Tropical cyclone prediction method, device, equipment and medium

Abstract

The invention provides a tropical cyclone prediction method and device, equipment and a medium, and the method comprises the steps: obtaining a historical sunspot data set and a historical tropical cyclone data set, so as to obtain a sample associated data pair set, and for any sample associated data pair, the sample associated data pair comprises sample sunspot data and sample tropical cyclone data; obtaining a to-be-trained candidate cyclone prediction model, and performing model training on the candidate cyclone prediction model through the sample association data pair set until the training is finished, thereby obtaining a trained target cyclone prediction model; and obtaining to-be-identified target sunspot data, and based on the target cyclonic prediction model and the target sunspot data, obtaining target predicted cyclonic data in the prediction time range. The method achieves the prediction of the tropical cyclone data through the sunspot data, improves the prediction efficiency and prediction accuracy of the tropical cyclone data, and provides an accurate data support for the disaster prevention and reduction of coastal regions.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a tropical cyclone prediction method, device, equipment and medium. Background Art

[0002] In reality, the earth's climate change has an important impact on people's work and life. Among them, the climate change caused by tropical cyclone climate in coastal areas may cause some types of disasters in coastal areas.

[0003] In related technologies, the relevant regularity analysis can be performed through historical tropical cyclone events that occurred within the historical time range to predict tropical cyclone data, but the accuracy is poor.

[0004] Therefore, it is very important to achieve accurate prediction of tropical cyclones. Summary of the invention

[0005] The purpose of this application is to solve one of the technical problems in the above technology at least to a certain extent.

[0006] The first aspect of the present application provides a tropical cyclone prediction method, comprising: obtaining a historical sunspot data set and a historical tropical cyclone data set to obtain a sample associated data pair set, wherein, for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data; obtaining a candidate cyclone prediction model to be trained, and performing model training on the candidate cyclone prediction model through the sample associated data pair set until the training is completed, thereby obtaining a trained target cyclone prediction model; obtaining target sunspot data to be identified, and obtaining target predicted cyclone data within a prediction time range based on the target cyclone prediction model and the target sunspot data.

[0007] The tropical cyclone prediction method provided in the first aspect of the present application also has the following technical features, including:

[0008] According to one embodiment of the present application, the historical sunspot data set and the historical tropical cyclone data set are obtained to obtain a set of sample associated data pairs, wherein, for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data, including: obtaining candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set; for any historical tropical cyclone data, obtaining candidate correlation parameters between the historical tropical cyclone data and each candidate associated sunspot data; according to each candidate correlation parameter, obtaining a sample tropical cyclone data set from the historical tropical cyclone data set, and obtaining sample sunspot data of each sample tropical cyclone data from the sample tropical cyclone data set; according to each sample tropical cyclone data and the sample sunspot data of each sample tropical cyclone data, the set of sample associated data pairs is obtained.

[0009] According to one embodiment of the present application, obtaining candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set includes: obtaining a first historical time period of each historical tropical cyclone data in the historical tropical cyclone data set; obtaining a second historical time period of each historical sunspot data in the historical sunspot data set; and obtaining candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set based on each first historical time period and each second historical time period.

[0010] According to an embodiment of the present application, obtaining candidate associated sunspot datasets of each historical tropical cyclone data from the historical sunspot dataset according to each first historical time period and each second historical time period includes: for any historical tropical cyclone data, obtaining the first historical year to which the first historical time period of the historical tropical cyclone data belongs, and the second historical year of the previous year of the first historical year; obtaining the first year sunspot dataset of the first historical year from the historical sunspot dataset according to each second historical time period to obtain the first monthly average sunspot number in the first historical year; obtaining the second year sunspot dataset of the second historical year from the historical sunspot dataset according to each second historical time period to obtain the annual average sunspot number and the second monthly average sunspot number in the second historical year; obtaining the candidate associated sunspot dataset of the historical tropical cyclone data according to the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number.

[0011] According to one embodiment of the present application, for any historical tropical cyclone data, obtaining candidate correlation parameters between the historical tropical cyclone data and each candidate associated sunspot data includes: for any candidate associated sunspot data, obtaining the Pearson correlation coefficient between the historical tropical cyclone data and the candidate associated sunspot data to obtain the candidate correlation parameters between the historical tropical cyclone data and the candidate associated sunspot data.

[0012] According to an embodiment of the present application, the method of obtaining a sample tropical cyclone data set from the historical tropical cyclone data set based on each candidate correlation parameter, and obtaining sample sunspot data of each sample tropical cyclone data from the sample tropical cyclone data set, includes: for any candidate correlation parameter, in response to the candidate correlation parameter being greater than or equal to a preset correlation parameter threshold, determining that the historical tropical cyclone data corresponding to the candidate correlation parameter is the sample tropical cyclone data, and determining that the candidate correlated sunspot data corresponding to the candidate correlation parameter is the sample sunspot data corresponding to the sample tropical cyclone data.

[0013] According to one embodiment of the present application, the candidate cyclone prediction model to be trained is obtained, and the candidate cyclone prediction model is trained by the set of sample associated data pairs until the training is completed to obtain a trained target cyclone prediction model, including: for any sample associated data pair, a training sample is obtained according to the sample sunspot data and the sample tropical cyclone data in the sample associated data pair; the candidate cyclone prediction model is trained according to the training sample to obtain a training loss of the candidate cyclone prediction model; the candidate cyclone prediction model is parameter-adjusted according to the training loss, and the next training sample is returned to continue training the candidate cyclone prediction model after the parameter adjustment, until the training is completed to obtain the trained target cyclone prediction model.

