Geomagnetic Ap index prediction method and system, storage medium and product
By combining the stability selection algorithm and the XGBoost model, a high-performance geomagnetic Ap index prediction model was constructed, which solved the problems of low accuracy and poor credibility of existing prediction methods, achieved more accurate and reliable prediction effects, and provided strong support for magnetic storm early warning.
Patent Information
- Application Number
- CN202411829049.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-12
- Publication Date
- 2025-03-25
AI Technical Summary
The existing geomagnetic Ap index prediction methods have low accuracy and poor confidence, and the model has limited effect when dealing with complex geomagnetic activities.
Using the geomagnetic Ap index prediction method based on stability selection and XGBoost model, the feature importance is calculated by obtaining and preprocessing the sample data set of the spatial environment, using the stability selection algorithm to determine the subset of the feature, and building the XGBoost model for training to achieve prediction.
It improves the accuracy, stability and reliability of geomagnetic Ap index prediction, enhances the prediction performance and generalization capabilities of the model, and provides more reliable magnetic storm warning prediction support.
Smart Images
Figure CN119691622A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of geomagnetic storm prediction, and in particular to a geomagnetic Ap index prediction method, device, storage medium and product based on stability selection and XGBoost model. Background Art
[0002] Geomagnetic storms, or geomagnetic storms for short, are phenomena of violent global disturbances of the Earth's magnetic field caused by solar surface activity. The impact of geomagnetic storms is present in many aspects. For example, in the fields of communication and navigation, geomagnetic storms can affect radio transmission and satellite navigation systems, leading to increased errors and, in severe cases, communication interruptions. In the power system, geomagnetic storms can increase the induced current in the transmission lines, causing overload of the power system, and thus causing power system failures. Therefore, early warning and prediction of geomagnetic storms are of great significance.
[0003] Since there is a strong correlation between the intensity of geomagnetic storm events and the geomagnetic Ap index, the geomagnetic Ap index is generally used to characterize the state of geomagnetic activity. The geomagnetic Ap index is an index of the global all-day geomagnetic disturbance intensity, ranging from 0 to 400. The larger the geomagnetic Ap index, the greater the magnitude of the geomagnetic disturbance.
[0004] However, some researchers currently use the autoregressive integrated moving average model to predict the future Ap index based on the past Ap index and solar wind speed. This method has limited predictive ability for transient events and does not work well during the more random stages of solar activity. The model only relies on the quality and accuracy of historical data and solar wind speed, so changes in the data may affect the prediction results. Some researchers use linear regression and multivariate regression models to analyze historical Ap data to predict the future Ap index. Although this method performs well in calm periods, it is limited in predicting complex geomagnetic activity.
[0005] In summary, the existing geomagnetic Ap index prediction methods either use limited features, or the data used often include a large number of features with complex feature relationships and noise, and not all features are useful for geomagnetic Ap index prediction, resulting in the constructed geomagnetic Ap index prediction model having low prediction performance and generalization ability, and the accuracy and credibility of the prediction results are also poor. Summary of the invention
[0006] The present invention provides a geomagnetic Ap index prediction method, system, storage medium and product, which are used to solve the problems of low accuracy and poor credibility in existing prediction methods.
[0007] In a first aspect, the present invention provides a method for predicting the geomagnetic Ap index, the method comprising: Acquire a sample data set of the space environment and perform preprocessing; wherein the sample data set includes multiple groups of sample data, each group of sample data includes a feature set related to geomagnetic changes and a historical geomagnetic Ap index value corresponding to the feature set; Calculating the feature importance of each feature in the feature set to obtain a feature importance value of each feature; Based on the feature importance value of each feature, a feature subset is determined using a stability selection algorithm; The feature subset is used as the input feature of the XGBoost model, the historical geomagnetic Ap index value corresponding to the feature subset is used as the label data output by the XGBoost model, the XGBoost model is trained, and a geomagnetic Ap index prediction model is obtained; The geomagnetic Ap index prediction model is used to predict the geomagnetic Ap index.
