Method and device for predicting productivity of low-permeability gas well

Abstract

The invention discloses a method and a device for predicting the productivity of a low-permeability gas well. The method comprises the following steps: acquiring data of the low-permeability gas well; preprocessing the acquired data of the low-permeability gas well to form a multi-source heterogeneous data set; constructing a multi-model stacking framework; using the multi-source heterogeneous data set to train the multi-model stacking architecture to obtain a prediction model; the prediction model outputs a prediction result; according to the method, the multi-model stacking technology is adopted, the defects of a single model are overcome by combining the advantages of models such as the random forest, the support vector machine and the lightweight neural network, the output of the sub-models is further optimized by introducing the second layer of sub-models, the prediction result is more accurate, and meanwhile the prediction efficiency is improved through the model stacking technology. And the non-linear expression capability of the deep learning model and the robustness of the traditional machine learning model are combined, so that the method can better adapt to the heterogeneity and stress sensitivity characteristics of the low-permeability gas reservoir.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and natural gas exploration and development, and in particular to a method and device for predicting the productivity of a low-permeability gas well. Background Art

[0002] Low-permeability gas reservoirs are widely distributed in the global oil and gas resources, but due to the strong reservoir heterogeneity, low permeability, and significant stress sensitivity, the production capacity prediction of low-permeability gas wells has become a major technical challenge in the development process. Traditional production capacity prediction methods are difficult to capture the nonlinear relationship between multiple variables, and fail to fully consider different types of input parameters, resulting in limited prediction accuracy. In recent years, machine learning methods have gradually been widely used, but the existing machine learning methods are relatively single and difficult to comprehensively utilize multidimensional data. In addition, the current fusion model is difficult to dynamically adjust the weights, resulting in poor model fusion effect and inability to balance local characteristics and global performance.

[0003] Traditional machine learning capacity prediction methods usually use single models such as support vector machines, random forests or neural networks, but due to their limitations, it is difficult to comprehensively utilize the characteristics of multidimensional data and the prediction accuracy is limited. While some current studies attempt to improve accuracy through model fusion, the fixed weighted average or voting method they use lacks adaptive adjustment capabilities and is difficult to handle the heterogeneity of low-permeability gas reservoirs. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies of the above-mentioned prior art and provide a method and device for predicting the productivity of a low-permeability gas well, which can be adapted to different types of data distribution and characteristics, enhance the robustness of the prediction model, and improve the accuracy and generalization ability of the prediction.

[0005] To achieve the above object, the technical solution of the present invention is: In a first aspect, the present invention provides a method for predicting the productivity of a low permeability gas well, comprising: Obtain low permeability gas well data; Preprocess the acquired low-permeability gas well data to form a multi-source heterogeneous data set; Construct a multi-model stacking architecture; the multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine, and lightweight neural network; the output of the first-layer sub-model is used as the input of the second-layer sub-model, and in the inter-layer optimization process, the performance of the first-layer sub-model is evaluated by using K-fold cross validation, and its prediction result is used as the data source for training the second-layer sub-model; Using the multi-source heterogeneous data sets to train the multi-model stacking architecture to obtain a prediction model; The prediction model outputs a prediction result.

[0006] Optionally, the preprocessing of the acquired low permeability gas well data includes: The KNN method is used to fill the missing data in the low permeability gas well data; the calculation formula of the KNN method is:

[0007] in, are the known values ​​of the K neighbors; The z-score method is used to detect and remove outliers in the low permeability gas well data; the calculation formula of the z-score method is:

[0008] in, is a data point, is the mean of the data set, is the standard deviation of the data set; Correlation analysis and principal component analysis were used to screen important features in low-permeability gas well data and remove redundant information. The data of different dimensions in the low permeability gas well data are normalized or standardized; the formula for the standardization is:

[0009] in, is the standardized data; is the original data; is the minimum value in the original data; is the maximum value of the original data.

[0010] Optionally, the use of the multi-source heterogeneous datasets to train the multi-model stacking architecture includes three levels: sub-model training, multi-model stacking training and second-layer sub-model training.