[0014] According to one embodiment of the present application, the target sunspot data to be identified is acquired, and based on the target cyclone prediction model and the target sunspot data, the target predicted cyclone data within the prediction time range is obtained, including: inputting the target sunspot data into the target cyclone prediction model; extracting data features of the target sunspot data according to the target cyclone prediction model, so as to output the target predicted cyclone data within the prediction time range according to the data features.

[0015] The second aspect of the present application provides a tropical cyclone prediction device, comprising: an acquisition module, used to acquire historical sunspot data sets and historical tropical cyclone data sets to obtain a set of sample associated data pairs, wherein, for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data; a training module, used to acquire a candidate cyclone prediction model to be trained, and perform model training on the candidate cyclone prediction model through the set of sample associated data pairs until the training is completed to obtain a trained target cyclone prediction model; a prediction module, used to acquire target sunspot data to be identified, and obtain target predicted cyclone data within a prediction time range based on the target cyclone prediction model and the target sunspot data.

[0016] The second aspect of the present application provides a tropical cyclone prediction device, which also has the following technical features, including:

[0017] According to one embodiment of the present application, the acquisition module is also used to: obtain candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set; for any historical tropical cyclone data, obtain candidate correlation parameters between the historical tropical cyclone data and each candidate associated sunspot data; obtain a sample tropical cyclone data set from the historical tropical cyclone data set according to each candidate correlation parameter, and obtain sample sunspot data of each sample tropical cyclone data from the sample tropical cyclone data set; obtain the sample associated data pair set according to each sample tropical cyclone data and the sample sunspot data of each sample tropical cyclone data.

[0018] According to one embodiment of the present application, the acquisition module is also used to: obtain the first historical time period of each historical tropical cyclone data in the historical tropical cyclone data set; obtain the second historical time period of each historical sunspot data in the historical sunspot data set; and obtain candidate associated sunspot data sets for each historical tropical cyclone data from the historical sunspot data set based on each first historical time period and each second historical time period.

[0019] According to one embodiment of the present application, the acquisition module is also used to: for any historical tropical cyclone data, obtain the first historical year to which the first historical time period of the historical tropical cyclone data belongs, and the second historical year of the previous year of the first historical year; according to each second historical time period, obtain the first year sunspot data set of the first historical year from the historical sunspot data set to obtain the first monthly average sunspot number in the first historical year; according to each second historical time period, obtain the second year sunspot data set of the second historical year from the historical sunspot data set to obtain the annual average sunspot number and the second monthly average sunspot number in the second historical year; according to the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number, obtain the candidate associated sunspot data set of the historical tropical cyclone data.

[0020] According to one embodiment of the present application, the acquisition module is also used to: for any candidate associated sunspot data, obtain the Pearson correlation coefficient between the historical tropical cyclone data and the candidate associated sunspot data to obtain the candidate correlation parameter between the historical tropical cyclone data and the candidate associated sunspot data.

[0021] According to one embodiment of the present application, the acquisition module is also used to: for any candidate correlation parameter, in response to the candidate correlation parameter being greater than or equal to a preset correlation parameter threshold, determine that the historical tropical cyclone data corresponding to the candidate correlation parameter is sample tropical cyclone data, and determine that the candidate correlated sunspot data corresponding to the candidate correlation parameter is the sample sunspot data corresponding to the sample tropical cyclone data.

[0022] According to one embodiment of the present application, the training module is also used to: for any sample-associated data pair, obtain a training sample based on the sample sunspot data and the sample tropical cyclone data in the sample-associated data pair; train the candidate cyclone prediction model based on the training sample to obtain the training loss of the candidate cyclone prediction model; adjust the parameters of the candidate cyclone prediction model based on the training loss, and return to obtain the next training sample to continue training the candidate cyclone prediction model with adjusted parameters until the training is completed to obtain the trained target cyclone prediction model.

[0023] According to one embodiment of the present application, the prediction module is also used to: input the target sunspot data into the target cyclone prediction model; extract data features of the target sunspot data according to the target cyclone prediction model, so as to output the target predicted cyclone data within the prediction time range according to the data features.

[0024] The third aspect embodiment of the present application provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the tropical cyclone prediction method provided in the first aspect of the present application.

[0025] The fourth aspect embodiment of the present application provides a non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to enable the computer to execute the tropical cyclone prediction method provided in the first aspect of the present application.

[0026] The fifth aspect of the present application provides a computer program product. When the instruction processor in the computer program product is executed, the tropical cyclone prediction method provided in the first aspect of the present application is executed.

[0027] The tropical cyclone prediction method and device provided in the present application obtain a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, obtain a candidate cyclone prediction model to be trained, and obtain corresponding sample associated data pairs from the set of sample associated data pairs, so as to perform model training of the candidate cyclone prediction model according to the sample associated data pairs until the training is completed, and obtain a trained target cyclone prediction model, and further, obtain target predicted cyclone data within the prediction time range according to the target sunspot data to be identified through the target cyclone prediction model. In the present application, the candidate cyclone prediction model is trained by sample associated data pairs including sample sunspot data and sample tropical cyclone data, so that the trained target cyclone prediction model realizes the prediction of tropical cyclone data through sunspot data, improves the prediction efficiency and prediction accuracy of tropical cyclone data, and reduces the computational cost of tropical cyclone prediction compared to the tropical cyclone prediction realized in the tropical cyclone value mode in the related art, provides accurate data support for disaster prevention and mitigation in coastal areas, and optimizes the prediction method and prediction effect of tropical cyclones.

