Intelligent early warning system and method based on simulation reasoning
Through deep learning technology, the analysis and significance optimization of rainfall timing correlation characteristics and significance optimization in the intelligent early warning system have solved the problem that traditional early warning systems cannot effectively explore the deep correlation of time series data, and improved the accuracy of rainfall prediction and the accuracy of early warning system. sex.
Patent Information
- Application Number
- CN202411192415.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-28
- Publication Date
- 2024-12-06
AI Technical Summary
Traditional early warning systems are based on statistical methods and simple mathematical models, and cannot fully explore deep correlations in complex time series data, resulting in limited accuracy of rainfall prediction.
Using an intelligent early warning system based on deep learning, through rainfall timing correlation feature extraction, significant information processing and timing propagation aggregation module, the cumulative precipitation prediction value of rainfall is automatically obtained and an early warning prompt is generated.
It improves the accuracy of rainfall prediction, can better simulate and predict the changing trend of rainfall, and enhances the accuracy of the early warning system.
Smart Images

Figure CN119091589A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of intelligent early warning technology, and specifically, to an intelligent early warning system and method based on simulation reasoning. Background Art
[0002] As the impact of global climate change becomes increasingly significant, the frequency and intensity of extreme weather events such as rainstorms and floods are increasing. These disasters not only pose a serious threat to people's lives and property, but also cause huge losses to the social economy. Therefore, the development of an efficient and accurate early warning system is crucial to mitigate the impact of disasters.
[0003] However, traditional early warning systems are often based on statistical methods and simple mathematical models for data analysis. This processing method cannot fully explore the deep correlations in complex time series data. In addition, traditional methods rely on limited historical data sets and relatively simple prediction models, which limits the accuracy of prediction results.
[0004] Therefore, an optimized early warning system is desired. Summary of the invention
[0005] This summary is provided to introduce concepts in a brief form that will be described in detail in the detailed description below. This summary is not intended to identify key features or essential features of the claimed technical solution, nor is it intended to limit the scope of the claimed technical solution.
[0006] In a first aspect, the present application provides an intelligent early warning system based on simulation reasoning, the system comprising:
[0007] The rainfall time series acquisition module is used to obtain the time queue of rainfall;
[0008] A rainfall time series correlation feature extraction module is used to input the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series correlation feature vectors;
[0009] A rainfall time series correlation feature maximum value extraction module is used to extract the maximum value of each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors to obtain a rainfall time series significant information representation vector;
[0010] A rainfall time series significant information activation processing module is used to activate the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector;
[0011] A rainfall time series correlation significant optimization module is used to weight each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors by taking the feature value of each position in the rainfall time series significant guide weight vector as a weight to obtain a sequence of significantly optimized rainfall time series correlation feature vectors;
[0012] A rainfall time series propagation aggregation module is used to input the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on the significant time series attenuation of node energy to obtain a rainfall time series propagation aggregation representation vector as a rainfall time series propagation aggregation representation feature;
[0013] The cumulative precipitation prediction module is used to obtain a cumulative precipitation prediction value based on the rainfall time series propagation aggregation representation characteristics, and determine whether to generate an early warning prompt based on the cumulative precipitation prediction value.
[0014] Optionally, the rainfall time series correlation feature extraction module is used to: input the rainfall time queue into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series correlation feature vectors.
[0015] Optionally, the rainfall time series significant information activation processing module is used to: input the rainfall time series significant information representation vector into a Sigmoid activation function to obtain the rainfall time series significant guidance weight vector.