[0011] Optionally, the multi-source heterogeneous data set is divided into a training set and a test set. The K-fold cross-validation method is used each time to divide the training set into K equal parts. K-1 parts of the data are used for training each time, and the remaining part is used as a validation set to complete the training of the multi-model stacking architecture.

[0012] Optionally, the multi-model stacking training includes: using the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network as the input features of the second-layer sub-model, and retaining the feature outputs of the three sub-models of random forest, support vector machine and lightweight neural network during the training process.

[0013] Optionally, the second-layer sub-model training includes: using XGBoost as the second-layer sub-model, and optimizing the number of XGBoost trees, learning rate and depth parameters through the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network.

[0014] Optionally, a genetic algorithm optimization method is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second-layer sub-model.

[0015] Optionally, the genetic algorithm optimization method is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second layer sub-model, including: Based on the initialization population method, multiple random weight combinations are generated, each of which represents the weight of a different sub-model in the second-layer sub-model; The root mean square error and determination coefficient are used as fitness functions to evaluate the quality of each weight combination. The weight combination that reaches the threshold of fitness will enter the next round. Select the weight combination whose fitness reaches the threshold for crossover, combine the weight parts of different individuals, and generate a new weight combination; By randomly adjusting the weight values ​​of some individuals, variations are introduced to maintain the diversity of the population; After each crossover and mutation operation, the new weight combination will undergo fitness evaluation. The weight combination that meets the requirements will be selected into the next generation and gradually optimized to eventually find an optimal or near-optimal weight score.

[0016] In a second aspect, the present invention provides a low permeability gas well productivity prediction device, comprising: A data acquisition module, used to acquire low permeability gas well data; A data preprocessing module is used to preprocess the acquired low-permeability gas well data to form a multi-source heterogeneous data set; A multi-model stacking architecture module is used to construct a multi-model stacking architecture; the multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine and lightweight neural network; the output of the first-layer sub-model is used as the input of the second-layer sub-model, and in the inter-layer optimization process, K-fold cross-validation is used to evaluate the performance of the first-layer sub-model, and its prediction result is used as the data source for training the second-layer sub-model.

[0017] A model training module uses the multi-source heterogeneous data set to train the multi-model stacking architecture to obtain a prediction model; The result output module outputs the prediction result based on the prediction model.

[0018] Optionally, the low permeability gas well productivity prediction device further includes: The dynamic weight optimization module is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second-layer sub-model by using the genetic algorithm optimization method.

[0019] Compared with the prior art, the present invention has the following beneficial effects: The present invention adopts multi-model stacking technology, which makes up for the shortcomings of a single model by combining the advantages of models such as random forest, support vector machine and lightweight neural network. The introduction of the second-layer sub-model further optimizes the output of the sub-model, making the prediction result more accurate. At the same time, through model stacking technology, combined with the nonlinear expression ability of deep learning models and the robustness of traditional machine learning models, it can better adapt to the heterogeneity and stress sensitivity characteristics of low-permeability gas reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A flow chart of a method for predicting the productivity of a low permeability gas well provided in an embodiment of the present application; Figure 2 A schematic diagram of a method for predicting the productivity of a low-permeability gas well provided in an embodiment of the present application; Figure 3 A comparison chart of the results of different single machine learning capacity prediction models and the low permeability gas well capacity prediction method based on multi-model stacking and weight adaptive optimization in this application; Figure 4 A schematic diagram of the composition of a low permeability gas well productivity prediction device provided in an embodiment of the present application; Figure 5 A schematic diagram of the composition of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0023] Example: In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0024] To facilitate the understanding of the embodiments of the present application, further explanation will be given below in conjunction with the drawings and specific embodiments. The embodiments do not constitute a limitation on the embodiments of the present application.