[0028] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through the practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0030] Figure 1 A schematic diagram of a tropical cyclone prediction method according to an embodiment of the present application;

[0031] Figure 2A schematic diagram of a flow chart of a tropical cyclone prediction method according to another embodiment of the present application;

[0032] Figure 3 A schematic diagram of a flow chart of a tropical cyclone prediction method according to another embodiment of the present application;

[0033] Figure 4 A schematic diagram of the structure of a tropical cyclone prediction device according to an embodiment of the present application;

[0034] Figure 5 It is a block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] Embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having 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 application, and should not be construed as limiting the present application.

[0036] The tropical cyclone prediction method, apparatus, device and medium of the embodiments of the present application are described below with reference to the accompanying drawings.

[0037] Figure 1 FIG. 1 is a flow chart of a tropical cyclone prediction method according to an embodiment of the present application. Figure 1 As shown, the method includes:

[0038] S101, acquiring a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data.

[0039] In an embodiment of the present application, a data set consisting of sunspot data collected within a set historical time range can be marked as a historical sunspot data set, and a set of tropical cyclone data collected within a set historical time range can be marked as a historical tropical cyclone data set.

[0040] The historical collection time ranges of the historical sunspot dataset and the historical tropical cyclone dataset may be the same historical time period.

[0041] Optionally, pre-set sample filtering conditions for sample sunspot data and sample filtering conditions for sample tropical cyclone data can be obtained. In this scenario, part of the sunspot data that matches the filtering conditions can be filtered out from the historical sunspot data set according to the sample filtering conditions for the sample sunspot data, and the part of the sunspot data can be determined as the sample sunspot data.

[0042] And, according to the sample screening condition of the sample tropical cyclone data, some tropical cyclone data are screened out from the historical tropical cyclone data set, and the part of tropical cyclone data is determined as the sample tropical cyclone data.

[0043] Optionally, for any sample sunspot data, sample tropical cyclone data that has a correlation with the sample sunspot data is screened out from all the sample tropical cyclone data, and the sample sunspot data and the sample tropical cyclone data are combined to obtain a combined correlated data pair, which can be marked as a sample correlated data pair.

[0044] Furthermore, the set consisting of each sample-associated data pair is marked as a sample-associated data pair set.

[0045] S102, obtaining a candidate cyclone prediction model to be trained, and performing model training on the candidate cyclone prediction model through a set of sample association data until the training is completed, thereby obtaining a trained target cyclone prediction model.

[0046] In the disclosed embodiments, tropical cyclone prediction may be achieved using a trained tropical cyclone prediction model, wherein the model to be trained may be marked as a candidate cyclone prediction model.

[0047] Optionally, the sample association data pairs used in the current training model training round can be obtained from the sample association data pair set, and the corresponding training samples can be obtained based on the sample association data pairs. Further, the training samples are input into the candidate cyclone prediction model to obtain the output results of the candidate cyclone prediction model.

[0048] Optionally, the training loss of the candidate cyclone prediction model in the current training round is obtained based on the output result, and the model parameters of the candidate cyclone prediction model are adjusted and optimized based on the training loss, and sample-associated data pairs continue to be obtained from the sample-associated data pair set as training samples, and the adjusted and optimized candidate cyclone prediction model continues to be trained until the training is completed, and the model obtained after the training is determined as the trained target cyclone prediction model.

[0049] S103, obtaining target sunspot data to be identified, and obtaining target predicted cyclone data within a prediction time range based on the target cyclone prediction model and the target sunspot data.

[0050] In an embodiment of the present application, the sunspot data used in tropical cyclone prediction can be marked as target sunspot data to be identified, wherein the target sunspot data can be obtained by monitoring the sunspot activity state and collecting it based on a preset data collection frequency.

[0051] Optionally, the collected target sunspot data can be input into a trained target cyclone prediction model, and the characteristic information of the target sunspot data can be extracted through the target cyclone prediction model. The tropical cyclone within the prediction time range is predicted based on the extracted characteristic information, and the corresponding prediction result is obtained, which is the target predicted cyclone data.

[0052] It should be noted that the target predicted cyclone data may include the occurrence frequency data corresponding to tropical cyclone events within the prediction time range, and may also include cyclone status information of each tropical cyclone event within the prediction time range, etc., which is not specifically limited here.

[0053] The tropical cyclone prediction method proposed in the present application obtains a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, obtains a candidate cyclone prediction model to be trained, and obtains corresponding sample associated data pairs from the set of sample associated data pairs, so as to perform model training of the candidate cyclone prediction model according to the sample associated data pairs until the training is completed, and obtain a trained target cyclone prediction model, and further, obtains target predicted cyclone data within the prediction time range according to the target sunspot data to be identified through the target cyclone prediction model. In the present application, the candidate cyclone prediction model is trained by sample associated data pairs including sample sunspot data and sample tropical cyclone data, so that the trained target cyclone prediction model realizes the prediction of tropical cyclone data through sunspot data, improves the prediction efficiency and prediction accuracy of tropical cyclone data, and reduces the computational cost of tropical cyclone prediction compared to the tropical cyclone prediction realized in the tropical cyclone value mode in the related art, provides accurate data support for disaster prevention and mitigation in coastal areas, and optimizes the prediction method and prediction effect of tropical cyclones.

