Regional precipitation prediction method and system based on big data

Abstract

The invention provides a regional precipitation prediction method based on big data, and belongs to the data processing technology, and the method comprises the steps: obtaining a plurality of first historical average precipitation indexes of a plurality of target precipitation sensors over the years based on the big data, and then building a neural network model with the correlation between the first historical average precipitation indexes and second historical average precipitation indexes, according to the method, a neural network model is constructed, data obtained during monitoring are imported into the model, and according to the difference condition between the monitored data and historical average data, the weight when the neural network model is input is selectively increased or reduced, so that a final prediction result is obtained. According to the regional precipitation prediction method based on the big data, the regional precipitation is predicted through the neural network model, when the data are input into the neural network model, the credibility of the data is judged firstly, if the difference between the data and historical average data is large, the credibility is low, and if the difference between the data and the historical average data is large, the credibility is low. And setting corresponding high and low weights according to the credibility of the data and inputting the high and low weights into the neural network model, so that the final result is more accurate.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a regional precipitation prediction method based on big data. Background Art

[0002] Regional precipitation prediction is an important research direction in meteorology, hydrology, agriculture and other fields. Accurate precipitation prediction is of great significance for water resources management, disaster warning, agricultural production and urban planning. With the development of big data technology, traditional precipitation prediction methods are gradually replaced by methods based on big data and artificial intelligence.

[0003] Statistical analysis based on historical precipitation data, such as time series analysis (ARIMA) and regression analysis, can predict the precipitation at a single location, but cannot predict the precipitation in surrounding locations. In actual precipitation, the precipitation situation is often correlated with the surrounding precipitation situation. Therefore, the prediction effect based only on statistical analysis of historical precipitation data is not good. Summary of the invention

[0004] The embodiments of the present application provide a regional precipitation prediction method and system based on big data to improve the above-mentioned problems.

[0005] In order to achieve the above objectives, this application adopts the following technical solutions: In a first aspect, an embodiment of the present application proposes a regional precipitation prediction method based on big data, the method comprising: Determine a target area, the target area is an area where precipitation prediction is required, and determine a source data area based on the target area, the source data area is a range area surrounding the target area, wherein a plurality of precipitation sensors distributed in the target area are defined as target precipitation sensors, and a plurality of precipitation sensors distributed in the source data area are defined as source precipitation sensors; Based on the big data, a plurality of first historical average precipitation indexes of a plurality of target precipitation sensors in previous years and a plurality of second historical average precipitation indexes of a plurality of source precipitation sensors in previous years are obtained; Inputting a plurality of first historical average precipitation indices and a plurality of second historical average precipitation indices as training sets into a target neural network model and performing iterative calculations, and obtaining a trained target neural network model, the target neural network model is used to output a plurality of predicted target precipitation indices according to the input plurality of source precipitation indices, wherein the plurality of first historical average precipitation indices have the same weight when input into the target neural network model; Acquire multiple source precipitation indices acquired by multiple source precipitation sensors in a first monitoring period; Comparing the plurality of source precipitation indices with the plurality of corresponding first historical average precipitation indices, and dividing the plurality of source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein the average difference between the plurality of source precipitation indices belonging to the high difference group and the plurality of corresponding first historical average precipitation indices is greater than the average difference between the plurality of source precipitation indices belonging to the low difference group and the plurality of corresponding first historical average precipitation indices; Inputting a plurality of source precipitation indices into the target neural network model, wherein the weights corresponding to the input of the plurality of source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of the plurality of source precipitation indices belonging to the low difference group; Based on the output results of the target neural network model, the predicted target precipitation index is obtained, and the predicted precipitation is obtained based on the target precipitation index.

[0006] In combination with the first aspect, optionally, multiple first historical average precipitation indices and multiple second historical average precipitation indices are input into a target neural network model as training sets and iteratively calculated, and a trained target neural network model is obtained, and the target neural network model is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Inputting a plurality of first historical average precipitation indexes as a plurality of dimensions into a target neural network model, wherein the plurality of dimensions constitute a feature vector of a second historical average precipitation index; Importing multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and performing iterative calculations, and comparing the result of each calculation with the multiple second historical average precipitation indices during the calculation process, and obtaining an error value; When the error value is less than the preset threshold, the iterative calculation is stopped and the trained target neural network model is output.

[0007] In combination with the first aspect, optionally, multiple first historical average precipitation indices and multiple second historical average precipitation indices are imported into the target neural network model and iteratively calculated, and during the calculation process, the result of each calculation is compared with the multiple second historical average precipitation indices, and an error value is obtained, including: An initializer is determined during the process of building a target neural network, and a plurality of identical weights are determined based on the initializer; The first historical average precipitation index is input into the input layer of the target neural network based on the initializer.

