Sun F10.7 index and flare intensity combined prediction method and system

Abstract

The invention relates to the technical field of solar activity monitoring and forecasting, and particularly discloses a solar F10.7 index and flare intensity combined prediction method and system, and the method comprises the steps: collecting the time sequence data of the solar F10.7 index observed in multiple continuous days in the past, carrying out the data preprocessing, and obtaining the primary feature vector sequence of the F10.7 index; collecting time sequence data of solar flare intensity observed in multiple continuous days in the past, and performing data preprocessing to obtain a solar flare primary feature vector sequence; performing cross fusion on the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fusion feature vector sequence of F10.7 index information and flare information; and predicting the values of the solar F10.7 index and the solar flare intensity in the next few days through a recursive equation according to the fusion feature vector sequence.

Description

Technical Field

[0001] The present invention relates to the technical field of solar activity monitoring and forecasting, and in particular to a method and system for jointly predicting the solar F10.7 index and flare intensity. Background Art

[0002] Changes in solar activity dominate changes in the near-Earth space environment. The solar F10.7 index is an important indicator for measuring the intensity of solar activity. In order to accurately predict disturbances in the Earth's space environment, it is necessary to know the specific value of the solar F10.7 index in advance. For example, in many models that predict changes in the Earth's atmosphere and ionosphere (such as the MSIS model and the IRI model), the F10.7 index needs to be input as a driving parameter. Therefore, the prediction of the F10.7 index is one of the important contents of solar activity monitoring and forecasting.

[0003] Flares are a short-term, sudden, intense energy release phenomenon on the surface of the sun. They are an intense manifestation of solar activity and have a significant impact on the Earth's space environment. According to the energy released, the intensity of flares is generally divided into five levels: A, B, C, M, and X. Flares with intensity levels M and X can often release a large amount of energy in the form of electromagnetics and high-energy particles in a short period of time, causing significant adverse effects on various systems of humans in space and on the ground. For example, flare outbreaks can lead to an increase in the electron concentration in the D region of the ionosphere, resulting in a significant increase in the ionosphere's absorption of radio waves, which can cause serious interference or even interruption to shortwave communications. In addition, flare outbreaks are often accompanied by high-energy proton events and coronal mass ejections. The higher the intensity level of the flare outbreak, the greater the possibility of a significant increase in the high-energy proton flux in the Earth's orbit, posing a serious threat to spacecraft and astronauts in orbit. At the same time, the scale of the coronal mass ejection accompanying the flare is often larger, which in turn triggers a series of Earth space environment disturbance events such as geomagnetic storms, high-energy electron storms, and ionospheric storms. Therefore, the intensity level of solar flare eruptions has become an important indicator of whether the near-Earth space environment is safe. In order to avoid and mitigate the adverse effects of solar flare eruptions, flare intensity prediction is another important part of solar activity monitoring and forecasting in addition to the F10.7 index prediction.

[0004] At present, the physical mechanism of solar activity has not been fully studied, and it is impossible to establish an accurate physical and mathematical model to accurately predict the solar F10.7 index and the intensity of flares in the future. Their predictions are mainly based on artificial experience, statistics and machine learning methods. Methods based on machine learning, especially neural networks, are increasingly used in the prediction of the F10.7 index and flares. However, the current forecasts of the solar F10.7 index and flare intensity are all based on independent forecasting methods. Studies have shown that the fluctuations of the solar F10.7 index and the solar flare eruption activities are both external manifestations caused by the evolution of the internal physical field of the sun, and there is an implicit correlation between the changes of the two. Existing forecasting methods fail to fully utilize this implicit correlation between the two. Summary of the invention

[0005] To achieve the purpose of the present invention, the present application provides a method for jointly predicting the solar F10.7 index and flare intensity, comprising:

[0006] Step S1: collect the solar F10.7 index time series data observed over multiple consecutive days in the past, and perform data preprocessing to obtain the F10.7 index primary feature vector sequence;

[0007] Step S2: collect the time series data of flare intensity observed over multiple consecutive days in the past, and perform data preprocessing to obtain the flare primary feature vector sequence;

[0008] Step S3: cross-fusing the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence of the F10.7 index information and the flare information;

[0009] Step S4: predicting the values ​​of the solar F10.7 index and flare intensity in the next few days through a recursive equation based on the fused feature vector sequence.