[0054] In the above embodiment, the prediction of tropical cyclones can also be combined with Figure 2 Further understanding, Figure 2 FIG. 1 is a flow chart of a tropical cyclone prediction method according to another embodiment of the present application. Figure 2 As shown, the method includes:

[0055] S201, acquiring a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data.

[0056] Optionally, candidate associated sunspot datasets for each historical tropical cyclone data are obtained from historical sunspot datasets.

[0057] In an embodiment of the present application, there is a certain degree of correlation between the activity of sunspots and tropical cyclone events on the earth. In this scenario, for any historical tropical cyclone data, part of the sunspot data that has a certain degree of correlation with the historical tropical cyclone data can be obtained from the historical sunspot data set, and the set consisting of this part of the sunspot data can be marked as a candidate associated sunspot data set that has an association with the historical tropical cyclone data.

[0058] As a possible implementation method, candidate associated sunspot data sets of each historical tropical cyclone data can be obtained from the time dimension, wherein the first historical time period of each historical tropical cyclone data in the historical tropical cyclone data set can be obtained, and the second historical time period of each historical sunspot data in the historical sunspot data set can be obtained.

[0059] In an embodiment of the present application, for any historical tropical cyclone data, the historical occurrence time period of the historical tropical cyclone event corresponding to the historical tropical cyclone data can be obtained, and the historical occurrence time is the first historical time period of the historical tropical cyclone data, thereby obtaining the first historical time period of each historical tropical cyclone data.

[0060] Furthermore, for any historical sunspot data, the historical activity time period of the historical sunspot activity corresponding to the historical sunspot data can be obtained, and the historical activity time period is the second historical time period of the historical sunspot data, thereby obtaining the second historical time period of each historical sunspot data.

[0061] Optionally, according to each first historical time period and each second historical time period, a candidate associated sunspot data set of each historical tropical cyclone data is obtained from the historical sunspot data set.

[0062] Among them, for any historical tropical cyclone data, the first historical year to which the first historical time period of the historical tropical cyclone data belongs and the second historical year of the previous year of the first historical year are obtained.

[0063] In an embodiment of the present application, for any historical tropical cyclone data, the natural year to which the first historical time period corresponding to the historical tropical cyclone data belongs can be marked as the first historical year to which the historical tropical cyclone data belongs.

[0064] And, the previous year of the first historical year is marked as the corresponding second historical year.

[0065] Optionally, according to each second historical time period, a first year sunspot dataset of a first historical year is obtained from the historical sunspot dataset to obtain a first monthly average sunspot number in the first historical year.

[0066] In an embodiment of the present application, part of the second historical time period belonging to the first historical year can be obtained from each second historical time period to obtain the historical sunspot data of each of the part of the second historical time period, and the part of the historical sunspot data can be marked as the first year sunspot data set of the first historical year.

[0067] In this scenario, the first year's sunspot data set can be algorithmically processed according to the monthly average sunspot number algorithm in the relevant technology, so as to obtain the corresponding monthly average sunspot number according to the algorithm processing result, and the sunspot number can be marked as the first monthly average sunspot number corresponding to the first year's sunspot data set in the first historical year.

[0068] Optionally, according to each second historical time period, a second year sunspot dataset for a second historical year is obtained from the historical sunspot dataset to obtain an average annual sunspot number and a second monthly average sunspot number for the second historical year.

[0069] In an embodiment of the present application, a portion of the second historical time period belonging to the second historical year can be obtained from each second historical time period to obtain the historical sunspot data of each of the portion of the second historical time periods, thereby obtaining a second year sunspot data set composed of the portion of the historical sunspot data.

[0070] Furthermore, the second year's sunspot data set can be algorithmically processed according to the algorithm for the monthly average sunspot number in the relevant technology, so as to obtain the corresponding monthly average sunspot number according to the algorithm processing result, and the sunspot number can be marked as the second monthly average sunspot number corresponding to the second year's sunspot data set in the second historical year.

[0071] And, according to the algorithm of the annual average sunspot number in the relevant technology, the sunspot data set of the second year is processed by algorithm, so as to obtain the corresponding annual average sunspot number according to the algorithm processing result.

[0072] Optionally, a candidate associated sunspot data set of historical tropical cyclone data is obtained according to the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number.

[0073] In an embodiment of the present application, for any historical tropical cyclone data, the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number corresponding to the historical tropical cyclone data can be obtained, and the data set composed of the three can be determined as the candidate associated sunspot data set corresponding to the historical tropical cyclone data.

[0074] Optionally, for any historical tropical cyclone data, a candidate correlation parameter between the historical tropical cyclone data and each candidate correlated sunspot data is obtained.

[0075] In an embodiment of the present application, according to the correlation parameter acquisition algorithm in the relevant technology, any historical tropical cyclone data and each candidate associated sunspot data in the corresponding candidate associated sunspot data set can be algorithmically processed separately to obtain the correlation parameters between each candidate associated sunspot data set and the historical tropical cyclone data, which are marked as candidate correlation parameters.

[0076] As a possible implementation method, the candidate correlation parameter can be a Pearson correlation coefficient. In this scenario, for any candidate associated sunspot data, the Pearson correlation coefficient between the historical tropical cyclone data and the candidate associated sunspot data is obtained to obtain the candidate correlation parameter between the historical tropical cyclone data and the candidate associated sunspot data.

[0077] In the embodiment of the present application, the Pearson correlation coefficient between each candidate associated sunspot data set and the historical tropical cyclone data can be obtained according to the formula shown below:

[0078]

[0079] In the above formula, ρ X，Y represents the Pearson correlation coefficient between X and Y, X represents any historical tropical cyclone data, Y represents any candidate associated sunspot number in the candidate associated sunspot data set corresponding to X, μ X represents the expected value of X, μ Y represents the expected value of Y, σ x represents the standard deviation of X, σ y represents the standard deviation of Y.