[0008] In combination with the first aspect, optionally, multiple first historical average precipitation indices and multiple second historical average precipitation indices are imported into the target neural network model and iteratively calculated, and during the calculation process, the result of each calculation is compared with the multiple second historical average precipitation indices, and an error value is obtained, including: If each result of each calculation during the calculation process is greater than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0009] In combination with the first aspect, optionally, multiple first historical average precipitation indices and multiple second historical average precipitation indices are imported into the target neural network model and iteratively calculated, and during the calculation process, the result of each calculation is compared with the multiple second historical average precipitation indices, and an error value is obtained, including: If each result of each calculation during the calculation process is less than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0010] In combination with the first aspect, optionally, a plurality of source precipitation indices are input into the target neural network model, wherein the weights corresponding to the input of the plurality of source precipitation indices belonging to the high difference group are less than the weights corresponding to the input of the plurality of source precipitation indices belonging to the low difference group, including: Get the initializer and modify it based on different weights; Multiple source precipitation indices are input into the input layer of the target neural network based on the modified initializer.

[0011] In combination with the first aspect, optionally, multiple first historical average precipitation indices and multiple second historical average precipitation indices are input into a target neural network model as training sets and iteratively calculated, and a trained target neural network model is obtained, and the target neural network model is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Total The first historical average precipitation index, the weight corresponding to each first historical average precipitation index is , the value input to the target neural network model each time satisfy:

[0012] in, It is the first historical average precipitation index; Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, including: When the source precipitation index belongs to the high difference group, the value input to the target neural network model each time satisfy: ; When the source precipitation index belongs to the low difference group, the value input to the target neural network model each time satisfy:

[0013] in, is a source precipitation index belonging to the high difference group, is a source precipitation index belonging to the low-difference group, is the corresponding weight of the source precipitation index belonging to the high difference group, is the corresponding weight of the source precipitation index belonging to the low-discrepancy group, and > .

[0014] In combination with the first aspect, optionally, multiple source precipitation indices are input into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are less than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, satisfying:

[0015] in, is the number of source precipitation indices belonging to the high-discrepancy group, is the number of source precipitation indices belonging to the high discrepancy group.

[0016] In the second aspect, the present application proposes a regional precipitation prediction system based on big data, the system comprising: The first submodule is used to determine a target area, which is an area where precipitation prediction is required, and determine a source data area based on the target area, which is a range area surrounding the target area, wherein a plurality of precipitation sensors distributed in the target area are defined as target precipitation sensors, and a plurality of precipitation sensors distributed in the source data area are defined as source precipitation sensors; The second submodule is used to obtain a plurality of first historical average precipitation indexes of a plurality of target precipitation sensors in previous years and a plurality of second historical average precipitation indexes of a plurality of source precipitation sensors in previous years based on big data; The third submodule is used to input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain the trained target neural network model, the target neural network model is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model; A fourth submodule, the fourth submodule is used to obtain multiple source precipitation indices obtained by multiple source precipitation sensors in the first monitoring period; The fifth submodule is used to compare the multiple source precipitation indices with the multiple corresponding first historical average precipitation indices, and divide the multiple source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein the average difference between the multiple source precipitation indices belonging to the high difference group and the multiple corresponding first historical average precipitation indices is greater than the average difference between the multiple source precipitation indices belonging to the low difference group and the multiple corresponding first historical average precipitation indices; The sixth submodule inputs a plurality of source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of source precipitation indices belonging to the low difference group; The seventh submodule obtains the predicted target precipitation index based on the output result of the target neural network model, and obtains the predicted precipitation based on the target precipitation index.

[0017] In conjunction with the second aspect, optionally, the system may also be configured as: Input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain a trained target neural network model, which is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Inputting a plurality of first historical average precipitation indexes as a plurality of dimensions into a target neural network model, wherein the plurality of dimensions constitute a feature vector of a second historical average precipitation index; Importing multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and performing iterative calculations, and during the calculation process, comparing the result of each calculation with the multiple second historical average precipitation indices, and obtaining an error value; When the error value is less than the preset threshold, the iterative calculation is stopped and the trained target neural network model is output.

[0018] In conjunction with the second aspect, optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: An initializer is determined during the process of building a target neural network, and a plurality of identical weights are determined based on the initializer; The first historical average precipitation index is input into the input layer of the target neural network based on the initializer.

[0019] In conjunction with the second aspect, optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: If each result of each calculation during the calculation process is greater than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0020] In conjunction with the second aspect, optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: If each result of each calculation during the calculation process is less than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0021] In conjunction with the second aspect, optionally, the system may also be configured as: Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, including: Get the initializer and modify it based on different weights; Multiple source precipitation indices are input into the input layer of the target neural network based on the modified initializer.