[0025] Figure 1The flowchart of the method for predicting the productivity of a low permeability gas well provided in the embodiment of the present application is as follows: Figure 1 As shown, the method for predicting the production capacity of a low permeability gas well provided in the embodiment of the present application mainly includes the following steps: 110, obtaining low permeability gas well data; 120, preprocessing the acquired low permeability gas well data to form a multi-source heterogeneous data set; 130, building a multi-model stacking architecture; Specifically, the multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine, and lightweight neural network; Random forest (RF): uses its powerful nonlinear modeling ability to capture the relationship between various parameters and production capacity. Support vector machine (SVR): enhances the model's adaptability to small sample data sets and the ability of complex high-dimensional data; Lightweight neural network (LightNN): enhances the ability to capture complex nonlinear feature relationships through deep learning. In this way, by combining the advantages of models such as random forest, support vector machine, and lightweight neural network, the shortcomings of a single model are made up.

[0026] The output of the first layer sub-model is used as the input of the second layer sub-model, and in the inter-layer optimization process, the performance of the first layer sub-model is evaluated using K-fold cross validation, and its prediction results are used as the data source for training the second layer sub-model. In this way, the output of the sub-model is further optimized by introducing the second layer sub-model, making the prediction results more accurate.

[0027] 140, using the multi-source heterogeneous data set to train the multi-model stacking architecture to obtain a prediction model; 150. The prediction model outputs a prediction result.

[0028] It can be seen that this method adopts multi-model stacking technology to make up for the shortcomings of a single model by combining the advantages of models such as random forest, support vector machine and lightweight neural network. The introduction of the second-layer sub-model further optimizes the output of the sub-model, making the prediction result more accurate. At the same time, through model stacking technology, combined with the nonlinear expression ability of deep learning models and the robustness of traditional machine learning models, it can better adapt to the heterogeneity and stress sensitivity characteristics of low permeability gas reservoirs.

[0029] In a specific implementation, the above step 120 includes: extracting reservoir parameters and production dynamic data from the production data of the low permeability gas well, and after preprocessing, unifying the format to form a multi-source heterogeneous data set. The preprocessing includes: The KNN method is used to fill in missing data to ensure data integrity. The calculation formula is:

[0030] in, are the known values ​​of the K neighbors.

[0031] The z-score method is used to detect and remove outliers. The calculation formula is:

[0032] in, is a data point, is the mean of the data set, is the standard deviation of the data set.

[0033] The first-year cumulative gas production of the gas well is taken as the target value, and the rest are characteristic values. Correlation analysis and principal component analysis (PCA) methods are used to screen important features, remove redundant information, and reduce data dimensions.

[0034] Pearson correlation coefficient calculation formula:

[0035] in, is the correlation coefficient value; For variables The corresponding different values; For variables The average value of For variables The corresponding different values; For variables The average value of is the number of variables.

[0036] Normalize or standardize data of different dimensions to enhance the stability of subsequent model training. The standardization formula is:

[0037] in, is the standardized data; is the original data; is the minimum value in the original data; is the maximum value of the original data.

[0038] In a specific implementation, the multi-source heterogeneous data set is used to train the multi-model stacking architecture, including three levels: sub-model training, multi-model stacking training, and second-layer sub-model training, so that it can adapt to different types of data distributions and characteristics and enhance the robustness of the resulting prediction model.

[0039] In a specific embodiment, the multi-source heterogeneous data set is divided into a training set and a test set. The K-fold cross-validation method is used each time to divide the training set into K equal parts. K-1 parts of data are used for training each time, and the remaining part is used as a validation set to complete the training of the multi-model stacking architecture. In this way, by looping K times, it is ensured that each part of the data is used as a validation set, avoiding the impact of data distribution bias on model performance evaluation.

[0040] In a specific embodiment, the multi-model stacking training includes: using the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network as the input features of the second-layer sub-model, and retaining the feature outputs of the three sub-models during the training process. In this way, by "stacking" the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network, a new feature set is formed, and the prediction ability is enhanced by combining the outputs of different models.