[0080] In this scenario, the Pearson correlation coefficient between each candidate associated sunspot dataset and the historical tropical cyclone data can be obtained based on the above calculation formula and marked as the candidate correlation degree parameter between each candidate associated sunspot dataset and the historical tropical cyclone data.

[0081] Optionally, according to each candidate correlation parameter, a sample tropical cyclone data set is obtained from a historical tropical cyclone data set, and sample sunspot data of each sample tropical cyclone data is obtained from the sample tropical cyclone data set.

[0082] Among them, for any candidate correlation parameter, in response to the candidate correlation parameter being greater than or equal to a preset correlation parameter threshold, the historical tropical cyclone data corresponding to the candidate correlation parameter is determined to be the sample tropical cyclone data, and the candidate correlated sunspot data corresponding to the candidate correlation parameter is determined to be the sample sunspot data corresponding to the sample tropical cyclone data.

[0083] In an embodiment of the present application, for any candidate correlation parameter, if the candidate correlation parameter is greater than or equal to a preset correlation parameter threshold, it can be determined that there is a certain degree of strong correlation between the historical tropical cyclone data corresponding to the candidate correlation parameter and the candidate associated sunspot data.

[0084] In this scenario, when the historical tropical cyclone data is marked as sample tropical cyclone data, the candidate associated sunspot data can be marked as sample sunspot data that is strongly correlated with the sample tropical cyclone data.

[0085] It should be noted that the relevance parameter threshold corresponding to the candidate relevance parameter may be set to 0.4 or other values, which are not specifically limited here.

[0086] Optionally, a set of sample associated data pairs is obtained based on each sample tropical cyclone data and sample sunspot data of each sample tropical cyclone data.

[0087] In an embodiment of the present application, for any sample tropical cyclone data, sample sunspot data that is strongly correlated with the sample tropical cyclone data can be obtained, and the two can be combined into an associated data pair, which is a sample associated data pair including the sample tropical cyclone data and the sample sunspot data.

[0088] Furthermore, the set consisting of all sample-associated data pairs is marked as a sample-associated data pair set.

[0089] S202, obtaining a candidate cyclone prediction model to be trained, and performing model training on the candidate cyclone prediction model through a set of sample association data until the training is completed, thereby obtaining a trained target cyclone prediction model.

[0090] In an embodiment of the present application, a cyclone prediction model that needs to be trained can be marked as a candidate cyclone prediction model, wherein the candidate cyclone prediction model can be constructed based on a regression model, a decision tree model, or a neural network model, which is not specifically limited here.

[0091] Optionally, for any sample-associated data pair, a training sample is obtained according to sample sunspot data and sample tropical cyclone data in the sample-associated data pair.

[0092] In an embodiment of the present application, for any sample-associated data pair, the sample tropical cyclone data included in the sample-associated data pair can be used as label data for the sample sunspot data included in the sample-associated data pair, and then the sample sunspot data and the sample tropical cyclone data are processed according to the sample construction method in the relevant technology, so as to obtain sample data constructed by the two as training samples for the candidate cyclone prediction model.

[0093] Optionally, the candidate cyclone prediction model is trained according to the training samples to obtain the training loss of the candidate cyclone prediction model, and the parameters of the candidate cyclone prediction model are adjusted according to the training loss, and the next training sample is returned to continue training the candidate cyclone prediction model with adjusted parameters until the training is completed to obtain a trained target cyclone prediction model.

[0094] In an embodiment of the present application, the training samples can be input into the candidate cyclone prediction model to obtain the output results of the candidate cyclone prediction model based on the training samples, and the label data and output results in the training samples can be algorithmically processed according to the loss value acquisition algorithm in the relevant technology, so as to obtain the loss value generated by the candidate cyclone prediction model in the current training round according to the result of the algorithm processing, which is marked as the training loss of the current training round.

[0095] Furthermore, the model parameters of the candidate cyclone prediction model are adjusted according to the training loss, and the candidate cyclone prediction model after the parameter adjustment is obtained by the next training loss to continue the model training until the training is completed.

[0096] Among them, the corresponding model training end conditions can be set according to the training rounds. For the current round of model training, if the total training rounds accumulated by the model training of this round meet the model training end conditions, the training of the candidate cyclone prediction model can be ended, and the model obtained in the last training round can be determined as the trained target cyclone prediction model.

[0097] In addition, corresponding model training end conditions can be set according to the training output results. For the current round of model training, if the output results of the model training in this round meet the model training end conditions, the training of the candidate cyclone prediction model can be ended, and the model obtained in the last training round can be determined as the trained target cyclone prediction model.

[0098] S203, obtaining target sunspot data to be identified, and obtaining target predicted cyclone data within a prediction time range based on the target cyclone prediction model and the target sunspot data.

[0099] Optionally, the target sunspot data may be input into a target cyclone prediction model, and data features of the target sunspot data may be extracted according to the target cyclone prediction model, so as to output target predicted cyclone data within a prediction time range according to the data features.

[0100] In an embodiment of the present application, the sunspot data required for predicting tropical cyclone data within a prediction time range may be marked as target sunspot data, and the target sunspot data may be input into a trained target cyclone prediction model.