[0022] In conjunction with the second aspect, optionally, the system may also be configured as: Input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain a trained target neural network model, which is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Total The first historical average precipitation index, the weight corresponding to each first historical average precipitation index is , the value input to the target neural network model each time satisfy:

[0023] in, It is the first historical average precipitation index; Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, including: When the source precipitation index belongs to the high difference group, the value input to the target neural network model each time satisfy: ; When the source precipitation index belongs to the low difference group, the value input to the target neural network model each time satisfy:

[0024] in, is a source precipitation index belonging to the high difference group, is a source precipitation index belonging to the low-difference group, is the corresponding weight of the source precipitation index belonging to the high difference group, is the corresponding weight of the source precipitation index belonging to the low-discrepancy group, and > .

[0025] In conjunction with the second aspect, optionally, the system may also be configured as: Multiple source precipitation indices are input into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are less than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, satisfying:

[0026] in, is the number of source precipitation indices belonging to the high-discrepancy group, is the number of source precipitation indices belonging to the high discrepancy group.

[0027] A third aspect of an embodiment of the present invention provides an electronic device, the 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 method proposed in the first aspect of the embodiment of the present invention.

[0028] A fourth aspect of an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method provided in the first aspect of the embodiment of the present invention.

[0029] In summary, the above method and system have the following technical effects: The present application proposes a regional precipitation prediction method based on big data, which obtains multiple first historical average precipitation indexes of multiple target precipitation sensors in previous years and multiple second historical average precipitation indexes of multiple source precipitation sensors in previous years based on big data, and then establishes a neural network model with the correlation between the first historical average precipitation index and the second historical average precipitation index, and imports the data obtained during monitoring into the model. During the import process, according to the difference between the monitoring data and the historical average data, the weight when inputting the neural network model is selectively increased or decreased to obtain the final prediction result. The present application proposes a regional precipitation prediction method based on big data, which predicts regional precipitation through a neural network model. When inputting the data into the neural network model, the credibility of the data is first judged. If the difference with the historical average data is large, the credibility is low, and vice versa. According to the credibility of the data, the corresponding high and low weights are set and input into the neural network model, so that the final result is more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A schematic flow chart of a regional precipitation prediction method based on big data proposed in an embodiment of the present application. DETAILED DESCRIPTION

[0031] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0032] See also Figure 1 The present application embodiment proposes a regional precipitation prediction method based on big data, the method comprising steps S101-S107: S101: Determine a target area, which is an area where precipitation prediction is required, and determine a source data area based on the target area, which is a range area surrounding the target area, wherein several precipitation sensors distributed in the target area are defined as target precipitation sensors, and several precipitation sensors distributed in the source data area are defined as source precipitation sensors.

[0033] As can be understood, the target area refers to a specific geographical area where precipitation forecasting is planned. Once the target area is determined, the next step is to define a source data area based on this target area. The source data area is a larger area around the target area, which includes the target area and may also include some additional surrounding areas. In the target area, several precipitation sensors can be set up, which are specifically used to monitor the precipitation in the target area, so they are called target precipitation sensors. At the same time, in the source data area, we also set up several precipitation sensors, which are responsible for monitoring the precipitation in the source data area, so they are called source precipitation sensors. In this way, precipitation data in the target area and its surrounding areas can be collected to provide necessary data support for subsequent precipitation forecasts. At the same time, by combining the surrounding precipitation to predict the precipitation in the target area, the surrounding data is combined, which is more accurate than when only historical data is used to predict.

[0034] S102: Acquire multiple first historical average precipitation indices of multiple target precipitation sensors in previous years and multiple second historical average precipitation indices of multiple source precipitation sensors in previous years based on big data.

[0035] Specifically, historical precipitation data can be obtained from the target precipitation sensor and the source precipitation sensor. For each sensor, the average precipitation index is calculated annually. The average precipitation index of the target precipitation sensor is the first historical average precipitation index. The average precipitation index of the source precipitation sensor is the second historical average precipitation index. It can be understood that statistical analysis based on historical data can be used to build a prediction model to obtain more accurate data.

[0036] S103: Input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain the trained target neural network model, which is used to output multiple predicted target precipitation indices based on the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model.

[0037] It can be understood that multiple first historical average precipitation indices, i.e., historical data of target precipitation sensors, and multiple second historical average precipitation indices, i.e., historical data of source precipitation sensors, are integrated into one data set, while ensuring that the time range of the data is consistent and aligned in time.

