Crop yield meteorological disaster loss risk prediction method

Abstract

The invention provides a crop yield meteorological disaster loss risk prediction method. The method comprises the following steps: acquiring original crop yield meteorological disaster loss sequence data; adding the obtained original crop yield meteorological disaster loss sequence data by adopting an interpolation method; performing transformation processing on the interpolated crop yield meteorological disaster loss sequence data by adopting inverse Logistic to obtain transformed crop yield meteorological disaster loss sequence data; decomposing the interpolated and transformed crop yield meteorological disaster loss sequence data into a series of intrinsic mode function components and residual trend terms by adopting an empirical mode decomposition algorithm; predicting each intrinsic mode function component by using a support vector machine model, and fitting residual trend terms by using a polynomial fitting method; and adding the prediction result of each intrinsic mode function component and the prediction result of each trend component, and then carrying out Logistic transformation to obtain a final crop yield meteorological disaster loss prediction result.

Description

Technical Field

[0001] The present invention relates to the technical field of disaster risk prediction, and particularly relates to a method for predicting the risk of meteorological disaster losses in crop yields. Background Art

[0002] The prediction of the risk of meteorological disaster losses in crop yields is a crucial link in agricultural production, which helps farmers and agricultural management departments take measures in advance to reduce the impact of meteorological disasters on crops.

[0003] According to the prediction results of the risk of meteorological disaster losses in crop yields, the planting structure and layout of crops can be reasonably planned to reduce the impact of disasters on agricultural production; disaster prevention and mitigation measures such as irrigation, drainage, heat preservation, and sunshading can also be formulated in advance to reduce the damage degree of meteorological disasters on crop yields.

[0004] Weather methods, climatological methods, phenological methods, numerical simulation methods, satellite remote sensing methods, and statistical methods are often used for the prediction of the risk of meteorological disaster losses in crop yields. There are various methods for predicting the risk of meteorological disaster losses in crop yields, and each method has its scope of application and limitations, and the accuracy of its prediction needs to be improved. Summary of the Invention

[0005] The present invention provides a method for predicting the risk of meteorological disaster losses in crop yields to solve the above technical problems.

[0006] The technical solution adopted by the present invention is as follows:

[0007] A method for predicting the risk of meteorological disaster losses in crop yields includes the following steps: obtaining the original sequence data of meteorological disaster losses in crop yields; increasing the obtained original sequence data of meteorological disaster losses in crop yields by using an interpolation method to obtain the interpolated sequence data of meteorological disaster losses in crop yields; performing an inverse Logistic transformation on the obtained interpolated sequence data of meteorological disaster losses in crop yields to obtain the transformed sequence data of meteorological disaster losses in crop yields; decomposing the interpolated and transformed sequence data of meteorological disaster losses in crop yields into a series of intrinsic mode function components and a residual trend term by using an empirical mode decomposition algorithm; predicting each intrinsic mode function component by using a support vector machine model, and fitting the residual trend term by using a polynomial fitting method; adding the prediction results of each intrinsic mode function component and the prediction result of the trend component, and then performing a Logistic transformation to obtain the final prediction result of the meteorological disaster losses in crop yields.

[0008] Performing an inverse Logistic transformation on the obtained interpolated sequence data of meteorological disaster losses in crop yields, and its transformation formula is:

[0009]

[0010] In the formula, y is the interpolated sequence data of the meteorological disaster losses of crop yields, and y * is the sequence data of the meteorological disaster losses of crop yields after interpolation and transformation; L is the upper bound of the original sequence data of the meteorological disaster losses of crop yields obtained, which is determined according to the maximum value of the disaster loss index; k is the steepness of the disaster loss curve, and k takes 0.5.

[0011] The mean absolute error MAE and the root mean square error RMSE are selected as the evaluation indexes for the accuracy of the prediction results, and their expressions are as follows:

[0012] Mean absolute error:

[0013] Root mean square error:

[0014] In the formula, y i represents the measured value of the disaster loss, represents the predicted value of the disaster loss, and n represents the length of the predicted value.