[0010] In some specific embodiments, in step S1, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the solar F10.7 index time series data; in step S2, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the flare intensity time series data.

[0011] In some specific embodiments, step S3 includes:

[0012] Step S31: using a transformation matrix and a bias vector to transform each F10.7 index primary eigenvector to the vector space where the flare primary eigenvector sequence is located, and using a selu function to calculate a nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector;

[0013] Step S32: normalizing the nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector to obtain the contribution of the flare primary eigenvector to the F10.7 index primary eigenvector;

[0014] Step S33: using the contribution of the flare primary feature vector to the F10.7 index primary feature vector as a weight coefficient, linearly combining the flare primary feature vectors to obtain the F10.7 index secondary feature vector that incorporates the flare information;

[0015] Step S34: Concatenate the F10.7 index primary feature vector and the F10.7 index secondary feature vector to determine the F10.7 index tertiary feature vector.

[0016] In some specific embodiments, step S3 further includes:

[0017] Step S35: using a transformation matrix and a bias vector to transform each flare primary eigenvector to the vector space where the F10.7 index tertiary eigenvector is located, and using a tanh function to calculate a nonlinear correlation factor between the flare primary eigenvector and the F10.7 index tertiary eigenvector;

[0018] Step S36: normalizing the nonlinear correlation factor between the primary flare feature vector and the third-level flare feature vector of the F10.7 index to obtain the contribution of the third-level flare feature vector of the F10.7 index to the primary flare feature vector;

[0019] Step S37: using the contribution of the F10.7 index three-level feature vector to the flare primary feature vector as a weight coefficient, linearly combining the F10.7 index three-level feature vector to obtain the flare secondary feature vector that incorporates the F10.7 index feature information;

[0020] Step S38: concatenate the flare primary feature vector and the flare secondary feature vector to obtain a flare tertiary feature vector.

[0021] In some specific embodiments, step S3 further includes:

[0022] The three-level feature vector of the F10.7 index and the three-level feature vector of the flare are transformed through a layer of fully connected neural network to obtain an intermediate feature vector, and the intermediate feature vector is spliced ​​with the solar F10.7 index time series data and the flare intensity time series data to obtain a fused feature vector sequence.

[0023] In some specific embodiments, in step S4, the recursive equation is determined according to the following formula:

[0024]

[0025] H t =o t *tanh(C t )

[0026] f t =W 5 o t

[0027] X t =softmax(W 6 o t )

[0028] Among them, f t The sun index is F10.7; X t is the flare intensity; H t is the implicit feature vector generated in the recursive process; i t is the weight vector of the current time feature information; C t is the weighted value of all historical feature information up to the current time; σ is the sigmoid function; is the intermediate eigenvector; is the feature information vector of the current time; t is the current time; W 1 ,W 2 ,W 3 ,W 4 ,W 5 ,W 6 are the values ​​of the transformation matrix respectively; b 1 ,b 2 ,b 3 ,b 3 ,b 4 are the values ​​of the vector parameters respectively.

[0029] To achieve the same invention purpose, the present application also provides a solar F10.7 index and flare intensity joint prediction system, including:

[0030] Solar F10.7 index time series collection module: used to collect solar F10.7 index time series data observed over multiple consecutive days in the past, and perform data preprocessing to obtain the primary feature vector sequence of the F10.7 index;

[0031] Flare intensity time series data collection module: used to collect the time series data of flare intensity observed over multiple consecutive days in the past, and perform data preprocessing to obtain the flare primary feature vector sequence;

[0032] Feature vector cross-fusion module: used for cross-fusion of the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence of the F10.7 index information and the flare information;

[0033] Solar F10.7 index and flare intensity prediction module: used to predict the values ​​of solar F10.7 index and flare intensity in the next few days through a recursive equation based on the fused feature vector sequence.