[0038] Of course, in some embodiments, the data set can also be divided into a training set and a test set, for example, 80% for training and 20% for testing. The training set is used to train the model, and the test set is used to evaluate the model performance. The specific training set setting is not limited in this application.

[0039] After that, select a suitable neural network structure, such as a fully connected neural network (FCN), a recurrent neural network (RNN), or a long short-term memory network (LSTM). The number of nodes in the input layer should be consistent with the number of input features. For example, if there are N source precipitation indices and M target precipitation indices, the number of input layer nodes is N+M. Of course, the number of hidden layers and the number of nodes can also be set according to requirements. You can start with a simple structure and adjust it step by step. The number of nodes in the output layer should be consistent with the number of target precipitation indices. That is, if there are M precipitation sensors distributed in the target area, if you need to predict M target precipitation indices, the number of nodes in the output layer is M.

[0040] It should be noted that in this process, since it is in the construction stage of the neural network, no actual prediction function will be performed. Therefore, in the input layer, it is necessary to ensure that multiple first historical average precipitation indices have the same weight.

[0041] Specifically, as an implementation method, multiple first historical average precipitation indices can be input into the target neural network model as multiple dimensions, and the multiple dimensions constitute the characteristic vector of the second historical average precipitation index. Then, the multiple first historical average precipitation indices and the multiple second historical average precipitation indices are imported into the target neural network model and iteratively calculated. During the calculation process, the result of each calculation is compared with the multiple second historical average precipitation indices, and the error value is obtained.

[0042] Exemplarily, specifically, as an implementation method, let the feature vector be x=[ , , ],in , and Respectively represent the normalized first historical average precipitation indexes, and the neural network training model is assumed to be f( ),in Represents the model parameters so that the model output value f( ) can be close to the true value y of multiple second historical average precipitation indices. Then the training process of the neural network training model can be achieved by minimizing the loss function, that is: L( ) = Where N represents the number of training samples. and Represent the input features and output values ​​of the i-th sample respectively. This loss function can measure the error between the model prediction value and the true value. The model parameters can be updated through optimization algorithms such as back propagation algorithm and gradient descent. , thereby improving the accuracy of the model.

[0043] Exemplarily, in order to ensure the same weights, in this embodiment, an initializer is determined during the construction of the target neural network, and multiple identical weights are determined based on the initializer, and then the first historical average precipitation index is input into the input layer of the target neural network based on the initializer. Of course, in other embodiments, it can also be implemented by initializing the weight matrix, etc., which is not limited in this application.

[0044] When the error value is less than the preset threshold, the iterative calculation is stopped and the trained target neural network model is output.

[0045] It is understandable that a custom training loop is used to check the loss value after each iteration and decide whether to stop training. As for how to obtain the error value, in this embodiment, if each result of each calculation during the calculation process is greater than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0046] Similarly, if each result of each calculation during the calculation process is less than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0047] It is understandable that in this way, the error value can be ensured to be always greater than 0. It is understandable that when the error value is less than the preset value, the training process of the neural network training model is stopped. This is because when the error value is less than a certain threshold, the model has predicted the data accurately enough, and continuing to train the model will not be effective and may cause problems such as overfitting.

[0048] S104: Acquire multiple source precipitation indices acquired by multiple source precipitation sensors in the first monitoring cycle.

[0049] It can be understood that the trained neural network model has been obtained in the process of steps S101-S103. Therefore, in the actual precipitation prediction process, the multiple source precipitation indices obtained by the multiple source precipitation sensors can be used to predict the precipitation index of the target area. The prediction time can also be the first cycle, which is used to verify the value obtained by the target precipitation sensor. The prediction time can also be the second cycle, that is, the cycle after the first cycle. Compared with directly using the historical data of the target precipitation sensor for prediction, it can better combine the actual precipitation distribution.

[0050] S105: Compare multiple source precipitation indices with multiple corresponding first historical average precipitation indices, and divide the multiple source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein the average difference between the multiple source precipitation indices belonging to the high difference group and the multiple corresponding first historical average precipitation indices is greater than the average difference between the multiple source precipitation indices belonging to the low difference group and the multiple corresponding first historical average precipitation indices.

[0051] Specifically, ensure that the source precipitation index and the first historical average precipitation index are aligned in the time dimension, and compare the source precipitation index at each time point with the corresponding first historical average precipitation index. It can be understood that the larger the difference, the greater the difference between the precipitation at that point and the historical expectation, that is, the lower the credibility of the reference. Similarly, the smaller the difference, the higher the credibility of the reference.

[0052] Therefore, in this embodiment, the multiple source precipitation indices can be divided into a high difference group and a low difference group according to the comparison result. The number of data in the high difference group and the low difference group can be the same or different, which is not limited in this embodiment.