[0015] Advantages of the present invention:

[0016] For the risk prediction method of the meteorological disaster losses of crop yields of the present invention, the SVM + interpolation + EMD method can be preferably used for predicting the risk of meteorological disaster losses. First, the interpolation method is used to increase the original sequence data of the meteorological disaster losses of crop yields and perform inverse Logistic transformation processing. Then, the empirical mode decomposition EMD method is used to decompose the interpolated and transformed original meteorological disaster loss data sequence into a series of intrinsic mode function components and residual trend terms. Next, the support vector machine SVM is used to predict each intrinsic mode function component, and the polynomial fitting method is used to fit the residual trend terms. Finally, the prediction results are summed and Logistic transformed to obtain the final prediction result of the meteorological disaster losses of crop yields, realizing the accurate prediction of the risk of meteorological disaster losses of crop yields. Description of the drawings

[0017] Figure 1 is the flow chart of the risk prediction method of the meteorological disaster losses of crop yields in the embodiment of the present invention;

[0018] Figure 2 is the spatial distribution map of the risk of meteorological disaster losses of potatoes in Gansu Province predicted based on different prediction schemes. Detailed implementation manners

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0020] As Figure 1 shown, a method for predicting the risk of meteorological disaster losses in crop yields according to an embodiment of the present invention includes the following steps:

[0021] S1: Obtain the original sequence data of meteorological disaster losses in crop yields.

[0022] S2: Use an interpolation method to increase the obtained original sequence data of meteorological disaster losses in crop yields to obtain the interpolated sequence data of meteorological disaster losses in crop yields.

[0023] The interpolation method is a technique for estimating new data points within a given range using existing discrete data points to form a long sequence of data. In the assessment of agricultural meteorological disaster losses, the original data may not be dense enough due to various reasons (such as limitations of observation equipment, insufficient observation frequency, etc.). At this time, the interpolation method can be used to increase the data points, thereby obtaining a smoother and more detailed data sequence. Common interpolation methods include nearest-neighbor interpolation, linear interpolation, cubic spline function interpolation, and piecewise cubic Hermite interpolation, etc. For the meteorological disaster loss data of crop yields in the present invention, the piecewise cubic Hermite interpolation method is used.

[0024] Exemplarily, for the 38 samples (original sequence data of meteorological disaster losses in crop yields) obtained, by inserting 4 values between two sample points, the interpolated original meteorological disaster loss sequence data (186 samples) is obtained.

[0025] S3: Perform an inverse Logistic transformation on the obtained interpolated sequence data of meteorological disaster losses in crop yields to obtain the transformed sequence data of meteorological disaster losses in crop yields.

[0026] Among them, the inverse Logistic transformation is the inverse process of the Logistic transformation, mainly used to restore the data after the Logistic transformation back to the original scale. When the interpolated original meteorological disaster loss sequence data is decomposed, the value range of the subsequence may change, which will lead to negative values in the subsequence prediction. Therefore, the inverse Logistic transformation is first performed on the interpolated original meteorological disaster loss sequence data.

[0027] Among them, the obtained interpolated original disaster loss sequence data is processed by inverse Logistic transformation, and its transformation formula is:

[0028]

[0029] In the formula, y is the interpolated original meteorological disaster loss sequence data, and y * is the meteorological disaster loss sequence after interpolation and transformation; L is the upper bound of the interpolated original disaster loss sequence, which is determined according to the maximum value of the disaster loss index; k is the steepness of the disaster loss curve, and k takes 0.5.

[0030] S4: Use the empirical mode decomposition algorithm to decompose the interpolated and transformed meteorological disaster loss sequence data of crop yields into a series of intrinsic mode function components and residual trend terms.