[0041] In a specific embodiment, the second layer sub-model training includes: using XGBoost as the second layer sub-model, optimizing the number of XGBoost trees, learning rate, depth and other hyper parameters through the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network input. Through training, XGBoost can fully integrate the prediction advantages of the three sub-models and improve the overall performance.

[0042] In a preferred embodiment, after each training, the genetic algorithm optimization method is used to dynamically adjust the weights of the three sub-models in the second-layer sub-model. In this way, the weight of the model is dynamically adjusted by adopting the weight adaptive optimization method strategy, so as to adapt to the changes in complex reservoir heterogeneity and improve the accuracy and generalization ability of the prediction model's capacity prediction, including: Based on the population initialization method, multiple random weight combinations are generated, each combination represents the weight of a different sub-model in the second-layer sub-model; in this embodiment, the weight combination of the three sub-models of random forest, support vector machine and lightweight neural network is [w1, w2, w3], where each w represents the weight of a sub-model.

[0043] The root mean square error RMSE and the coefficient of determination R² are used as fitness functions to evaluate the quality of each weight combination. The weight combination that reaches the threshold of fitness will enter the next round. The root mean square error formula is:

[0044] in, is the sample size; is the actual value; is the predicted value.

[0045] The coefficient of determination formula is:

[0046] in, is the sample size; is the actual value; is the predicted value; is the average value.

[0047] Select the weight combination whose fitness reaches the threshold for crossover, combine the weight parts of different individuals, and generate a new weight combination; By randomly adjusting the weight values ​​of certain individuals, mutations are introduced and the diversity of the population is maintained; this process is equivalent to "gene recombination" and can produce more diverse candidate solutions.

[0048] After each crossover and mutation operation, the new weight combination will undergo fitness evaluation. The weight combination that meets the requirements will be selected into the next generation and gradually optimized to eventually find an optimal or near-optimal weight score.

[0049] Through the above steps, when new data is input or reservoir conditions change, the genetic algorithm can be restarted. Through this dynamic adjustment process, the model will adaptively adjust the weights according to the new data and changing reservoir characteristics to maintain high-precision prediction.

[0050] In this way, through the above method, the genetic algorithm can optimize the weights of each sub-model in the multi-model stack, adapt to the data changes caused by reservoir heterogeneity, and enable the prediction model to maintain good generalization ability under different geological conditions or production environments, thereby improving the accuracy of production capacity prediction.

[0051] like Figure 3 As shown, there is a comparison chart of the results of different single machine learning capacity prediction models and the low permeability gas well capacity prediction method based on multi-model stacking and weight adaptive optimization of the present application. The results show that the low permeability gas well capacity prediction method based on multi-model stacking and weight adaptive optimization proposed in the present invention has significantly improved the prediction accuracy compared with the traditional single machine learning capacity prediction model, providing a new methodological basis for the capacity prediction of low permeability gas wells.

[0052] like Figure 4 As shown, the embodiment of the present application further provides a low permeability gas well productivity prediction device 400, comprising: A data acquisition module 410 is used to acquire low permeability gas well data; The data preprocessing module 420 is used to preprocess the acquired low permeability gas well data to form a multi-source heterogeneous data set; The multi-model stacking architecture module 430 is used to construct a multi-model stacking architecture; the multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine and lightweight neural network; the output of the first-layer sub-model is used as the input of the second-layer sub-model, and in the inter-layer optimization process, K-fold cross-validation is used to evaluate the performance of the first-layer sub-model, and its prediction result is used as the data source for training the second-layer sub-model.

[0053] A model training module 440 uses the multi-source heterogeneous data set to train the multi-model stacking architecture to obtain a prediction model; The result output module 450 outputs the prediction result based on the prediction model.