[0008] Furthermore, the feature set includes: longitude, latitude, F10.7 index, geomagnetic component, and Dst index.
[0009] Furthermore, the preprocessing includes: missing value processing, outlier processing, and encoding of feature categories.
[0010] Furthermore, obtaining the feature importance value includes: Dividing the sample data set into a plurality of subsets by cross-validation; each subset includes one or more groups of sample data; In each training, a subset of the data is randomly selected as the validation set, and the remaining subset is used as the training set. Cross-validation and model training are repeated multiple times, and the initial feature importance value of each feature obtained in each training is recorded. Calculate multiple initial feature importance values of a feature to obtain the feature importance value of the feature; and so on to obtain the feature importance value of each feature.
[0011] Furthermore, the feature subset is determined by: By randomly sampling in the sample data set and training to evaluate the stability of the feature, the feature stability value of each feature is obtained; Based on the feature importance value and the feature stability value of each feature, a feature subset is determined.
[0012] Further, based on the feature importance value and the feature stability value of each feature, determining the feature subset further includes: Filtering a first feature set from the feature set according to a preset feature importance value threshold; Filtering a second feature set from the feature set according to a preset feature stability value threshold; The overlapping features of the first feature set and the second feature set are combined as a feature subset.
[0013] Furthermore, training the XGBoost model includes: During the training process, the XGBoost model parameters are adjusted by continuously learning the relationship between the input features and the output label data; The geomagnetic Ap index prediction model was evaluated using a cross-validation method, and the XGBoost model parameters were optimized based on the evaluation results.
[0014] In a second aspect, the present invention provides a geomagnetic Ap index prediction system, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of any of the above methods.
[0015] In a third aspect, the present invention provides a computer-readable storage medium having a computer program / instruction stored thereon, wherein the computer program / instruction, when executed by a processor, implements the steps of any of the above-mentioned methods.
[0016] In a fourth aspect, the present invention provides a computer program product, comprising a computer program / instruction, characterized in that the computer program / instruction implements the steps of any of the above-mentioned methods when executed by a processor.
[0017] In general, the present invention provides a geomagnetic Ap index prediction method, system, storage medium and product. The technical solution conceived by the present invention can achieve the following beneficial effects compared with the prior art: First, the present invention combines the stability selection algorithm with the XGBoost model to fully mine the feature information in the data, constructs a high-performance, highly reliable, and predictive-effective geomagnetic Ap index prediction model, and identifies stable and important features through random sampling and repeated modeling. It can effectively reduce the possible randomness and noise in the feature selection process, realize the prediction of the geomagnetic Ap index, and improve the accuracy, stability, reliability, and prediction performance of the geomagnetic Ap index prediction, providing reliable support for the early warning prediction of magnetic storms.
[0018] Second, the present invention uses a cross-validation method to evaluate the XGBoost model for different data sets. It has the ability to process high-dimensional data, can automatically process missing values, and effectively captures the interaction effects between nonlinear features and complex feature relationships through a tree structure model to verify the generalization ability and prediction accuracy of the geomagnetic Ap index prediction model on unknown data; at the same time, based on the feedback information in actual applications, the model is continuously iteratively updated, optimized and improved, and the gradient boosting mechanism enables the model to better reduce errors in each iteration, thereby improving the prediction ability and applicability of the geomagnetic Ap index prediction model.
[0019] Third, the XGBoost model provided by the present invention can flexibly adjust parameters to adapt to different data characteristics and prediction requirements, so that the geomagnetic Ap index prediction model can perform well in different geomagnetic activity stages (such as quiet periods and active periods), thereby improving the applicability of the geomagnetic Ap index prediction model. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0021] Figure 1 It is a method flow diagram of a geomagnetic Ap index prediction method, system, storage medium and product provided by the present invention; Figure 2 It is a schematic diagram of a sample data set of a geomagnetic Ap index prediction method, system, storage medium and product provided by the present invention; Figure 3 It is a characteristic subset schematic diagram of a geomagnetic Ap index prediction method, system, storage medium and product provided by the present invention. DETAILED DESCRIPTION
[0022] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings and embodiments of the present invention. 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 ordinary technicians in this field without creative work belong to the scope of protection of the present invention.