[0031] Empirical mode decomposition (EMD) is a powerful tool for analyzing non-linear and non-stationary signals. The empirical mode decomposition algorithm decomposes a complex, non-stationary, non-linear time series into the sum of several intrinsic mode function (IMF) components with obvious periodicity and the remaining trend term (RES). Each component is independent of each other, there is no loss during the decomposition process, and the sum is equal to the original time series. Here, according to the EMD principle, a complex, non-stationary, non-linear disaster loss time series can be decomposed into several components with frequency characteristics (IMF components) and a trend term (RES). Then, predictions are made for each IMF component and trend term.

[0032] S5: Use the support vector machine model to predict each intrinsic mode function component, and use the polynomial fitting method to fit the residual trend term.

[0033] Among them, the support vector machine (SVM) is based on statistical theory and is a small-sample machine learning method. It adopts the principle of minimizing structural risk, maps data in a low-dimensional space and linearly inseparable data to a high-dimensional space to make it linearly separable, and then makes predictions on the data in the high-dimensional space. This method can effectively avoid the local extreme value problem, prevent data overfitting, and significantly improve the prediction accuracy.

[0034] Specifically, the time series data of each IMF component and trend component is divided into three parts: the first 60% of the data is used as the training sample to construct the prediction model, the remaining 35% of the data is used as the validation sample to verify the model accuracy and correct its hyperparameters, and the remaining 5% of the data is used as the test data to verify the model performance. The IMF components are predicted by SVM, and the trend components are modeled and predicted by a first-degree polynomial.

[0035] S6: Sum up the prediction results of each intrinsic mode function component and the prediction result of the trend component, and then perform a Logistic transformation to obtain the final prediction result of the meteorological disaster loss of crop yields.

[0036] In S6, the mean absolute error (MAE) and the root mean square error (RMSE) can be selected as the evaluation indicators for the accuracy of the prediction results. Each error indicator can be used to measure the deviation degree between the predicted value and the actual value. The smaller the error value, the higher the prediction accuracy. The expressions are as follows:

[0037] Mean absolute error:

[0038] Root mean square error:

[0039] In the formula, y i represents the measured value of the disaster loss, represents the predicted value of the disaster loss, and n represents the length of the predicted value.

[0040] The following takes the annual prediction of the meteorological disaster loss risk of potato yields in Gansu Province as an example for illustration:

[0041] Obtain the sequence data of the meteorological disaster loss of potato yields in each county of Gansu Province. After interpolation, the number of disaster loss samples increases from 38 to 186. The interpolated disaster loss curve is consistent with the original disaster loss curve. In addition, among different years, the average fluctuation coefficient representing the disaster loss risk varies greatly, with great uncertainty.

[0042] Adopt the empirical mode decomposition (EMD) method to decompose the original sequence of the average fluctuation coefficient into multiple subsequences (IMF components and trend terms). The IMF components have obvious periodic and frequency characteristics, while the trend term shows a significant downward trend, indicating that the disaster loss fluctuation gradually decreases.

[0043] Then divide each IMF component data sequence into a training set (60% of the samples), a validation set (35% of the samples), and a test set (5% of the samples). First, use the support vector machine (SVM) algorithm to construct a model based on the training samples, and then use the validation set to calibrate and evaluate the model. Secondly, use a linear regression equation to simulate the trend term. Finally, after transformation, the prediction results of the training set and the validation set are obtained. Finally, this step is implemented on the potato yield disaster loss data of each county in Gansu Province to obtain the prediction results of the meteorological disaster loss risk of potato yields in each county of Gansu Province.

[0044] Table 1 shows the fitting accuracy of the training set and the validation set for predicting the meteorological disaster loss risk of potato yields in each county of Gansu Province using different prediction schemes.