[0101] In this scenario, the target cyclone prediction model can be used to extract the data features carried in the target sunspot data, and then the tropical cyclone data can be predicted based on the extracted data features, thereby outputting the corresponding prediction results as the target predicted cyclone data within the prediction time range.

[0102] The tropical cyclone prediction method proposed in the present application trains a candidate cyclone prediction model by using sample associated data including sample sunspot data and sample tropical cyclone data, so that the trained target cyclone prediction model can realize the prediction of tropical cyclone data through sunspot data, thereby improving the prediction efficiency and prediction accuracy of tropical cyclone data. Compared with the tropical cyclone prediction realized in the tropical cyclone value mode in the related technology, the computational cost of tropical cyclone prediction is reduced, accurate data support is provided for disaster prevention and mitigation in coastal areas, and the prediction method and prediction effect of tropical cyclones are optimized.

[0103] To better understand the above embodiments, Figure 3 Further understanding, Figure 3 FIG. 4 is a flow chart of a tropical cyclone prediction method according to another embodiment of the present application.

[0104] like Figure 3 As shown, it can be obtained from Figure 3 The historical tropical cyclone dataset is obtained from the yearbook data shown in Figure 1, where Figure 3 The yearbook data shown may be typhoon yearbook data or historical tropical cyclone yearbook data, which is not specifically limited here.

[0105] like Figure 3 As shown in FIG. 1 , in the historical sunspot data set, the candidate associated sunspot data sets of each historical tropical cyclone data can be obtained according to each historical sunspot data and each historical tropical cyclone data, wherein the candidate associated sunspot data sets include Figure 3 The first monthly average sunspot number, the second monthly average sunspot number, and the annual average sunspot number are shown.

[0106] Optionally, according to the calculation formula of the Pearson correlation coefficient proposed in the above embodiment, algorithm processing can be performed on any historical tropical cyclone data, and the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number corresponding to the historical tropical cyclone data are respectively performed to obtain the Pearson correlation coefficients between the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number and the historical tropical cyclone data as candidate correlation parameters.

[0107] In this scenario, sunspot data that is strongly correlated with any historical tropical cyclone data can be obtained from a candidate associated sunspot data set corresponding to any historical tropical cyclone data according to the candidate correlation parameter, wherein the historical tropical cyclone data is determined to be sample tropical cyclone data, and the sunspot data is sample sunspot data that is strongly correlated with the sample tropical cyclone data.

[0108] Furthermore, sample associated data pairs including sample tropical cyclone data and corresponding sample sunspot data are formed, and then a sample associated data pair set consisting of all sample associated data pairs is obtained.

[0109] like Figure 3 As shown, the training samples of the candidate cyclone prediction model can be obtained according to the set of sample associated data pairs, and the candidate cyclone prediction model can be trained according to the training samples, wherein the model parameters of the candidate cyclone prediction model can be adjusted and optimized according to the training loss, and the next training sample can be returned to continue training the candidate cyclone prediction model with adjusted parameters until the training is completed, and a trained target cyclone prediction model can be obtained, and then the target prediction cyclone data within the prediction time range can be output according to the target cyclone prediction model based on the target sunspot data.

[0110] The tropical cyclone prediction method proposed in the present application trains a candidate cyclone prediction model by using sample associated data including sample sunspot data and sample tropical cyclone data, so that the trained target cyclone prediction model can realize the prediction of tropical cyclone data through sunspot data, thereby improving the prediction efficiency and prediction accuracy of tropical cyclone data. Compared with the tropical cyclone prediction realized in the tropical cyclone value mode in the related technology, the computational cost of tropical cyclone prediction is reduced, accurate data support is provided for disaster prevention and mitigation in coastal areas, and the prediction method and prediction effect of tropical cyclones are optimized.

[0111] Corresponding to the tropical cyclone prediction methods proposed in the above-mentioned embodiments, an embodiment of the present application also proposes a tropical cyclone prediction device. Since the tropical cyclone prediction device proposed in the embodiment of the present application corresponds to the tropical cyclone prediction methods proposed in the above-mentioned embodiments, the implementation method of the above-mentioned tropical cyclone prediction method is also applicable to the tropical cyclone prediction device proposed in the embodiment of the present application, and will not be described in detail in the following embodiments.

[0112] Figure 4 FIG. 1 is a schematic diagram of the structure of a tropical cyclone prediction device according to an embodiment of the present application. Figure 4 As shown, a tropical cyclone prediction device 400 includes an acquisition module 41, a training module 42 and a prediction module 43, wherein:

[0113] An acquisition module 41 is used to acquire a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data;

[0114] The training module 42 is used to obtain a candidate cyclone prediction model to be trained, and to perform model training on the candidate cyclone prediction model through a set of sample association data until the training is completed, thereby obtaining a trained target cyclone prediction model;

[0115] The prediction module 43 is used to obtain the target sunspot data to be identified, and obtain the target predicted cyclone data within the prediction time range based on the target cyclone prediction model and the target sunspot data.

[0116] In an embodiment of the present application, the acquisition module 41 is also used to: obtain candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set; obtain candidate correlation parameters between the historical tropical cyclone data and each candidate associated sunspot data for any historical tropical cyclone data; obtain a sample tropical cyclone data set from the historical tropical cyclone data set according to each candidate correlation parameter, and obtain sample sunspot data of each sample tropical cyclone data from the sample tropical cyclone data set; obtain a set of sample associated data pairs according to each sample tropical cyclone data and the sample sunspot data of each sample tropical cyclone data.