[0023] It should be noted that, in the description of the embodiments of the present invention, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a method or step including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such method or step. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the method or step including the element.
[0024] In the first aspect, the present invention proposes a method for predicting the geomagnetic Ap index, such as Figure 1 As shown, the method includes: Step S101: Acquire a sample data set of a spatial environment and perform preprocessing.
[0025] The sample data set includes multiple groups of sample data, and each group of sample data includes a feature set related to geomagnetic changes and a historical geomagnetic Ap index value corresponding to the feature set.
[0026] The feature set in the sample data is the basis for constructing the geomagnetic Ap index prediction model. The features are related to geomagnetic changes. As an embodiment, the feature set includes: longitude, latitude, F10.7 index, geomagnetic component, and Dst index. The geomagnetic component represents the physical quantity of the direction and magnitude of the earth's magnetic field, including By, Bz, and Bx, which respectively represent the vectors of the geomagnetic field intensity at a point on the surface in the x component, y component, and z component.
[0027] It should be noted that the sample data set can be directly obtained from the website of the Space Environment Forecast Center, and a set of sample data corresponds to one day.
[0028] Preprocessing includes: missing value processing, outlier processing, and encoding feature categories.
[0029] Missing value processing is to fill or delete missing data points to ensure data integrity; outlier processing is to modify or delete abnormal data points to ensure data normality; encoding the category of features is to convert the categorical data of the features into a digital form that the model can understand, such as one-hot encoding or label encoding.
[0030] How to select features that have a significant impact on the target variable from a large number of features is very important, which is related to whether the performance and interpretability of the prediction model can be improved, and then directly affects the accuracy and credibility of the prediction results. Therefore, the present invention calculates the feature importance and feature stability, selects the most important feature subset, and uses this as the model input feature to train the model to obtain a prediction model.
[0031] Step S102: Calculate the feature importance of each feature in the feature set to obtain the feature importance value of each feature.
[0032] It should be noted that there are many ways to calculate feature importance, such as recursive feature elimination, random forest algorithm, LASSO regression method, etc.
[0033] Preferably, the present invention uses a cross-validation method to calculate the feature importance of each feature in the feature set to obtain the feature importance value of each feature. Cross-validation is a commonly used method for evaluating model performance. The present invention can more comprehensively evaluate model performance and reduce the deviation introduced by different data divisions through multiple cross-validations.
[0034] More specifically, cross-validation is used to divide the sample data set into multiple subsets; each subset includes one or more groups of sample data; in each training, a part of the subset is randomly selected as the validation set, and the remaining subset is used as the training set, and cross-validation and model training are repeated multiple times, and the initial feature importance value of each feature obtained in each training is recorded; multiple initial feature importance values of a feature are calculated to obtain the feature importance value of the feature; and so on, the feature importance value of each feature is obtained.
[0035] In other words, through multiple trainings, feature importance values of different features can be obtained to ensure a more stable feature importance value ranking. For each feature, multiple feature importance values corresponding to the multiple trainings are statistically calculated to obtain the final feature importance value of the feature for subsequent feature selection.
[0036] Among them, the statistical calculation method can be to calculate the average value, median, or variance of multiple feature importance values; finally, the calculated result is used as the final feature importance value, so as to ensure that the obtained feature importance value ranking is more robust and reliable.
[0037] Step S103: Based on the feature importance value of each feature, a feature subset is determined using a stability selection algorithm. In other words, the most important feature subset is determined using a stability selection algorithm.
[0038] As a feature selection method, stability selection has the advantages of reducing the risk of overfitting, improving the generalization ability of the model, retaining the most relevant feature information, being applicable to high-dimensional data, and being general and flexible. The core idea is that if a feature is selected under different feature subsets and model settings, then the feature is likely to be a stable and important feature.
[0039] The stability selection algorithm includes an algorithm based on random repeated sampling, that is, by randomly sampling and training in the sample data set to evaluate the stability of the feature, and according to the algorithm results, the final feature subset is determined.