[0045] Table 1: Prediction accuracy statistics (expressed as mean error ± standard deviation (SD)) of the disaster loss risk (average fluctuation coefficient) in Gansu Province based on different prediction schemes

[0046]

[0047] Obviously, as can be seen from Table 1, both the interpolation and empirical mode decomposition (EMD) algorithms can improve the prediction accuracy of the support vector machine (SVM), and interpolation has a stronger enhancement effect.

[0048] Specifically, for the potato disaster loss data in each county of Gansu Province, the average MAE and SD of the training set using SVM are 6.8 and 4.21 respectively. However, after interpolation, the average MAE and SD are 1.96 and 1.95 respectively, and after EMD decomposition, the average MAE and SD are 2.69 and 2.03 respectively. If these two methods are used in combination, the average MAE and SD are 1.37 and 0.64 respectively. From the statistical results, the average MAE of the training set is reduced by 71.2% (interpolation), 60.4% (EMD), and 79.9% (SVM + interpolation + EMD); in addition, the average RMSE of the SVM, SVM + interpolation, SVM + EMD, and SVM + interpolation + EMD schemes are 8.35, 2.75, 3.63, and 2.48 respectively, that is, the average RMSE of the training set is reduced by 67.1% (SVM + interpolation), 56.5% (SVM + EMD), and 70.3% (SVM + interpolation + EMD).

[0049] For the validation set, the average MAE of the SVM + interpolation, SVM + EMD, and SVM + interpolation + EMD schemes is reduced by 55.4%, 39.9%, and 50.3% respectively, while the average RMSE is reduced by 69.0%, 42.5%, and 61.0% respectively.

[0050] In summary, the SVM + interpolation + EMD method can be preferably used to predict the disaster loss risk.

[0051] Use the SVM + interpolation + EMD scheme to conduct annual prediction of the potato disaster loss risk:

[0052] Use the prediction models constructed by each disaster loss risk prediction scheme to predict the average fluctuation coefficient of potato yield per unit area in each county of Gansu Province in 2021 and 2022. The spatial distribution of the potato disaster loss risk predicted based on different prediction schemes is as Figure 2 shown.

[0053] From Figure 2It can be seen that whether using interpolation, the separate EMD method, or the SVM + interpolation + EMD combination, their spatial distributions are clearly more in line with the spatial distribution of the original fluctuation coefficients than the separate SVM model. Especially after adopting the SVM + interpolation + EMD technique, the high-value and low-value regions can be more accurately reflected. When using SVM alone, the predicted average fluctuation coefficient is significantly greater than the actual value. In terms of prediction accuracy, the overall performance of the SVM + interpolation + EMD scheme is better than that of using a single SVM model and can produce more accurate prediction results at the regional scale.

[0054] Based on different schemes, the correlations between the true values and predicted values of the average fluctuation coefficients of potato yields in Gansu Province are shown in Table 2. After adopting various schemes, the correlation between the true values and predicted values has increased significantly. After combining SVM with interpolation and EMD techniques, the correlation coefficient in 2021 is 0.569, reaching a highly significant level.

[0055] Table 2 Correlations between the true values and predicted values of the average fluctuation coefficients in 2021 and 2022 predicted based on different prediction schemes

[0056]

[0057] As can be seen from Table 2, when adopting the scheme of adding only one technique, the correlation did not improve significantly in 2022. However, when combining these two methods, the correlation increased significantly, and the correlation coefficient reached 0.515, reaching a highly significant level.

[0058] In summary, according to the method for predicting the risk of meteorological disaster losses in crop yields of the embodiments of the present invention, the SVM + interpolation + EMD method can be preferably used to predict the risk of meteorological disaster losses. First, the interpolation method is used to increase the original meteorological disaster loss sequence data of crop yields and perform inverse Logistic transformation processing. Then, the empirical mode decomposition (EMD) method is used to decompose the interpolated and transformed original meteorological disaster loss data sequence into a series of intrinsic mode function components and a residual trend term. Next, the support vector machine (SVM) is used to predict each intrinsic mode function component, and the polynomial fitting method is used to fit the residual trend term. Finally, the prediction results are summed and Logistic-transformed to obtain the final prediction result of the risk of meteorological disaster losses in crop yields, realizing the accurate prediction of the risk of meteorological disaster losses in crop yields.