[0034] In some specific embodiments, the feature vector cross fusion module is used to perform the following steps:

[0035] Step S31: using a transformation matrix and a bias vector to transform each F10.7 index primary eigenvector to the vector space where the flare primary eigenvector sequence is located, and using a selu function to calculate a nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector;

[0036] Step S32: normalizing the nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector to obtain the contribution of the flare primary eigenvector to the F10.7 index primary eigenvector;

[0037] Step S33: using the contribution of the flare primary feature vector to the F10.7 index primary feature vector as a weight coefficient, linearly combining the flare primary feature vectors to obtain the F10.7 index secondary feature vector that incorporates the flare information;

[0038] Step S34: Concatenate the F10.7 index primary feature vector and the F10.7 index secondary feature vector to determine the F10.7 index tertiary feature vector.

[0039] In some specific embodiments, the feature vector cross fusion module is further used to perform the following steps:

[0040] Step S35: using a transformation matrix and a bias vector to transform each flare primary eigenvector to the vector space where the F10.7 index tertiary eigenvector is located, and using a tanh function to calculate a nonlinear correlation factor between the flare primary eigenvector and the F10.7 index tertiary eigenvector;

[0041] Step S36: normalizing the nonlinear correlation factor between the primary flare feature vector and the third-level flare feature vector of the F10.7 index to obtain the contribution of the third-level flare feature vector of the F10.7 index to the primary flare feature vector;

[0042] Step S37: using the contribution of the F10.7 index three-level feature vector to the flare primary feature vector as a weight coefficient, linearly combining the F10.7 index three-level feature vector to obtain the flare secondary feature vector that incorporates the F10.7 index feature information;

[0043] Step S38: concatenate the flare primary feature vector and the flare secondary feature vector to obtain a flare tertiary feature vector.

[0044] In some specific embodiments, the feature vector cross fusion module is further used to perform the following steps:

[0045] The three-level feature vector of the F10.7 index and the three-level feature vector of the flare are transformed through a layer of fully connected neural network to obtain an intermediate feature vector, and the intermediate feature vector is spliced ​​with the solar F10.7 index time series data and the flare intensity time series data to obtain a fused feature vector sequence.

[0046] Beneficial effects of the above technical solution:

[0047] The present application provides a method and system for jointly predicting the solar F10.7 index and flare intensity, which can fully explore the implicit connection between the solar F10.7 index and flares, and use the historical observation data of many days before the forecast day to predict the solar F10.7 index and flare intensity in the next few days. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without creative work.

[0049] Figure 1 A schematic flow chart of a method for jointly predicting the solar F10.7 index and flare intensity provided by an embodiment of the present invention;

[0050] Figure 2 A schematic diagram of the structure of a solar F10.7 index and flare intensity joint prediction system provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0052] Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements with the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0053] Embodiment 1

[0054] An embodiment of the present invention provides a method for jointly predicting the solar F10.7 index and the flare intensity, referring to Figure 1 As shown, including:

[0055] Step S1: Collect the solar F10.7 index time series data f observed over several consecutive days in the past i (i=1……T), and perform data preprocessing to obtain the F10.7 index primary feature vector sequence F i (i=1……T).

[0056] Specifically, by constructing a joint prediction model for the solar F10.7 index and flares, and using historical observation data from T days (T=81) before the forecast day, the solar F10.7 index and flare intensity for the next three days are predicted.

[0057] Step S2: Collect the time series data x of the flare intensity observed over multiple consecutive days in the past j (j=1...T), and perform data preprocessing to obtain the flare primary feature vector sequence X j (j=1…T).

[0058] In a specific embodiment of the present invention, in step S1, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the solar F10.7 index time series data; in step S2, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the flare intensity time series data.