[0040] It should be noted that the stability selection algorithm is described as a feature selection method for high-dimensional data, which can not only reduce the overfitting problem in the case of high-dimensional data and small sample size, but also reduce the data dimension by evaluating the importance of features on different data subsets, retaining the most representative and discriminative feature information, which helps to improve the accuracy, efficiency and interpretability of the prediction model.
[0041] As an embodiment, a method for obtaining a feature subset includes: obtaining a feature stability value of each feature by randomly sampling in a sample data set and training to evaluate the stability of the feature; and determining a feature subset based on the feature importance value and the feature stability value of each feature.
[0042] Specifically, this dataset includes variables: Date, Area, coronal hole thickness (CH_Th), PchT30, PchT60, PchT90, PchN30, PchN60, PchN90, By and Bz components of the interplanetary magnetic field (By, Bz), solar wind speed (V), and Ap index (Ap). These variables are used together to study and predict geomagnetic activity, among which the Ap index is the main prediction target, which is used to quantify the intensity of geomagnetic activity.
[0043] Among them, the steps of data preprocessing include: first, read and preliminarily understand the data structure, then check and process missing values, and choose to delete or fill missing values. Next, select and standardize the features needed for analysis and modeling according to the characteristics of the data set. Ensure that the target variable (Ap index) corresponds to the feature data, and finally save the processed data set to provide clean and standardized data for subsequent modeling and analysis. After evaluating the importance of features through the stability selection method combined with the XGBoost model, it was found that the solar wind speed (V) was confirmed as a feature that made an important contribution to the prediction results. Although the By component (By) of the interplanetary magnetic field is of moderate importance, it still has a certain impact on the prediction. This method of combining stability selection and XGBoost model effectively screens out the most critical features for Ap index prediction, helps to improve the prediction accuracy and stability of the model, and provides a reliable data basis for further analysis and application.
[0044] It should be noted that feature stability is evaluated through multiple random resampling and feature selection, thereby reducing the overfitting of the model to noise and redundant features. Feature selection is repeated on different feature subsets, and the frequency of each feature being selected is counted, and the probability of being selected is the feature stability value.
[0045] As a specific embodiment of the present invention: First, the sample dataset is separated by features and labels.
[0046] Then, the sample data set is randomly resampled multiple times. In each resampling, a portion of features and labels are randomly selected from the sample data set, and a portion of features are extracted from the sample data set with replacement each time to form a new feature set. Among them, resampling can use methods such as cross-validation and bootstrap. The more repetitions, the more accurate the feature stability can be evaluated.
[0047] Next, a feature selection algorithm is applied to each resampled feature set for feature selection, wherein the feature selection method can be Lasso regression, logistic regression, etc., or a tree model-based method such as random forest or gradient boosting tree.
[0048] Finally, repeat the process several times to obtain multiple feature selection results, and count the frequency of each feature being selected, which is the feature stability value. Generally speaking, features with a feature stability value higher than a certain threshold are considered to be important features with high stability. Therefore, based on the preset feature stability value threshold, a feature subset with high stability can be screened out to better adapt to different data distributions and prediction model settings, thereby improving the generalization performance of the prediction model.
[0049] As an embodiment, based on the feature importance value and the feature stability value of each feature, determining the feature subset further includes: According to a preset feature importance value threshold, a first feature set is screened out from the feature set; according to a set feature stability value threshold, a second feature set is screened out from the feature set; and overlapping features of the first feature set and the second feature set are combined as a feature subset.
[0050] It should be noted that, in addition to setting the threshold, features can also be sorted from large to small according to feature importance values or feature stability values, and filtered based on the sorting.
[0051] As another embodiment, the method for obtaining a feature subset may also include: filtering out a first feature set from the feature set according to a preset feature importance value threshold; obtaining a feature stability value of each feature by randomly sampling and training in the first data set to evaluate the stability of the feature; filtering out a second feature set from the feature set according to a preset feature stability value threshold, and using the second feature set as the feature subset.