[0054] As a preferred embodiment of the low permeability gas well productivity prediction device on the market, the device also includes: The dynamic weight optimization module is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second-layer sub-model using the genetic algorithm optimization method, including: Based on the initialization population method, multiple random weight combinations are generated, each of which represents the weight of a different sub-model in the second-layer sub-model; The root mean square error and determination coefficient are used as fitness functions to evaluate the quality of each weight combination. The weight combination that reaches the threshold of fitness will enter the next round. Select the weight combination whose fitness reaches the threshold for crossover, combine the weight parts of different individuals, and generate a new weight combination; By randomly adjusting the weight values ​​of some individuals, variations are introduced to maintain the diversity of the population; After each crossover and mutation operation, the new weight combination will undergo fitness evaluation. The weight combination that meets the requirements will be selected into the next generation and gradually optimized to eventually find an optimal or near-optimal weight score.

[0055] Optionally, the preprocessing of the acquired low permeability gas well data includes: The KNN method is used to fill the missing data in the low permeability gas well data; the calculation formula of the KNN method is:

[0056] in, are the known values ​​of the K neighbors.

[0057] The z-score method is used to detect and remove outliers in the low permeability gas well data; the calculation formula of the z-score method is:

[0058] in, is a data point, is the mean of the data set, is the standard deviation of the data set.

[0059] Correlation analysis and principal component analysis were used to screen important features in low-permeability gas well data and remove redundant information. The data of different dimensions in the low permeability gas well data are normalized or standardized; the formula for the standardization is:

[0060] in, is the standardized data; is the original data; is the minimum value in the original data; is the maximum value of the original data.

[0061] Optionally, the use of the multi-source heterogeneous datasets to train the multi-model stacking architecture includes three levels: sub-model training, multi-model stacking training and second-layer sub-model training.

[0062] Optionally, the multi-source heterogeneous data set is divided into a training set and a test set. The K-fold cross-validation method is used each time to divide the training set into K equal parts. K-1 parts of the data are used for training each time, and the remaining part is used as a validation set to complete the training of the multi-model stacking architecture.

[0063] Optionally, the multi-model stacking training includes: using the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network as input features of the second-layer sub-model, and retaining the feature outputs of the three sub-models during the training process.

[0064] Optionally, the second-layer sub-model training includes: using XGBoost as the second-layer sub-model, and optimizing the number, learning rate, and depth parameters of XGBoost through the prediction results of the three sub-models of random forest, support vector machine, and lightweight neural network.

[0065] It should be noted that the low permeability gas well productivity prediction device provided in the embodiment of the present application can execute the low permeability gas well productivity prediction method provided in any embodiment of the present application, and has the corresponding functions and beneficial effects of the execution method.

[0066] like Figure 5As shown, an embodiment of the present application provides an electronic device, including a processor 111, a communication interface 112, a memory 113 and a communication bus 114, wherein the processor 111, the communication interface 112, and the memory 113 communicate with each other through the communication bus 114; the memory 113 is used to store computer programs; the processor 111 is used to implement the steps of the low permeability gas well production capacity prediction method provided by any of the aforementioned method embodiments when executing the program stored in the memory 113.

[0067] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the method for predicting the productivity of a low-permeability gas well provided in any one of the aforementioned method embodiments are implemented.

[0068] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0069] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable ordinary technicians in the field to understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made based on the essence of the content of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for predicting the productivity of a low permeability gas well, characterized in that: include: Obtain low permeability gas well data; Preprocess the acquired low-permeability gas well data to form a multi-source heterogeneous data set; Build a multi-model stacking architecture; The multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine and lightweight neural network; the output of the first-layer sub-model is used as the input of the second-layer sub-model, and in the inter-layer optimization process, the performance of the first-layer sub-model is evaluated by K-fold cross validation, and its prediction result is used as the data source for training the second-layer sub-model; The genetic algorithm optimization method is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second-layer sub-model, including: based on the initialization population method, multiple random weight combinations are generated, each combination represents the weight of different sub-models in the second-layer sub-model; the root mean square error and determination coefficient are used as fitness functions to evaluate the quality of each weight combination, and the weight combination with the fitness reaching the threshold enters the next round; the weight combination with the fitness reaching the threshold is selected for crossover, and the weight parts of different individuals are combined to generate a new weight combination; by randomly adjusting the weight values ​​of some individuals, mutations are introduced to maintain the diversity of the population; after each crossover and mutation operation, the new weight combination will be evaluated for fitness, and the weight combination that meets the requirements will be selected into the next generation and gradually optimized to finally find an optimal or approximately optimal weight score; Using the multi-source heterogeneous data sets to train the multi-model stacking architecture to obtain a prediction model; The prediction model outputs a prediction result.

