool model, and calculating and outputting a future water quality parameter result; and inputting a future water quality parameter result into the trained LSTM-WQI model to predict a future river WQI value so as to obtain a prediction result. According to the invention, the machine learning algorithm is coupled with the hydrological model to construct the water quality evaluation model which is convenient to use and adapts to local conditions, and efficient and accurate prediction of future water quality is realized.

The invention discloses an idle land recognition method, device and equipment and a storage medium, and belongs to the technical field of machine learning, and the method comprises the steps: obtaining a first land feature set based on the land text data of a plurality of pieces of land; screening the first land feature set to obtain a second land feature set; semi-supervised learning cooperative training is carried out on a first classifier and a second classifier based on the second land feature set, the first classifier and the second classifier are used for idle land identification, the first classifier is realized based on a rotating forest algorithm, and the second classifier is realized based on an extreme gradient lifting algorithm; wherein in the semi-supervised learning cooperative training process, the prediction labels are screened based on a preset idle land classification rule. The method can realize accurate and efficient idle land identification.

The invention discloses a remote driving risk early warning method based on communication time delay perception, which comprises the following steps: firstly, collecting related data of a target communication link, carrying out preprocessing and feature extraction on the data, fitting a nonlinear relationship between time delay and multiple features based on an extreme gradient lifting XGboost algorithm, and predicting a change condition of the time delay in real time; constructing a communication risk evaluation model, and quantifying a communication risk value based on the time delay change condition; predicting a vehicle trajectory according to the vehicle dynamics model and the driver execution instruction sequence; calculating a lane departure risk and an obstacle collision risk based on the vehicle trajectory, the environmental risk value being the sum of the two; and comprehensively considering an environment risk value and a communication risk value, and judging whether to trigger an alarm according to a designed safety threshold. According to the method, an environment risk evaluation mode is redesigned on the basis of considering the communication time delay, and the environment risk and the communication risk are comprehensively incorporated into an early warning framework, so that better safety guarantee is provided for a remote driving system.

The invention discloses a risk indicator construction method and system for new energy output prediction, and belongs to the technical field of data processing and power system management, and the method comprises the steps: obtaining the preprocessing data of a plurality of new energy stations, carrying out the spatial-temporal feature analysis, and generating a spatial-temporal feature set; obtaining meteorological prediction data, and performing medium and long term output prediction through an extreme gradient boosting tree based on the space-time feature set; comparing the output prediction result with historical real output data to obtain a prediction error, and performing probability coupling modeling and quantile regression prediction in combination with a meteorological element evolution sequence to generate a dynamic quantile value; and determining a dynamic risk interval, and constructing a graded early warning index. According to the method, probabilistic coupling modeling and quantile regression prediction are adopted, prediction errors are deeply associated with dynamic evolution of meteorological elements, predicted uncertain risks can be quantized and graded, and decision support is provided for optimal scheduling and risk management of a power grid.

The application provides a stone composition multi-modal prediction method and system based on urine analysis, and belongs to the technical field of medical detection and artificial intelligence. The method comprises the following steps: acquiring a multi-modal data set of urine of a stone patient; identifying a crystal type; generating a metabolic feature matrix; extracting features from the crystal type identification result to generate a crystal feature vector; performing standardization processing on the metabolic feature matrix and the crystal feature vector to obtain a fusion feature matrix; inputting the fusion feature matrix into an extreme gradient boosting tree model to output an initial probability of a stone composition; inputting the initial probability of the stone composition into a bidirectional long short-term memory network model to output a probability distribution of the stone composition, and completing stone composition prediction. The application breaks through the limitations of traditional technologies in terms of accuracy, efficiency, cost, applicability and the like, and provides a feasible solution for early diagnosis, personalized treatment and popularization of primary medical treatment of urinary stones.