[0059] In the description of the present invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include one or more of such features. The meaning of "a plurality" is two or more, unless specifically defined otherwise.

[0060] In the present invention, unless otherwise clearly defined and limited, terms such as "mounted", "connected", "coupled", "fixed", etc. shall be construed in a broad sense. For example, it may be a fixed connection, a detachable connection, or integrated; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium, and it may be the internal communication of two components or the interaction relationship between two components. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0061] In the present invention, unless otherwise clearly defined and limited, the first feature being "on" or "under" the second feature may be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on top of" the second feature may be that the first feature is directly above or obliquely above the second feature, or merely indicates that the first feature has a higher horizontal height than the second feature. The first feature being "under", "beneath" and "underneath" the second feature may be that the first feature is directly below or obliquely below the second feature, or merely indicates that the first feature has a lower horizontal height than the second feature.

[0062] In the description of this specification, the description with reference to terms such as "an embodiment", "some embodiments", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not have to be directed to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0063] Any process or method description represented in a flowchart or otherwise described herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where functions may be executed in a manner that is not shown or discussed, including in a substantially simultaneous manner according to the functions involved or in a reverse order, which should be understood by those skilled in the art to which the embodiments of the present invention pertain.

[0064] The logic and / or steps represented in a flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing a logical function, and can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpretation, or other appropriate processing as necessary, and then stored in a computer memory.

[0065] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above 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 in hardware, as in another embodiment, any one of the following techniques known in the art or a combination thereof can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), and the like.

[0066] Those of ordinary skill in the art can understand that all or part of the steps carried out in implementing the above-described embodiment methods can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0067] In addition, in each of the embodiments of the present invention, the functional units can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above-mentioned integrated module can be implemented in the form of hardware or in the form of a software functional module. When the above-mentioned integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0068] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting the risk of crop yield loss due to meteorological disasters, characterized in that: The following steps are involved: Obtain the original crop yield meteorological disaster loss series data; The interpolation method is used to increase the original crop yield meteorological disaster loss series data to obtain the interpolated crop yield meteorological disaster loss series data; The interpolated crop yield meteorological disaster loss series data are transformed by inverse Logistic to obtain the transformed crop yield meteorological disaster loss series data; The empirical mode decomposition algorithm is used to decompose the interpolated and transformed crop yield meteorological disaster loss series data into a series of intrinsic mode function components and residual trend terms; The support vector machine model is used to predict each intrinsic mode function component, and the polynomial fitting method is used to fit the residual trend term; The prediction results of each intrinsic mode function component and the prediction results of the trend component are added together, and then a Logistic transformation is performed to obtain the final prediction results of crop yield meteorological disaster losses.

2. The method for predicting crop yield meteorological disaster loss risk according to claim 1, characterized in that: The interpolated crop yield meteorological disaster loss series data is transformed using inverse Logistic transformation, and the transformation formula is: In the formula, y ， is the interpolated crop yield meteorological disaster loss series data, y * is the interpolated and transformed crop yield meteorological disaster loss series data; L is the upper bound of the original crop yield meteorological disaster loss series data, which is determined according to the maximum value of the disaster loss index; k is the steepness of the disaster loss curve, and k is 0.

5.

3. The method for predicting the risk of crop yield meteorological disaster losses according to claim 2, characterized in that: The mean absolute error MAE and root mean square error RMSE are selected as the prediction result accuracy evaluation indicators, and their expressions are as follows: Mean absolute error: Root Mean Square Error: In the formula, y i represents the measured value of disaster losses, represents the predicted value of disaster loss, and n represents the length of the predicted value.

Images