2. The method for predicting the productivity of a low permeability gas well according to claim 1, characterized in that: The preprocessing of the acquired low permeability gas well data includes: The KNN method is used to fill the missing data in the low permeability gas well data; the calculation formula of the KNN method is: ； in, are the known values ​​of the K neighbors; The z-score method is used to detect and remove outliers in the low permeability gas well data; the calculation formula of the z-score method is: ； in, is a data point, is the mean of the data set, is the standard deviation of the data set; Correlation analysis and principal component analysis were used to screen important features in low-permeability gas well data and remove redundant information. The data of different dimensions in the low permeability gas well data are normalized or standardized; the formula for the standardization is: ； in, is the standardized data; is the original data; is the minimum value in the original data; is the maximum value of the original data.

3. The method for predicting the productivity of a low permeability gas well according to claim 1, characterized in that: The use of the multi-source heterogeneous datasets to train the multi-model stacking architecture includes three levels: sub-model training, multi-model stacking training, and second-layer sub-model training.

4. The method for predicting the productivity of a low permeability gas well according to claim 1 or 3, characterized in that: The multi-source heterogeneous data set is divided into a training set and a test set. The K-fold cross-validation method is used each time to divide the training set into K equal parts. K-1 parts of the data are used for training each time, and the remaining part is used as a validation set to complete the training of the multi-model stacking architecture.

5. The method for predicting the productivity of a low permeability gas well according to claim 3, characterized in that: The multi-model stacking training includes: using the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network as the input features of the second-layer sub-model, and retaining the feature outputs of the three sub-models of random forest, support vector machine and lightweight neural network during the training process.

6. The method for predicting the productivity of a low permeability gas well according to claim 3, characterized in that: The second-layer sub-model training includes: using XGBoost as the second-layer sub-model, and optimizing the number of XGBoost trees, learning rate and depth parameters through the prediction results of the three sub-models of random forest, support vector machine and lightweight neural network.

7. A low permeability gas well productivity prediction device, characterized in that: include: A data acquisition module, used to acquire low permeability gas well data; A data preprocessing module is used to preprocess the acquired low-permeability gas well data to form a multi-source heterogeneous data set; Multi-model stacking architecture module, used to build multi-model stacking architecture; The multi-model stacking architecture includes a first-layer sub-model and a second-layer sub-model; the first-layer sub-model includes three sub-models: random forest, support vector machine and lightweight neural network; the output of the first-layer sub-model is used as the input of the second-layer sub-model, and in the inter-layer optimization process, the performance of the first-layer sub-model is evaluated by K-fold cross validation, and its prediction result is used as the data source for training the second-layer sub-model; A model training module uses the multi-source heterogeneous data set to train the multi-model stacking architecture to obtain a prediction model; A result output module, outputting prediction results based on the prediction model; The dynamic weight optimization module is used to dynamically adjust the weights of the three sub-models of random forest, support vector machine and lightweight neural network in the second-layer sub-model by using the genetic algorithm optimization method, including: generating multiple random weight combinations based on the initialization population method, each combination represents the weight of a different sub-model in the second-layer sub-model; The root mean square error and determination coefficient are used as fitness functions to evaluate the quality of each weight combination. The weight combination that reaches the threshold fitness will enter the next round; the weight combination that reaches the threshold fitness is selected for crossover, and the weight parts of different individuals are combined to generate a new weight combination; by randomly adjusting the weight values ​​of some individuals, mutations are introduced to maintain the diversity of the population; after each crossover and mutation operation, the new weight combination will undergo fitness evaluation, and the weight combination that meets the requirements will be selected into the next generation and gradually optimized to eventually find an optimal or near-optimal weight score.

Images