[0053] S106: Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group.

[0054] It can be understood that, since the average difference between the multiple source precipitation indices belonging to the high difference group and the multiple corresponding first historical average precipitation indices is large, the multiple source precipitation indices belonging to the high difference group have a low reference credibility, and therefore, when they are input into the target neural network model, a relatively low weight can be set. Conversely, since the average difference between the multiple source precipitation indices belonging to the low difference group and the multiple corresponding first historical average precipitation indices is small, the multiple source precipitation indices belonging to the high difference group have a high reference credibility, and therefore, when they are input into the target neural network model, a relatively high weight can be set.

[0055] Specifically, an initializer can be obtained, the initializer can be modified based on different weights, and then multiple source precipitation indices can be input into the input layer of the target neural network based on the modified initializer. Of course, different weights can also be set in other ways according to different actual models, which is not limited in this embodiment.

[0056] In order to ensure that the change in weights does not cause an overall increase or decrease in the actual results, it is necessary to ensure that the sum remains unchanged before and after the weight adjustment. For example, when constructing a neural network, there are The first historical average precipitation index, the weight corresponding to each first historical average precipitation index is , the value input to the target neural network model each time satisfy:

[0057] in, It is the first historical average precipitation index; When the source precipitation index belongs to the high difference group, the value input to the target neural network model each time satisfy: ; When the source precipitation index belongs to the low difference group, the value input to the target neural network model each time satisfy:

[0058] in, is a source precipitation index belonging to the high difference group, is a source precipitation index belonging to the low-difference group, is the corresponding weight of the source precipitation index belonging to the high difference group, is the corresponding weight of the source precipitation index belonging to the low-discrepancy group, and > .

[0059] S107: Based on the output result of the target neural network model, obtain the predicted target precipitation index, and obtain the predicted precipitation based on the target precipitation index.

[0060] It is understandable that after obtaining the predicted target precipitation index, the target precipitation index and the precipitation amount can be converted to each other, and how to convert is not limited in this embodiment. The predicted precipitation amount can be directly or indirectly obtained through the predicted target precipitation index.

[0061] The embodiment of the present application proposes a regional precipitation prediction method based on big data, which obtains multiple first historical average precipitation indexes of multiple target precipitation sensors in previous years and multiple second historical average precipitation indexes of multiple source precipitation sensors in previous years based on big data, and then establishes a neural network model with the correlation between the first historical average precipitation index and the second historical average precipitation index, and imports the data obtained during monitoring into the model. During the import process, according to the difference between the monitoring data and the historical average data, the weight when inputting the neural network model is selectively increased or decreased to obtain the final prediction result. The embodiment of the present application proposes a regional precipitation prediction method based on big data, which predicts regional precipitation through a neural network model. When inputting the data into the neural network model, the credibility of the data is first judged. If the difference with the historical average data is large, the credibility is low, and vice versa. According to the credibility of the data, the corresponding high and low weights are set and input into the neural network model, so that the final result is more accurate.

[0062] Based on the same inventive concept, the embodiment of the present application also proposes a regional precipitation prediction system based on big data, which includes: The first submodule is used to determine a target area, which is an area where precipitation prediction is required, and determine a source data area based on the target area, which is a range area surrounding the target area, wherein a plurality of precipitation sensors distributed in the target area are defined as target precipitation sensors, and a plurality of precipitation sensors distributed in the source data area are defined as source precipitation sensors; The second submodule is used to obtain a plurality of first historical average precipitation indexes of a plurality of target precipitation sensors in previous years and a plurality of second historical average precipitation indexes of a plurality of source precipitation sensors in previous years based on big data; The third submodule is used to input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain the trained target neural network model, the target neural network model is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model; A fourth submodule, the fourth submodule is used to obtain multiple source precipitation indices obtained by multiple source precipitation sensors in the first monitoring period; The fifth submodule is used to compare the multiple source precipitation indices with the multiple corresponding first historical average precipitation indices, and divide the multiple source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein the average difference between the multiple source precipitation indices belonging to the high difference group and the multiple corresponding first historical average precipitation indices is greater than the average difference between the multiple source precipitation indices belonging to the low difference group and the multiple corresponding first historical average precipitation indices; The sixth submodule inputs a plurality of source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of source precipitation indices belonging to the low difference group; The seventh submodule obtains the predicted target precipitation index based on the output result of the target neural network model, and obtains the predicted precipitation based on the target precipitation index.