The invention provides a stone component multi-modal prediction method and system based on urine analysis, and belongs to the technical field of medical detection and artificial intelligence crossing. The method comprises the following steps: acquiring a multi-modal data set of urine of a stone patient; identifying the crystal type; generating a metabolic characteristic matrix; performing feature extraction on the crystal type identification result to generate a crystal feature vector; standardizing the metabolism characteristic matrix and the crystallization characteristic vector to obtain a fusion characteristic matrix; inputting the fusion feature matrix into an extreme gradient boosting tree model, and outputting the initial probability of the calculus component; and inputting the initial probability of the calculus component into the bidirectional long-short-term memory network model, and outputting probability distribution of the calculus component to complete calculus component prediction. The method breaks through the limitation of the traditional technology in the aspects of accuracy, efficiency, cost, applicability and the like, and provides a feasible solution for early diagnosis, personalized treatment and basic medical popularization of urinary calculi.

The invention discloses an NEP estimation method based on feature selection and machine learning, and belongs to the technical field of ecological environment monitoring and carbon cycle research, and the method comprises the steps: obtaining remote sensing image data and environment factors of a target region, and carrying out the preprocessing of the obtained remote sensing image data, and obtaining remote sensing factors; based on the remote sensing factor and the environmental factor, obtaining a net ecosystem productivity NEP feature influencing multi-source data fusion of the target area, performing screening, obtaining a preset number of key factors, and performing training through an extreme gradient lifting regression algorithm; analyzing the training result of the extreme gradient lifting regression algorithm by adopting a Shapril addition interpretation method; the extreme gradient lifting regression algorithm is verified through the evaluation indexes, and target area NEP estimation is completed; according to the method, high-precision estimation and driving mechanism analysis of the NEP of the complex mountain ecosystem are realized by fusing multi-source data and a machine learning technology, and clear mechanism explanation is provided for ecological research.

The invention discloses a soil texture remote sensing mapping method based on time window screening. The method comprises the following steps: step 1, collecting and testing a soil sample; 2, screening a remote sensing image time window; step 3, acquiring and preprocessing an environment covariable; 4, constructing a multi-source data set; 5, constructing a soil texture prediction model; 6, multi-temporal remote sensing covariable combination optimization modeling is carried out; and step 7, soil texture mapping based on the optimal multi-temporal remote sensing combination. And fusing the multi-source environment covariable and the multi-temporal remote sensing data, and constructing a soil texture prediction model based on the optimal multi-temporal model combination by adopting an extreme gradient lifting algorithm. The problems of insufficient representativeness and weak generalization ability of single-time-phase remote sensing data are solved. On one hand, the generalization ability of the model is improved, and on the other hand, quantitative analysis of the key environmental factors is realized in combination with the SHAP technology, the interpretability of the model is enhanced, and meanwhile, the soil texture prediction precision is effectively improved.

The invention relates to the field of laser welding monitoring, and provides a laser welding penetration state prediction method based on a multi-modal composite machine learning model, which comprises the following steps: S1, acquiring a multi-modal signal in a laser welding process by multiple sensors; s2, carrying out preprocessing on the multi-mode signal; s3, establishing a multi-modal data set based on the preprocessed data, and dividing the multi-modal data set into a training set and a test set; and S4, establishing a machine learning model of a composite neural network (NN) and an Extreme Gradient Boosting (XGBoost) algorithm, training on the training set, and verifying the prediction performance of the model on the laser welding penetration state on the test set. The invention aims to fully fuse different modal information in the laser welding process and combine the advantages of different machine learning algorithms to improve the accuracy and robustness of penetration state monitoring.

The invention provides a loan risk assessment method which can be applied to the technical field of artificial intelligence. The method comprises the steps that under the condition of user authorization, structured data and time series data of a user are acquired, the structured data are identity attributes, financial qualification and loan application information of the user, and the time series data are historical credit behavior data of the user; inputting the structured data into a preset random forest model, and outputting a first predicted value; inputting the structured data into a preset limit gradient lifting model, and outputting a second predicted value; inputting the time sequence data into a preset long-short-term memory network model, and outputting a third predicted value; performing weighted fusion on the first predicted value, the second predicted value and the third predicted value to obtain a combined predicted value; and outputting a final loan risk assessment result according to the joint prediction value. The invention further provides a loan risk assessment device and equipment, a storage medium and a program product.

The invention discloses a spiral anchor bearing capacity prediction method based on machine learning, and solves the problems that in the prior art, numerical analysis consumes long time, actuarial precision is limited by precise description of soil constitutive, the operation difficulty is high, and large-scale application is difficult. According to the method, methods of algorithm and database establishment, model training and hyper-parameter optimization, model evaluation, characteristic important analysis and the like are adopted, an ensemble learning algorithm with high popularity is utilized to improve the limit gradient to establish a screw anchor bearing capacity prediction model, and the screw anchor bearing capacity is rapidly and accurately predicted based on a machine learning model.