[0117] In an embodiment of the present application, the acquisition module 41 is also used to: obtain the first historical time period of each historical tropical cyclone data in the historical tropical cyclone data set; obtain the second historical time period of each historical sunspot data in the historical sunspot data set; and obtain candidate associated sunspot data sets for each historical tropical cyclone data from the historical sunspot data set based on each first historical time period and each second historical time period.

[0118] In an embodiment of the present application, the acquisition module 41 is also used to: for any historical tropical cyclone data, obtain the first historical year to which the first historical time period of the historical tropical cyclone data belongs, and the second historical year of the previous year of the first historical year; according to each second historical time period, obtain the first year sunspot data set of the first historical year from the historical sunspot data set to obtain the first monthly average sunspot number in the first historical year; according to each second historical time period, obtain the second year sunspot data set of the second historical year from the historical sunspot data set to obtain the annual average sunspot number and the second monthly average sunspot number in the second historical year; according to the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number, obtain the candidate associated sunspot data set of the historical tropical cyclone data.

[0119] In the embodiment of the present application, the acquisition module 41 is also used to: for any candidate associated sunspot data, obtain the Pearson correlation coefficient between the historical tropical cyclone data and the candidate associated sunspot data to obtain the candidate correlation parameter between the historical tropical cyclone data and the candidate associated sunspot data.

[0120] In an embodiment of the present application, the acquisition module 41 is also used to: for any candidate correlation parameter, in response to the candidate correlation parameter being greater than or equal to a preset correlation parameter threshold, determine that the historical tropical cyclone data corresponding to the candidate correlation parameter is sample tropical cyclone data, and determine that the candidate associated sunspot data corresponding to the candidate correlation parameter is sample sunspot data corresponding to the sample tropical cyclone data.

[0121] In an embodiment of the present application, the training module 42 is also used to: for any sample-associated data pair, obtain a training sample based on the sample sunspot data and the sample tropical cyclone data in the sample-associated data pair; train a candidate cyclone prediction model based on the training sample to obtain the training loss of the candidate cyclone prediction model; adjust the parameters of the candidate cyclone prediction model based on the training loss, and return to obtain the next training sample to continue training the candidate cyclone prediction model with the adjusted parameters until the training is completed to obtain a trained target cyclone prediction model.

[0122] In the embodiment of the present application, the prediction module 43 is also used to: input the target sunspot data into the target cyclone prediction model; extract the data features of the target sunspot data according to the target cyclone prediction model, and output the target predicted cyclone data within the prediction time range according to the data features.

[0123] The tropical cyclone prediction device proposed in the present application obtains a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, obtains a candidate cyclone prediction model to be trained, and obtains corresponding sample associated data pairs from the set of sample associated data pairs, so as to perform model training of the candidate cyclone prediction model according to the sample associated data pairs until the training is completed, and obtain a trained target cyclone prediction model. Further, the target cyclone prediction model is used to obtain target predicted cyclone data within the prediction time range according to the target sunspot data to be identified. In the present application, the candidate cyclone prediction model is trained by sample associated data pairs including sample sunspot data and sample tropical cyclone data, so that the trained target cyclone prediction model realizes the prediction of tropical cyclone data through sunspot data, improves the prediction efficiency and prediction accuracy of tropical cyclone data, and reduces the computational cost of tropical cyclone prediction compared to the tropical cyclone prediction realized in the tropical cyclone value mode in the related art, provides accurate data support for disaster prevention and mitigation in coastal areas, and optimizes the prediction method and prediction effect of tropical cyclones.

[0124] To achieve the above embodiments, the present application also provides an electronic device, a computer-readable storage medium and a computer program product.

[0125] Figure 5 A block diagram of an electronic device according to an embodiment of the present application, such as Figure 5 As shown, the device 500 includes a memory 51, a processor 52, and a computer program stored in the memory 51 and executable on the processor 52. When the processor 52 executes the program instructions, the execution is realized. Figures 1 to 3 A tropical cyclone prediction method of an embodiment.

[0126] In order to implement the above embodiments, the present application also provides a non-transitory computer-readable storage medium storing computer instructions, the computer instructions are used to enable the computer to execute Figures 1 to 3 A tropical cyclone prediction method of an embodiment.

[0127] In order to implement the above embodiment, the present application also provides a computer program product. When the instruction processor in the computer program product executes, Figures 1 to 3 A tropical cyclone prediction method of an embodiment.

[0128] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", 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 application. In this specification, the schematic representations of the above terms do 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. In addition, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of the different embodiments or examples, without contradiction.

[0129] In addition, the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In the description of this application, the meaning of "plurality" is at least two, such as two, three, etc., unless otherwise clearly and specifically defined.

[0130] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by technicians in the technical field to which the embodiments of the present application belong.

[0131] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by an instruction execution system, device or apparatus (such as a computer-based system, a system including a processor, or other system that can fetch instructions from an instruction execution system, device or apparatus and execute the instructions), or in combination with these instruction execution systems, devices or apparatuses. For the purpose of this specification, "computer-readable medium" can be any device that can contain, store, communicate, propagate or transmit a program for use by an instruction execution system, device or apparatus, or in combination with these instruction execution systems, devices or apparatuses. More specific examples of computer-readable media (a non-exhaustive list) include the following: an electrical connection with one or more wires (electronic device), a portable computer disk box (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disk read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing in other suitable ways if necessary, and then stored in a computer memory.

[0132] It should be understood that the various parts of the present application can be implemented by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, it can be implemented by any one of the following technologies known in the art or their combination: a discrete logic circuit having a logic gate circuit for implementing a logic function for a data signal, a dedicated integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0133] A person skilled in the art may understand that all or part of the steps in the method for implementing the above-mentioned embodiment may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiment.