[0059] Step S3: Cross-fusing the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence Z of the F10.7 index information and the flare information j (j=1…T).

[0060] In a specific embodiment of the present invention, step S3 includes:

[0061] Step S31: Use transformation matrix W x and the bias vector b x Each F10.7 index primary eigenvector F i Transformed into the flare primary feature vector sequence X jThe vector space where the F10.7 index primary eigenvector F is located is calculated using the selu function i With the flare primary eigenvector X j The nonlinear correlation factor β i,j :

[0062] β i,j = selu(W x F i +X j +b x )

[0063] Step S32: The nonlinear correlation factor β between the primary eigenvector of the F10.7 index and the primary eigenvector of the flare i,j Perform normalization processing to obtain the primary characteristic vector X of the flare j The primary eigenvector F of the F10.7 index i Contribution

[0064]

[0065] Step S33: Taking the primary eigenvector X of the flare j Contribution to the primary eigenvector of F10.7 index The flare primary feature vectors are linearly combined to obtain the F10.7 index secondary feature vector that incorporates the flare information.

[0066]

[0067] Step S34: The F10.7 index primary feature vector F i (i=1...T) and the F10.7 index secondary eigenvector Splicing to determine the third-level eigenvector of F10.7 index

[0068]

[0069] In a specific embodiment of the present invention, step S3 further includes:

[0070] Step S35: Use the transformation matrix W f and the bias vector b f Each flare primary eigenvector X j Transform the third-level eigenvector to the F10.7 index The flare primary eigenvector X is calculated using the tanh function j The third-level eigenvector with the F10.7 index The nonlinear correlation factor α j,i :

[0071]

[0072] Step S36: Nonlinear correlation factor α between the primary eigenvector of the flare and the tertiary eigenvector of the F10.7 index j,i Perform normalization processing to obtain the three-level feature vector of the F10.7 index The primary eigenvector X of the flare j Contribution

[0073]

[0074] Step S37: Using the F10.7 index to obtain the third-level feature vector Contribution to the primary eigenvector of the flare The three-level feature vectors of the F10.7 index are linearly combined to obtain the secondary feature vector of the flare that incorporates the feature information of the F10.7 index.

[0075]

[0076] Step S38: Substituting the primary eigenvector X j The flare secondary eigenvector Splicing to get the three-level feature vector of the flare

[0077]

[0078] In a specific embodiment of the present invention, step S3 further includes:

[0079] The F10.7 index three-level eigenvector and the flare level three eigenvector After transformation through a layer of fully connected neural network, the intermediate feature vector is obtained The intermediate feature vector The solar F10.7 index time series data f j (j=1...T) and the time series data x of the flare intensity j (j=1...T) concatenate to obtain the fused feature vector sequence Z j (j=1...T), where

[0080] In a specific embodiment of the present invention, in step S4, the fused feature vector Z from 1 to T days is used. j(j=1...T), recursively predict the solar F10.7 index f for days T+1, T+2 and T+3 t and the flare intensity X t , the recursive equation is determined according to the following formula:

[0081]

[0082] H t =o t *tanh(C t )

[0083] f t =W 5 o t

[0084] X t =softmax(W 6 o t )

[0085] Among them, f t The sun index is F10.7; X t is the flare intensity; H t is the implicit feature vector generated in the recursive process; i t is the weight vector of the current time feature information; C t is the weighted value of all historical feature information up to the current time; σ is the sigmoid function; is the intermediate eigenvector; is the feature information vector of the current time; t is the current time; W 1 ,W 2 ,W 3 ,W 4 ,W 5 ,W 6 are the values ​​of the transformation matrix respectively; b 1 ,b 2 ,b 3 ,b 3 ,b 4 are the values ​​of the vector parameters respectively. Step S4: according to the fusion feature vector sequence Z j (j=1…T) The recursive equation is used to predict the values ​​of the solar F10.7 index and flare intensity in the next few days.