[0063] Optionally, the system may also be configured as: Input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain a trained target neural network model, which is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Inputting a plurality of first historical average precipitation indexes as a plurality of dimensions into a target neural network model, wherein the plurality of dimensions constitute a feature vector of a second historical average precipitation index; Importing multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and performing iterative calculations, and comparing the result of each calculation with the multiple second historical average precipitation indices during the calculation process, and obtaining an error value; When the error value is less than the preset threshold, the iterative calculation is stopped and the trained target neural network model is output.

[0064] Optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: An initializer is determined during the process of building a target neural network, and a plurality of identical weights are determined based on the initializer; The first historical average precipitation index is input into the input layer of the target neural network based on the initializer.

[0065] Optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: If each result of each calculation during the calculation process is greater than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0066] Optionally, the system may also be configured as: Import multiple first historical average precipitation indices and multiple second historical average precipitation indices into the target neural network model and perform iterative calculations. During the calculation process, compare the result of each calculation with the multiple second historical average precipitation indices and obtain error values, including: If each result of each calculation during the calculation process is less than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each result of each calculation, is the corresponding historical average precipitation index for each second.

[0067] Optionally, the system may also be configured as: Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, including: Get the initializer and modify it based on different weights; Multiple source precipitation indices are input into the input layer of the target neural network based on the modified initializer.

[0068] Optionally, the system may also be configured as: Input multiple first historical average precipitation indices and multiple second historical average precipitation indices as training sets into the target neural network model and perform iterative calculations, and obtain a trained target neural network model, which is used to output multiple predicted target precipitation indices according to the input multiple source precipitation indices, wherein the multiple first historical average precipitation indices have the same weight when input into the target neural network model, including: Total The first historical average precipitation index, the weight corresponding to each first historical average precipitation index is , the value input to the target neural network model each time satisfy:

[0069] in, It is the first historical average precipitation index; Inputting multiple source precipitation indices into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are smaller than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, including: When the source precipitation index belongs to the high difference group, the value input to the target neural network model each time satisfy: ; When the source precipitation index belongs to the low difference group, the value input to the target neural network model each time satisfy:

[0070] in, is a source precipitation index belonging to the high difference group, is a source precipitation index belonging to the low-difference group, is the corresponding weight of the source precipitation index belonging to the high difference group, is the corresponding weight of the source precipitation index belonging to the low-discrepancy group, and > .

[0071] Optionally, the system may also be configured as: Multiple source precipitation indices are input into the target neural network model, wherein the weights corresponding to the input of multiple source precipitation indices belonging to the high difference group are less than the weights corresponding to the input of multiple source precipitation indices belonging to the low difference group, satisfying:

[0072] in, is the number of source precipitation indices belonging to the high-discrepancy group, is the number of source precipitation indices belonging to the high discrepancy group.

[0073] The present application proposes a regional precipitation prediction system based on big data, which obtains multiple first historical average precipitation indexes of multiple target precipitation sensors in previous years and multiple second historical average precipitation indexes of multiple source precipitation sensors in previous years based on big data, and then establishes a neural network model with the correlation between the first historical average precipitation index and the second historical average precipitation index, and imports the data obtained during monitoring into the model. During the import process, according to the difference between the monitoring data and the historical average data, the weight when inputting the neural network model is selectively increased or decreased to obtain the final prediction result. The present application proposes a regional precipitation prediction system based on big data, which predicts regional precipitation through a neural network model. When inputting the data into the neural network model, the credibility of the data is first judged. If the difference with the historical average data is large, the credibility is low, and vice versa. According to the credibility of the data, the corresponding high and low weights are set and input into the neural network model, so that the final result is more accurate.

[0074] Based on the same inventive concept, an embodiment of the present application further proposes an electronic device, the electronic device comprising: At least one processor; and a memory in communication with 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 automatic overheating protection method for a universal testing machine based on an embodiment of the present application.

[0075] In addition, to achieve the above-mentioned purpose, an embodiment of the present application further proposes a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the automatic overheating protection method based on a universal testing machine of an embodiment of the present application.

[0076] The following is a detailed introduction to the various components of electronic equipment: The processor is the control center of the electronic device, which can be a processor or a general term for multiple processing elements. For example, the processor is one or more central processing units (CPUs), or an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of the present invention, such as one or more microprocessors (digital signal processors, DSPs), or one or more field programmable gate arrays (FPGAs).

[0077] Optionally, the processor can perform various functions of the electronic device by running or executing a software program stored in the memory, and calling data stored in the memory.

[0078] The memory is used to store the software program for executing the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can refer to the above method embodiment and will not be repeated here.