[0134] In addition, each functional unit in each embodiment of the present application may be integrated into a processing module, or each unit may exist physically separately, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.

[0135] The storage medium mentioned above may be a read-only memory, a magnetic disk or an optical disk, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limiting the present application. A person of ordinary skill in the art may change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A tropical cyclone prediction method, characterized in that: The method comprises: Acquire a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein, for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data; Acquire a candidate cyclone prediction model to be trained, and perform model training on the candidate cyclone prediction model through the sample association data pair set until the training is completed, thereby obtaining a trained target cyclone prediction model; Target sunspot data to be identified is acquired, and based on the target cyclone prediction model and the target sunspot data, target predicted cyclone data within a prediction time range is obtained.

2. The method according to claim 1, characterized in that The acquiring of the historical sunspot data set and the historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data, including: Acquire candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data sets; For any historical tropical cyclone data, obtaining candidate correlation parameters between the historical tropical cyclone data and each candidate correlated sunspot data; According to each candidate correlation parameter, a sample tropical cyclone data set is obtained from the historical tropical cyclone data set, and sample sunspot data of each sample tropical cyclone data is obtained from the sample tropical cyclone data set; The sample associated data pair set is obtained according to each sample tropical cyclone data and the sample sunspot data of each sample tropical cyclone data.

3. The method according to claim 2, characterized in that The step of obtaining candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data sets includes: Obtaining a first historical time period of each historical tropical cyclone data in the historical tropical cyclone data set; Acquire a second historical time period of each historical sunspot data in the historical sunspot data set; According to each first historical time period and each second historical time period, a candidate associated sunspot data set of each historical tropical cyclone data is obtained from the historical sunspot data set.

4. The method according to claim 3, characterized in that The step of obtaining, according to each first historical time period and each second historical time period, candidate associated sunspot data sets of each historical tropical cyclone data from the historical sunspot data set comprises: For any historical tropical cyclone data, obtain a first historical year to which a first historical time period of the historical tropical cyclone data belongs, and a second historical year that is the previous year of the first historical year; According to each second historical time period, obtaining a first-year sunspot data set of the first historical year from the historical sunspot data set to obtain a first monthly average sunspot number in the first historical year; According to each second historical time period, obtaining a second year sunspot data set for the second historical year from the historical sunspot data set to obtain an annual average sunspot number and a second monthly average sunspot number for the second historical year; The candidate associated sunspot data set of the historical tropical cyclone data is obtained according to the first monthly average sunspot number, the second monthly average sunspot number and the annual average sunspot number.

5. The method according to claim 2, characterized in that: The step of obtaining, for any historical tropical cyclone data, candidate correlation parameters between the historical tropical cyclone data and each candidate correlated sunspot data comprises: For any candidate associated sunspot data, the Pearson correlation coefficient between the historical tropical cyclone data and the candidate associated sunspot data is obtained to obtain the candidate correlation degree parameter between the historical tropical cyclone data and the candidate associated sunspot data.

6. The method according to claim 2, characterized in that The step of obtaining a sample tropical cyclone data set from the historical tropical cyclone data set according to each candidate correlation parameter, and obtaining sample sunspot data of each sample tropical cyclone data from the sample tropical cyclone data set, comprises: For any candidate correlation parameter, in response to the candidate correlation parameter being greater than or equal to a preset correlation parameter threshold, the historical tropical cyclone data corresponding to the candidate correlation parameter is determined to be the sample tropical cyclone data, and the candidate correlated sunspot data corresponding to the candidate correlation parameter is determined to be the sample sunspot data corresponding to the sample tropical cyclone data.

7. The method according to claim 1, characterized in that The step of obtaining a candidate cyclone prediction model to be trained, and performing model training on the candidate cyclone prediction model through the set of sample association data until the training is completed to obtain a trained target cyclone prediction model includes: For any sample-associated data pair, obtaining a training sample according to the sample sunspot data and the sample tropical cyclone data in the sample-associated data pair; Training the candidate cyclone prediction model according to the training samples to obtain a training loss of the candidate cyclone prediction model; The parameters of the candidate cyclone prediction model are adjusted according to the training loss, and the next training sample is returned to continue training the candidate cyclone prediction model after the parameter adjustment until the training is completed, so as to obtain the trained target cyclone prediction model.

8. The method according to claim 1, characterized in that The step of acquiring the target sunspot data to be identified and obtaining the target predicted cyclone data within the prediction time range based on the target cyclone prediction model and the target sunspot data includes: Inputting the target sunspot data into the target cyclone prediction model; The data features of the target sunspot data are extracted according to the target cyclone prediction model, so as to output the target predicted cyclone data within the prediction time range according to the data features.

9. A tropical cyclone prediction device, characterized in that: The device comprises: An acquisition module, used to acquire a historical sunspot data set and a historical tropical cyclone data set to obtain a set of sample associated data pairs, wherein for any sample associated data pair, the sample associated data pair includes sample sunspot data and sample tropical cyclone data; A training module, used for obtaining a candidate cyclone prediction model to be trained, and performing model training on the candidate cyclone prediction model through the set of sample association data until the training is completed, thereby obtaining a trained target cyclone prediction model; The prediction module is used to obtain the target sunspot data to be identified, and based on the target cyclone prediction model and the target sunspot data, obtain the target predicted cyclone data within the prediction time range.

10. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 8.

11. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1 to 8.

Images