[0086] The present application provides a method and system for jointly predicting the solar F10.7 index and flare intensity, which can fully explore the implicit connection between the solar F10.7 index and flares, and use the historical observation data of many days before the forecast day to predict the solar F10.7 index and flare intensity in the next few days.

[0087] Embodiment 2

[0088] An embodiment of the present invention provides a solar F10.7 index and flare intensity joint prediction system, referring to Figure 2 As shown, including:

[0089] The solar F10.7 index time series collection module 10 is used to collect the solar F10.7 index time series data observed over the past several consecutive days, and perform data preprocessing to obtain the F10.7 index primary feature vector sequence;

[0090] Flare intensity time series data collection module 20: used to collect the time series data of flare intensity observed for multiple consecutive days in the past, and perform data preprocessing to obtain the flare primary feature vector sequence;

[0091] A feature vector cross-fusion module 30 is used to cross-fuse the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence of the F10.7 index information and the flare information;

[0092] The solar F10.7 index and flare intensity prediction module 40 is used to predict the values ​​of the solar F10.7 index and flare intensity in the next few days through a recursive equation according to the fused feature vector sequence.

[0093] In a specific embodiment of the present invention, in the solar F10.7 index time series collection module 10, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the solar F10.7 index time series data; in the flare intensity time series data collection module 20, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the flare intensity time series data.

[0094] In a specific embodiment of the present invention, the feature vector cross fusion module 30 is used to perform the following steps:

[0095] Step S31: using a transformation matrix and a bias vector to transform each F10.7 index primary eigenvector to the vector space where the flare primary eigenvector sequence is located, and using a selu function to calculate a nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector;

[0096] Step S32: normalizing the nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector to obtain the contribution of the flare primary eigenvector to the F10.7 index primary eigenvector;

[0097] Step S33: using the contribution of the flare primary feature vector to the F10.7 index primary feature vector as a weight coefficient, linearly combining the flare primary feature vectors to obtain the F10.7 index secondary feature vector that incorporates the flare information;

[0098] Step S34: Concatenate the F10.7 index primary feature vector and the F10.7 index secondary feature vector to determine the F10.7 index tertiary feature vector.

[0099] In a specific embodiment of the present invention, the feature vector cross fusion module 30 is further used to perform the following steps:

[0100] Step S35: using a transformation matrix and a bias vector to transform each flare primary eigenvector to the vector space where the F10.7 index tertiary eigenvector is located, and using a tanh function to calculate a nonlinear correlation factor between the flare primary eigenvector and the F10.7 index tertiary eigenvector;

[0101] Step S36: normalizing the nonlinear correlation factor between the primary flare feature vector and the third-level flare feature vector of the F10.7 index to obtain the contribution of the third-level flare feature vector of the F10.7 index to the primary flare feature vector;

[0102] Step S37: using the contribution of the F10.7 index three-level feature vector to the flare primary feature vector as a weight coefficient, linearly combining the F10.7 index three-level feature vector to obtain the flare secondary feature vector that incorporates the F10.7 index feature information;

[0103] Step S38: concatenate the flare primary feature vector and the flare secondary feature vector to obtain a flare tertiary feature vector.

[0104] In a specific embodiment of the present invention, the feature vector cross fusion module 30 is further used to perform the following steps:

[0105] The three-level feature vector of the F10.7 index and the three-level feature vector of the flare are transformed through a layer of fully connected neural network to obtain an intermediate feature vector, and the intermediate feature vector is spliced ​​with the solar F10.7 index time series data and the flare intensity time series data to obtain a fused feature vector sequence.