The present application relates to a kind of carbon emission space refinement simulation and driving mechanism analysis method based on multi-source data and machine learning, including obtaining night light remote sensing, population density data, multi-source data such as interest point density, nonlinear spatial fitting model is constructed and trained by using information entropy weight method to carry out multi-source data fusion and construct economic energy comprehensive index, use city land use classification data as industry space restriction, combined with the total carbon emission of each industry calculated to build carbon emission space simulation base map, based on extreme gradient boosting tree, introduce SHAP framework for post-analysis, from global, local and spatial three scales, the nonlinear driving mechanism of carbon emission of multi-source city form element is analyzed, overcome the over-saturation defect of single night light data in city core area, realize the leap from pure space fitting to mechanism explanation, provide scientific basis and method support for high-density city accurate locking emission hotspot and making low-carbon policy.

The invention discloses a short message classification method and system, and relates to the technical field of artificial intelligence, and the method comprises the steps: carrying out the feature extraction of a short message text, and obtaining a multi-dimensional feature; and inputting the multi-dimensional features into a pre-trained extreme gradient boosting tree model to obtain a probability that the to-be-predicted short message belongs to a specified type, if the importance probability is greater than a preset threshold value, determining that the to-be-predicted short message is an important short message, and if the importance probability is less than the preset threshold value, determining that the to-be-predicted short message is a non-important short message. The problem that the importance of the short messages cannot be efficiently, accurately and automatically classified in the prior art is solved.

The invention provides a method and a device for predicting the effluent concentration of pollutants in a constructed wetland based on machine learning, and the technical key points are as follows: constructing a multi-feature data set by collecting data including environment operation conditions, water quality parameters and wetland conditions of the constructed wetland; an initial constructed wetland water quality prediction model is constructed, and the processed multi-feature data set is trained through four single-output machine learning models of a random forest, gradient lifting, limit gradient lifting and a gradient lifting decision tree; a plurality of indexes are adopted to evaluate the prediction performance of the four single-output machine learning models on the test set, and an optimal single-output model with the optimal comprehensive performance of the indexes is obtained; establishing a multi-target joint prediction model of multi-output regression by using the optimal single-output model, and predicting the water quality index of the effluent of the constructed wetland; and sorting the importance of different target variables which are explained by the multi-target joint prediction model and are influenced by the multi-target joint prediction model through an SHAP method.

The invention relates to the technical field of data processing and engineering project process management, in particular to a full-process intelligent management method and system for bridge and tunnel maintenance engineering. The method comprises the following steps: extracting a disease geometric contour based on bridge and tunnel inner wall three-dimensional point cloud data, and calculating a centroid coordinate, a volume and a surface contour area; constructing a space retrieval sphere by using a target disease centroid, screening an adjacent disease set, and calculating a local aggregation density by combining the volume and the centroid distance; calculating the average depth of the target disease, and deducing the overlapping sharing rate of the working plane; constructing a group topology consumption equivalent through linear analysis; and inputting the extreme gradient lifting model to output an optimal repair material dosage table and a material distribution operation instruction book. According to the method, the disease space distribution and aggregation degree can be accurately evaluated, the construction loss surface overlapping probability and the material sharing potential are scientifically evaluated, the material preparation accuracy is improved, waste and cost are reduced, and the whole-process intelligent management of bridge and tunnel maintenance engineering is realized.

The invention discloses a neonatal congenital heart disease screening method based on machine learning, and the method comprises the following steps: 1, designing a questionnaire according to literatures related to adverse birth result risk factors, collecting infant cardiac ultrasound examination data, and then preprocessing the data; step 2, randomly dividing the data obtained in the step 1 into a training set and a test set, and performing dimension reduction processing on features by using lASSO regression; and step 3, based on the features obtained in the step 2, training a prediction model by using machine learning algorithms such as a random forest, a support vector machine, a lightweight gradient elevator, logistic regression and extreme gradient lifting. According to the method, the model is optimized through combination of multi-step data processing and various machine learning algorithms, related features can be comprehensively and scientifically screened, the accuracy of the prediction model is improved, a more reliable and effective method is provided for screening the neonatal congenital heart disease, early-stage accurate screening is facilitated, and the screening efficiency and quality are improved.