[0052] Step S104: Using the feature subset as the input feature of the XGBoost model, using the historical geomagnetic Ap index value corresponding to the feature subset as the label data output by the XGBoost model, training the XGBoost model, and obtaining a geomagnetic Ap index prediction model.
[0053] The XGBoost model (eXtreme Gradient Boosting) is an integrated learning algorithm used to solve regression and classification problems. It is based on the gradient boosting framework and combines the advantages of multiple weak learners through boosting algorithms and decision tree integration, and can effectively handle complex data relationships.
[0054] Furthermore, training the XGBoost model includes: during the training process, adjusting the XGBoost model parameters by continuously learning the relationship between the input features and the output label data to improve the prediction ability of the geomagnetic Ap index prediction model.
[0055] In addition, the cross-validation method can be used to evaluate the geomagnetic Ap index prediction model, and the XGBoost model parameters can be optimized based on the evaluation results.
[0056] Specifically, the performance and generalization ability of the geomagnetic Ap index prediction model are evaluated using methods such as cross-validation, and the evaluation indicators include mean square error, accuracy, recall rate, F1 score, etc. According to the evaluation results, the XGBoost model parameters are optimized to further improve the performance of the geomagnetic Ap index prediction model. Finally, an independent test set is used to verify the optimized XGBoost model to evaluate the generalization ability and prediction accuracy of the model on unknown data. In addition, grid search, random search and other methods can be used to search for the optimal parameter combination.
[0057] Step S105: using the geomagnetic Ap index prediction model to predict the geomagnetic Ap index.
[0058] That is, the geomagnetic Ap index can be predicted by inputting the current feature subset data into the geomagnetic Ap index prediction model.
[0059] It should be noted that when a new sample data set is obtained, the geomagnetic Ap index prediction model can also be iteratively updated using the new sample data set, and the geomagnetic Ap index prediction model can be continuously optimized and improved to improve the prediction ability and applicability of the model.
[0060] As a specific embodiment of the present invention, Figure 2 As shown, it is a schematic diagram of a sample data set of a geomagnetic Ap index prediction method, system, storage medium and product provided by the present invention. The sample data set is 3136 rows × 13 columns of data, 3136 rows represent 3136 groups of sample data; 13 columns represent 12 columns of feature sets and 1 column of geomagnetic Ap index values.
[0061] First, the sample data set (12 columns of data and 1 column of geomagnetic Ap index values) was repeatedly cross-validated and trained, and the feature importance values of each training were recorded; through multiple trainings, the feature importance values of different features can be obtained.
[0062] Next, based on the feature importance value of each feature, a feature subset is determined using a stability selection algorithm, such as Figure 3 As shown, the most important feature subsets are obtained: geomagnetic component By, geomagnetic component Bz and solar wind speed V. That is, among these 12 columns of data, only geomagnetic component By, geomagnetic component Bz and solar wind speed V have some connection with the geomagnetic Ap index value, and only 628 sets of sample data among 3136 sets of sample data can be used for training.
[0063] Finally, the geomagnetic component By, geomagnetic component Bz and solar wind speed V data in 628 groups of sample data are used as the input features of the XGBoost model, and the geomagnetic Ap index values corresponding to the features in 628 groups of sample data are used as the label data output by the XGBoost model. The XGBoost model is trained and optimized, and finally the geomagnetic Ap index prediction model is obtained.
[0064] In a second aspect, the present invention provides a geomagnetic Ap index prediction system, comprising a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of any of the above methods. The technical features of the system are consistent with those of the above methods and are not described in detail herein.
[0065] In a third aspect, the present invention provides a computer-readable storage medium having a computer program / instruction stored thereon, wherein the computer program / instruction, when executed by a processor, implements the steps of any of the above methods. The technical features of the storage medium are consistent with the technical features of the above method and are not described in detail herein.
[0066] In a fourth aspect, the present invention provides a computer program product, including a computer program / instruction, characterized in that the computer program / instruction implements the steps of any of the above methods when executed by a processor. The technical features of the computer program product are consistent with the technical features of the above method, and are not repeated here.