[0106] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

[0107] Each embodiment in this specification is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referenced to each other. The embodiments of the present invention are described with reference to the flowcharts and / or block diagrams of the methods, terminal devices (systems), and computer program products according to the embodiments of the present invention. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing terminal device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing terminal device generate instructions for implementing the functions in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing terminal device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing terminal device, so that a series of operation steps are executed on the computer or other programmable terminal device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable terminal device provide for implementing the process in the process. Figure 1 A process or multiple processes and / or boxes Figure 1The steps of the functions specified in one or more boxes. Although the preferred embodiments of the embodiments of the present invention have been described, those skilled in the art may make additional changes and modifications to these embodiments once they know the basic creative concept. Therefore, the attached claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the embodiments of the present invention. Finally, it should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variant thereof are intended to cover non-exclusive inclusion, so that the process, method, article or terminal device including a series of elements includes not only those elements, but also other elements that are not explicitly listed, or also includes elements inherent to such process, method, article or terminal device. In the absence of more restrictions, the elements defined by the sentence "including one..." do not exclude the existence of other identical elements in the process, method, article or terminal device including the elements.

[0108] The method and device provided by the present invention are introduced in detail above. Specific examples are used in this article to illustrate the principle and implementation mode of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation mode and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

[0109] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", "one specific embodiment" or "some examples" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for jointly predicting the solar F10.7 index and flare intensity, characterized in that: include: Step S1: collect the solar F10.7 index time series data observed over multiple consecutive days in the past, and perform data preprocessing to obtain the F10.7 index primary feature vector sequence; Step S2: collect the time series data of flare intensity observed over multiple consecutive days in the past, and perform data preprocessing to obtain the flare primary feature vector sequence; Step S3: cross-fusing the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence of the F10.7 index information and the flare information; Step S4: Based on the fused feature vector sequence, the values ​​of the solar F10.7 index and flare intensity in the next few days are predicted by a recursive equation.

2. The method for jointly predicting the solar F10.7 index and flare intensity according to claim 1, characterized in that: In step S1, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the solar F10.7 index time series data; in step S2, a four-layer bidirectional long short-term memory neural network and a single-layer fully connected neural network are used to preprocess the flare intensity time series data.

3. The method for jointly predicting the solar F10.7 index and flare intensity according to claim 1, characterized in that: Step S3 includes: Step S31: using a transformation matrix and a bias vector to transform each F10.7 index primary eigenvector to the vector space where the flare primary eigenvector sequence is located, and using a selu function to calculate a nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector; Step S32: normalizing the nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector to obtain the contribution of the flare primary eigenvector to the F10.7 index primary eigenvector; Step S33: using the contribution of the flare primary feature vector to the F10.7 index primary feature vector as a weight coefficient, linearly combining the flare primary feature vectors to obtain the F10.7 index secondary feature vector that incorporates the flare information; Step S34: Concatenate the F10.7 index primary feature vector and the F10.7 index secondary feature vector to determine the F10.7 index tertiary feature vector.

4. The method for jointly predicting the solar F10.7 index and flare intensity according to claim 3, characterized in that: Step S3 also includes: Step S35: using a transformation matrix and a bias vector to transform each flare primary eigenvector to the vector space where the F10.7 index tertiary eigenvector is located, and using a tanh function to calculate a nonlinear correlation factor between the flare primary eigenvector and the F10.7 index tertiary eigenvector; Step S36: normalizing the nonlinear correlation factor between the primary flare feature vector and the third-level flare feature vector of the F10.7 index to obtain the contribution of the third-level flare feature vector of the F10.7 index to the primary flare feature vector; Step S37: using the contribution of the F10.7 index three-level feature vector to the flare primary feature vector as a weight coefficient, linearly combining the F10.7 index three-level feature vector to obtain the flare secondary feature vector that incorporates the F10.7 index feature information; Step S38: concatenate the flare primary feature vector and the flare secondary feature vector to obtain a flare tertiary feature vector.

5. The method for jointly predicting the solar F10.7 index and flare intensity according to claim 4, characterized in that: Step S3 also includes: The three-level feature vector of the F10.7 index and the three-level feature vector of the flare are transformed through a layer of fully connected neural network to obtain an intermediate feature vector, and the intermediate feature vector is spliced ​​with the solar F10.7 index time series data and the flare intensity time series data to obtain a fused feature vector sequence.