The invention provides a steel smelting yield dynamic analysis and prediction method and system, and the method comprises the steps: S1, obtaining data according to the demands of a user, and carrying out the preprocessing of the data, and obtaining a preprocessing result; s2, according to the preprocessing result, analyzing to obtain influence factors of the steel smelting yield; s3, fitting a training model according to the influence factors; s4, outputting a prediction result of the model, and evaluating the model according to the prediction result to obtain an evaluation result; and S5, selecting a corresponding model according to the evaluation result, and predicting the steel smelting yield. According to the method, the yield influence factors are determined by combining the result of the limit gradient lifting model and big data experience, and the prediction result is generated through the multi-layer perceptron model, so that the difference of actual output is shortened, and decision deployment of steel smelting production is facilitated.

The invention provides an injection-production well pipe detection period prediction method, and belongs to the field of petroleum and natural gas industry. The method comprises the following steps: firstly, collecting actual pipe detection periods and related production data and state data of a plurality of injection-production wells, assigning values to the related state data, and constructing a data set; secondly, preprocessing the data set; then main control factors are screened; constructing an injection-production well pipe inspection period prediction model by using an extreme gradient lifting tree, and adjusting hyper-parameters; training the injection-production well pipe inspection period prediction model by using the preprocessed actual pipe inspection period of the injection-production well and the sample data of the main control factors of the injection-production well; calculating evaluation indexes of the injection-production well pipe inspection period prediction model for performance evaluation; then data of main control factors of the in-service injection and production well pipe inspection period are obtained; and finally, predicting the pipe inspection period of the in-service injection and production well by using the trained injection and production well pipe inspection period prediction model. By adopting the method, the pipe detection period of the injection-production well can be accurately predicted, the production decision is optimized, and the production safety is ensured.

The invention discloses a distribution network line fault trip prediction method and device, a medium and a product, and relates to the field of power distribution network operation and maintenance, and the method comprises the steps: obtaining fault data of a distribution network line; constructing a fault data set; according to the fault data set, constructing a trip prediction model based on an akaike information criterion; an exogenous variable corresponding to the tripping prediction model is determined; constructing a limit gradient lifting model based on a grid search method according to the exogenous variables and the historical fault trip times; constructing a fusion model according to the trip prediction model and the limit gradient lifting model; respectively determining prediction errors of the tripping prediction model and the fusion model, and determining a final prediction model according to the prediction errors; determining the number of fault tripping times of the month to be predicted by using the final prediction model; and regularly obtaining fault data of the distribution network line to regularly update the final prediction model. The monthly fault trip times of the distribution network line can be efficiently predicted, and the operation and maintenance reliability of the distribution network is improved.

The invention discloses a hydropower station peak capability evaluation method under an extreme scene. The method comprises the following steps: automatically classifying hydrological scenes of the hydropower station based on a ten-day runoff drought threshold value and a ten-day runoff flood threshold value of the hydropower station in a kurtosis period; based on the hydrological scene classification result of the hydropower station, simulating the operation process of the hydropower station by using a hydropower guarantee supply absorption operation model, and obtaining the peak duration of the hydropower station so as to construct a training set; and constructing a hydropower station peak capability evaluation model based on an extreme gradient lifting method, training the hydropower station peak capability evaluation model by using the training set to obtain a trained hydropower station peak capability evaluation model, and further obtaining a peak capability evaluation result and a corresponding visual analysis result of the hydropower station in a peak summering period. According to the method, the problem that the peak capacity of the hydropower station in extreme weather cannot be quantitatively evaluated by a traditional method is solved, the conversion from passive response to active pre-judgment is realized, and the scientificity and foresight of a power grid supply guarantee decision are improved.