[0079] Optionally, the memory may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM) or other types of dynamic storage devices that can store information and instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical disc, laser disc, optical disc, digital versatile disc, Blu-ray disc, etc.), a magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto. The memory may be integrated with the processor, or may exist independently and be coupled to the processor through an interface circuit of the electronic device, which is not specifically limited in the embodiments of the present invention.

[0080] A transceiver is used to communicate with a network device or a terminal device.

[0081] Optionally, the transceiver may include a receiver and a transmitter, wherein the receiver is used to implement a receiving function, and the transmitter is used to implement a sending function.

[0082] Optionally, the transceiver may be integrated with the processor, or may exist independently and be coupled to the processor via an interface circuit of the router, which is not specifically limited in the embodiment of the present invention.

[0083] In addition, the technical effects of the electronic device can refer to the technical effects of the data transmission method in the above method embodiment, which will not be repeated here.

[0084] It should be understood that the processor in the embodiment of the present invention may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.

[0085] It should also be understood that the memory in the embodiments of the present invention may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic random access memory (DRAM), synchronous DRAM (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0086] The above embodiments can be implemented in whole or in part by software, hardware (such as circuits), firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When a computer instruction or computer program is loaded or executed on a computer, a process or function according to an embodiment of the present invention is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, computer instructions can be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (such as infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0087] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0088] In the present invention, "at least one" means one or more, and "plurality" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0089] It should be understood that in various embodiments of the present invention, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0090] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

Claims

1. A regional precipitation prediction method based on big data, characterized in that: The method comprises: Determine a target area, the target area is an area where precipitation prediction is required, and determine a source data area based on the target area, the source data area is a range area surrounding the target area, wherein a plurality of precipitation sensors distributed in the target area are defined as target precipitation sensors, and a plurality of precipitation sensors distributed in the source data area are defined as source precipitation sensors; Acquire a plurality of first historical average precipitation indexes of a plurality of the target precipitation sensors in previous years and a plurality of second historical average precipitation indexes of a plurality of the source precipitation sensors in previous years based on the big data; Inputting a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices as training sets into a target neural network model and performing iterative calculations, and obtaining the trained target neural network model, wherein the target neural network model is used to output a plurality of predicted target precipitation indices according to the input plurality of source precipitation indices, wherein the plurality of the first historical average precipitation indices have the same weight when input into the target neural network model; Acquire a plurality of source precipitation indices acquired by a plurality of the source precipitation sensors in a first monitoring period; Comparing a plurality of the source precipitation indices with a plurality of corresponding first historical average precipitation indices, and dividing the plurality of the source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein an average difference between the plurality of the source precipitation indices belonging to the high difference group and a plurality of corresponding first historical average precipitation indices is greater than an average difference between the plurality of the source precipitation indices belonging to the low difference group and a plurality of corresponding first historical average precipitation indices; Inputting a plurality of the source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of the source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of the source precipitation indices belonging to the low difference group; Based on the output result of the target neural network model, the predicted target precipitation index is obtained, and the predicted precipitation is obtained based on the target precipitation index.

2. The regional precipitation prediction method based on big data according to claim 1 is characterized in that: Inputting a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices as training sets into a target neural network model and performing iterative calculations, and obtaining the trained target neural network model, wherein the target neural network model is used to output a plurality of predicted target precipitation indices according to the input plurality of source precipitation indices, wherein the plurality of the first historical average precipitation indices have the same weight when input into the target neural network model, including: Inputting a plurality of the first historical average precipitation indexes as a plurality of dimensions into the target neural network model, wherein the plurality of the dimensions constitute a feature vector of a second historical average precipitation index; Importing a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices into the target neural network model and performing iterative calculations, and during the calculation process, comparing the result of each calculation with the plurality of the second historical average precipitation indices, and obtaining an error value; When the error value is less than a preset threshold, the iterative calculation is stopped and the trained target neural network model is output.

3. The regional precipitation prediction method based on big data according to claim 2 is characterized in that: Importing a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices into the target neural network model and performing iterative calculations, and comparing the result of each calculation with the plurality of the second historical average precipitation indices during the calculation process, and obtaining an error value, including: Determining an initializer during the process of constructing the target neural network, and determining a plurality of identical weights based on the initializer; The first historical average precipitation index is input into the input layer of the target neural network based on the initializer.

4. The regional precipitation prediction method based on big data according to claim 3 is characterized in that: Importing a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices into the target neural network model and performing iterative calculations, and comparing the result of each calculation with the plurality of the second historical average precipitation indices during the calculation process, and obtaining an error value, including: If each result of each calculation during the calculation process is greater than each corresponding second historical average precipitation index, the error value satisfies: =( - ) in, is the error value, For each of the above results of each calculation, is the corresponding second historical average precipitation index for each.