6. The method for jointly predicting the solar F10.7 index and flare intensity according to claim 3, characterized in that: In step S4, the recursive equation is determined according to the following formula: H t =o t *tanh(C t ) F t =W5·o t X t =softmax(W6·o t ) Among them, f t The sun index is F10.7; X t is the flare intensity; H t is the implicit feature vector generated in the recursive process; i t is the weight vector of the current time feature information; C t is the weighted value of all historical feature information up to the current time; σ is the sigmoid function; is the intermediate eigenvector; is the feature information vector of the current time; t is the current time; W1, W2, W3, W4, W5, W6 are the values ​​of the transformation matrix respectively; b1, b2, b3, b4, b5 are the values ​​of the vector parameters respectively.

7. A solar F10.7 index and flare intensity joint prediction system, characterized in that: include: Solar F10.7 index time series collection module: used to collect solar F10.7 index time series data observed over multiple consecutive days in the past, and perform data preprocessing to obtain the primary feature vector sequence of the F10.7 index; Flare intensity time series data collection module: used to collect the time series data of flare intensity observed over multiple consecutive days in the past, and perform data preprocessing to obtain the flare primary feature vector sequence; Feature vector cross-fusion module: used for cross-fusion of the F10.7 index primary feature vector sequence and the flare primary feature vector sequence to obtain a fused feature vector sequence of the F10.7 index information and the flare information; Solar F10.7 index and flare intensity prediction module: used to predict the values ​​of solar F10.7 index and flare intensity in the next few days through a recursive equation based on the fused feature vector sequence.

8. The solar F10.7 index and flare intensity joint prediction system according to claim 7, characterized in that: The feature vector cross fusion module is used to perform the following steps: Step S31: using a transformation matrix and a bias vector to transform each F10.7 index primary eigenvector to the vector space where the flare primary eigenvector sequence is located, and using a selu function to calculate a nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector; Step S32: normalizing the nonlinear correlation factor between the F10.7 index primary eigenvector and the flare primary eigenvector to obtain the contribution of the flare primary eigenvector to the F10.7 index primary eigenvector; Step S33: using the contribution of the flare primary feature vector to the F10.7 index primary feature vector as a weight coefficient, linearly combining the flare primary feature vectors to obtain the F10.7 index secondary feature vector that incorporates the flare information; Step S34: Concatenate the F10.7 index primary feature vector and the F10.7 index secondary feature vector to determine the F10.7 index tertiary feature vector.

9. The solar F10.7 index and flare intensity joint prediction system according to claim 8, characterized in that: The feature vector cross fusion module is also used to perform the following steps: Step S35: using a transformation matrix and a bias vector to transform each flare primary eigenvector to the vector space where the F10.7 index tertiary eigenvector is located, and using a tanh function to calculate a nonlinear correlation factor between the flare primary eigenvector and the F10.7 index tertiary eigenvector; Step S36: normalizing the nonlinear correlation factor between the primary flare feature vector and the third-level flare feature vector of the F10.7 index to obtain the contribution of the third-level flare feature vector of the F10.7 index to the primary flare feature vector; Step S37: using the contribution of the F10.7 index three-level feature vector to the flare primary feature vector as a weight coefficient, linearly combining the F10.7 index three-level feature vector to obtain the flare secondary feature vector that incorporates the F10.7 index feature information; Step S38: concatenate the flare primary feature vector and the flare secondary feature vector to obtain a flare tertiary feature vector.

10. The solar F10.7 index and flare intensity joint prediction system according to claim 9, characterized in that: The feature vector cross fusion module is also used to perform the following steps: The three-level feature vector of the F10.7 index and the three-level feature vector of the flare are transformed through a layer of fully connected neural network to obtain an intermediate feature vector, and the intermediate feature vector is spliced ​​with the solar F10.7 index time series data and the flare intensity time series data to obtain a fused feature vector sequence.

Images