[0067] In summary, the present invention establishes a robust and high-performance geomagnetic Ap index prediction model based on stability selection and XGBoost model, realizes the prediction of geomagnetic Ap index, improves the accuracy, stability and reliability of geomagnetic Ap index prediction, and provides reliable support for early warning prediction of magnetic storms.
[0068] It should be noted that, for the above-mentioned various embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the present application is not limited by the described order of actions, because according to the present application, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present application.
[0069] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0070] In the several embodiments provided in this application, it should be understood that the disclosed method or system can be implemented in other ways. For example, the above-described embodiments are only schematic, and the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
[0071] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0072] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
[0073] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable memory. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or the whole or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a memory, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application.
[0074] Those skilled in the art will appreciate that all or part of the various circuits in the above embodiments may be completed by entering a program to instruct related hardware, and the program may be stored in a computer-readable memory, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.
[0075] The above is only an exemplary embodiment of the present disclosure, and the scope of the present disclosure cannot be limited thereto. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the specification and practicing the disclosure here, those skilled in the art will easily think of the implementation scheme of the present disclosure. This application is intended to cover any modification, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the technical field not recorded in the present disclosure. The description and examples are regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.
[0076] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0077] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.
Claims
1. A geomagnetic Ap index prediction method, characterized in that: The method comprises: Acquire a sample data set of the space environment and perform preprocessing; wherein the sample data set includes multiple groups of sample data, each group of sample data includes a feature set related to geomagnetic changes and a historical geomagnetic Ap index value corresponding to the feature set; Calculating the feature importance of each feature in the feature set to obtain a feature importance value of each feature; Based on the feature importance value of each feature, a feature subset is determined using a stability selection algorithm; The feature subset is used as an input feature of the XGBoost model, the historical geomagnetic Ap index value corresponding to the feature subset is used as label data output by the XGBoost model, the XGBoost model is trained, and a geomagnetic Ap index prediction model is obtained; The geomagnetic Ap index prediction model is used to predict the geomagnetic Ap index.
2. A geomagnetic Ap index prediction method according to claim 1, characterized in that: The feature set includes: longitude, latitude, F10.7 index, geomagnetic component, and Dst index.
3. A geomagnetic Ap index prediction method according to claim 1, characterized in that: The preprocessing includes: missing value processing, outlier processing, and encoding the feature category.
4. A geomagnetic Ap index prediction method according to claim 1, characterized in that: The feature importance value is obtained by: Dividing the sample data set into a plurality of subsets by cross-validation; each subset includes one or more groups of sample data; In each training, a subset of the data is randomly selected as the validation set, and the remaining subset is used as the training set. Cross-validation and model training are repeated multiple times, and the initial feature importance value of each feature obtained in each training is recorded. Calculate multiple initial feature importance values of a feature to obtain the feature importance value of the feature; and so on to obtain the feature importance value of each feature.
5. A geomagnetic Ap index prediction method according to claim 1, characterized in that: The method of determining the feature subset includes: By randomly sampling in the sample data set and training to evaluate the stability of the feature, the feature stability value of each feature is obtained; Based on the feature importance value and the feature stability value of each feature, a feature subset is determined.
6. A geomagnetic Ap index prediction method according to claim 5, characterized in that: Based on the feature importance value and the feature stability value of each feature, determining to obtain a feature subset further includes: Filtering a first feature set from the feature set according to a preset feature importance value threshold; Filtering a second feature set from the feature set according to a preset feature stability value threshold; The overlapping features of the first feature set and the second feature set are combined as a feature subset.
7. A geomagnetic Ap index prediction method according to claim 1, characterized in that: Training an XGBoost model involves: During the training process, the XGBoost model parameters are adjusted by continuously learning the relationship between the input features and the output label data; The geomagnetic Ap index prediction model was evaluated using a cross-validation method, and the XGBoost model parameters were optimized based on the evaluation results.
8. A geomagnetic Ap index prediction system, comprising a memory, a processor and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.
9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.
10. A computer program product comprising a computer program / instructions, characterized in that When the computer program / instruction is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.