The invention discloses a food detection data management system and method based on cloud computing, and the method comprises the steps: obtaining a structured data set, carrying out the preprocessing of the structured data set in a data processing module, and transmitting the preprocessed structured data set to a cloud platform; constructing a dynamic risk assessment model based on an AHP (Analytical Hierarchy Process) and an XGBoost extreme gradient boosting tree, and inputting the preprocessed structured data set into the dynamic risk assessment model for identification; marking the detected food according to the food safety risk index, and generating a unique Hash value of the structured data set by using an SHA-256 Hash algorithm to obtain a detection data Hash value; and writing the marked structured data set and the detection data hash value into a block chain for evidence storage, and deploying an intelligent contract to execute compliance verification management. Data sharing and cooperative work can be realized on a unified platform, an information island is broken, and the working efficiency is improved.

The invention discloses a physical information neural network-based magnetic element magnetic core loss prediction method, which is characterized in that a PINN-based model is constructed by taking a hybrid network architecture as a baseline algorithm, a Conv-LSTM (Convolutional Long Short-Term Memory Network), a PSD (Power Spectral Density) and an integrated learning method are combined, and the method comprises XGB (Limit Gradient Braking), GBR (Gradient Braking Regression) and RF (Random Forest). The objective of the invention is to solve the complexity of magnetic core loss prediction. In addition, the Steinmetz equation is improved to enhance the adaptability of the Steinmetz equation under complex conditions, and the improved Steinmetz equation is used as a physical constraint to be integrated into the neural network for magnetic core loss prediction. Based on a traditional data-driven loss item, a physical residual item is introduced as a regularization constraint, so that prediction can meet observation data distribution and also can conform to a physical rule. Experimental results show that the PINN-based model has good prediction performance on the magnetic core loss under complex conditions. And a reference is provided for magnetic core loss prediction in practical engineering.

The application provides a single crystal silicon edge break prediction method based on improved SSA and extreme gradient boosting, and relates to the technical field of data processing. The method comprises the following steps: calculating the difference between the current constant diameter growth stage data of the single crystal silicon to be predicted and the standard constant diameter growth stage data determined according to the weight coefficients formed by clustering the historical constant diameter growth stage data of a plurality of sample single crystal silicon, to obtain current difference constant diameter growth stage data; performing statistical feature extraction on the current difference constant diameter growth stage data to obtain current difference features; inputting the current difference features into a prediction model to obtain the edge break prediction result of the single crystal silicon to be predicted; and the prediction model is obtained by optimizing an extreme gradient boosting model based on the historical difference features and edge break labels of each sample single crystal silicon and an improved sparrow optimization algorithm (SSA). The application realizes efficient and accurate single crystal silicon edge break prediction by reducing the dependence on a mechanism model.

This invention relates to the field of machine learning technology, and in particular to a method for predicting material removal depth based on an improved XGBoost algorithm. The method first collects a grinding and polishing process dataset and divides it into training and testing sets. After preprocessing the dataset, an improved Egret Optimization Algorithm (IESOA) is used to optimize the hyperparameters of the Extreme Gradient Boosting Algorithm (XGBoost), thereby establishing a predictive model for material removal depth. Finally, the contribution of each grinding and polishing process parameter to the material removal depth is quantified based on the Shapley Additive Interpretation (SHAP) method, enabling visual analysis of the model's decision-making process. This invention can significantly improve the prediction accuracy of material removal depth in grinding and polishing processes, and clarifies the influence mechanism of each process parameter through interpretability analysis, possessing the dual advantages of high prediction accuracy and interpretable decision-making.

This invention discloses a method for simulating extreme precipitation using a CNN-LSTM model based on multi-source data fusion. First, multi-source data of the study area are acquired, and the regional extreme precipitation is calculated. An extreme gradient boosting model is used to select the key atmospheric circulation indices that have the most significant impact on extreme precipitation in the target area. The multi-source data undergoes unified preprocessing. Then, a spatiotemporal prediction model fusing convolutional neural networks and long short-term memory networks is constructed, and a deep learning prediction model is generated through joint training of multiple features. Finally, the preprocessed multi-source data and the deep learning model are used to perform spatiotemporal prediction of extreme precipitation. This invention integrates extreme precipitation data with atmospheric circulation indices and topographic DEM data into a unified input framework, enabling the model to simultaneously utilize spatiotemporal precipitation distribution characteristics, atmospheric circulation patterns, and topographic information to jointly predict precipitation, resulting in high prediction and simulation accuracy.