5. The regional precipitation prediction method based on big data according to claim 3 is characterized in that: Importing a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices into the target neural network model and performing iterative calculations, and comparing the result of each calculation with the plurality of the second historical average precipitation indices during the calculation process, and obtaining an error value, including: If each of the results calculated during the calculation process is less than each of the corresponding second historical average precipitation indices, the error value satisfies: =( - ) in, is the error value, For each of the above results of each calculation, is the corresponding second historical average precipitation index for each.

6. The regional precipitation prediction method based on big data according to claim 3 is characterized in that: Inputting a plurality of the source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of the source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of the source precipitation indices belonging to the low difference group, comprises: Obtaining the initializer, and modifying the initializer based on different weights; A plurality of source precipitation indices are input into the input layer of the target neural network based on the modified initializer.

7. The regional precipitation prediction method based on big data according to claim 6 is characterized in that: Inputting a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices as training sets into a target neural network model and performing iterative calculations, and obtaining the trained target neural network model, wherein the target neural network model is used to output a plurality of predicted target precipitation indices according to the input plurality of source precipitation indices, wherein the plurality of the first historical average precipitation indices have the same weight when input into the target neural network model, including: Total The first historical average precipitation index, and the weight corresponding to each of the first historical average precipitation index is , each time the value input to the target neural network model satisfy: in, is the first historical average precipitation index; Inputting a plurality of the source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of the source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of the source precipitation indices belonging to the low difference group, comprises: When the source precipitation index belongs to the high difference group, the value input to the target neural network model each time satisfy: ; When the source precipitation index belongs to the low difference group, the value input to the target neural network model each time satisfy: in, is one of the source precipitation indices belonging to the high difference group, is one of the source precipitation indices belonging to the low-discrepancy group, is the corresponding weight of the source precipitation index belonging to the high difference group, is the corresponding weight of the source precipitation index belonging to the low-discrepancy group, and > .

8. The regional precipitation prediction method based on big data according to claim 7 is characterized in that: A plurality of the source precipitation indices are input into the target neural network model, wherein the weight corresponding to the input of the plurality of the source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of the source precipitation indices belonging to the low difference group, satisfying: in, is the number of the source precipitation indices belonging to the high discrepancy group, is the number of the source precipitation indices belonging to the high difference group.

9. A regional precipitation prediction system based on big data, characterized in that: The system includes: A first submodule, the first submodule is used to determine a target area, the target area is an area where precipitation prediction is required, and determine a source data area based on the target area, the source data area is a range area surrounding the target area, wherein a plurality of precipitation sensors distributed in the target area are defined as target precipitation sensors, and a plurality of precipitation sensors distributed in the source data area are defined as source precipitation sensors; A second submodule, the second submodule is used to obtain a plurality of first historical average precipitation indexes of a plurality of target precipitation sensors in previous years and a plurality of second historical average precipitation indexes of a plurality of source precipitation sensors in previous years based on big data; A third submodule, the third submodule is used to input a plurality of the first historical average precipitation indices and a plurality of the second historical average precipitation indices as training sets into a target neural network model and perform iterative calculations, and obtain the trained target neural network model, the target neural network model is used to output a plurality of predicted target precipitation indices according to the input plurality of source precipitation indices, wherein the plurality of the first historical average precipitation indices have the same weight when input into the target neural network model; A fourth submodule, the fourth submodule is used to obtain a plurality of source precipitation indices obtained by the plurality of source precipitation sensors in the first monitoring period; A fifth submodule, the fifth submodule is used to compare the plurality of source precipitation indices with the plurality of corresponding first historical average precipitation indices, and divide the plurality of source precipitation indices into a high difference group and a low difference group according to the comparison result, wherein the average difference between the plurality of source precipitation indices belonging to the high difference group and the plurality of corresponding first historical average precipitation indices is greater than the average difference between the plurality of source precipitation indices belonging to the low difference group and the plurality of corresponding first historical average precipitation indices; A sixth submodule, wherein the sixth submodule inputs a plurality of the source precipitation indices into the target neural network model, wherein the weight corresponding to the input of the plurality of the source precipitation indices belonging to the high difference group is less than the weight corresponding to the input of the plurality of the source precipitation indices belonging to the low difference group; The seventh submodule obtains the predicted target precipitation index based on the output result of the target neural network model, and obtains the predicted precipitation based on the target precipitation index.

10. An electronic device, characterized in that: The electronic device comprises: at least one processor; And, a memory communicatively connected to at least one of the processors; wherein the memory stores instructions executable by at least one of the processors, and the instructions are executed by at least one of the processors so that at least one of the processors can execute the method proposed in any one of claims 1-8.

Images