[0016] Optionally, the rainfall time series propagation aggregation module includes: a significantly optimized rainfall energy significant descriptor calculation unit, which is used to calculate the energy significant descriptors of each significantly optimized rainfall time series associated feature vector in the sequence of the significantly optimized rainfall time series associated feature vector based on the mean and standard deviation of each significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall time series energy significant descriptors; a timestamp extraction unit, which is used to extract the timestamps of each significantly optimized rainfall time series associated feature vector in the sequence of the significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall timestamps; and a significantly optimized rainfall feature vector acquisition unit, which is used to use the significantly optimized rainfall time series associated feature vector corresponding to the current time point in the sequence of the significantly optimized rainfall time series associated feature vector as the current node feature vector and use the significantly optimized rainfall time series associated feature vectors corresponding to other time points as the historical node feature vector to obtain the current significantly optimized rainfall time series associated feature vector and the historical node feature vector. A sequence of significantly optimized rainfall time-series associated feature vectors; a significantly optimized rainfall energy time decay factor calculation unit, used to calculate the energy time decay factor of each historically significantly optimized rainfall time-series associated feature vector in the sequence of historically significantly optimized rainfall time-series associated feature vectors relative to the current significantly optimized rainfall time-series associated feature vector to obtain a sequence of significantly optimized rainfall energy time decay factors; a historically significantly optimized rainfall time-series propagation unit, used to use the sequence of significantly optimized rainfall time-series energy significant descriptors as positive adjustment factors and the sequence of significantly optimized rainfall energy time decay factors as negative adjustment factors to perform time-series propagation aggregation on the sequence of historically significantly optimized rainfall time-series associated feature vectors to obtain a historically significantly optimized rainfall time-series propagation aggregation representation feature vector; a vector fusion unit, used to fuse the historically significantly optimized rainfall time-series propagation aggregation representation feature vector and the current significantly optimized rainfall time-series associated feature vector to obtain the rainfall time-series propagation aggregation representation vector.
[0017] Optionally, the significantly optimized rainfall energy significant descriptor calculation unit is used to: respectively calculate the mean and standard deviation of the significantly optimized rainfall time series association feature vector to obtain the significantly optimized rainfall time series association feature mean and the significantly optimized rainfall time series association feature standard deviation; calculate the positional difference between the significantly optimized rainfall time series association feature vector and the significantly optimized rainfall time series association feature mean to obtain the significantly optimized rainfall time series association difference vector; calculate the fourth power of each eigenvalue in the significantly optimized rainfall time series association difference vector to obtain the significantly optimized rainfall time series association modulation difference vector; calculate the expected value of the significantly optimized rainfall time series association modulation difference vector to obtain the significantly optimized rainfall time series association expected value; divide the significantly optimized rainfall time series association expected value by the fourth power of the significantly optimized rainfall time series association feature standard deviation to obtain the significantly optimized rainfall time series energy significant descriptor.
[0018] Optionally, the significantly optimized rainfall energy time decay factor calculation unit is used to: subtract the significantly optimized rainfall timestamp corresponding to the current significantly optimized rainfall time series associated characteristic vector from the significantly optimized rainfall timestamp corresponding to each historical significantly optimized rainfall time series associated characteristic vector, and then round down to obtain a sequence of historical significantly optimized rainfall time span values; calculate the square of the sequence of historical significantly optimized rainfall time span values divided by the time attenuation inverse scaling parameter by position to obtain a sequence of historical significantly optimized rainfall time span square modulation values, and then multiply the sequence of historical significantly optimized rainfall time span square modulation values by the attenuation rate positive scaling parameter by position to obtain a sequence of significantly optimized rainfall time series span energy decay coefficients; use each significantly optimized rainfall time series span energy decay coefficient in the sequence of significantly optimized rainfall time series span energy decay coefficients as an exponential power, and calculate the natural exponential function value with the natural constant e as the base to obtain the sequence of significantly optimized rainfall energy time decay factors.
[0019] Optionally, the historical significant optimized rainfall time series propagation unit is used to: divide the sequence of the significant optimized rainfall time series energy significant descriptors by the sequence of the significant optimized rainfall energy time attenuation factors by position to obtain a sequence of significant optimized rainfall adjustment factors, and calculate the positionally weighted sum of the sequence of the significant optimized rainfall adjustment factors and the sequence of the historical significant optimized rainfall time series associated feature vectors to obtain the historical significant optimized rainfall time series propagation aggregation representation feature vector.
[0020] Optionally, the cumulative precipitation prediction module includes: a decoding unit, used to input the rainfall time series propagation aggregation representation vector into a decoder-based cumulative precipitation predictor to obtain a cumulative precipitation prediction decoding value; and an early warning determination unit, used to determine whether to generate the early warning prompt based on a comparison between the cumulative precipitation prediction decoding value and a preset threshold.
[0021] In a second aspect, the present application provides an intelligent early warning method based on simulation reasoning, the method comprising:
[0022] Get the time queue of rainfall;
[0023] Inputting the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors;
[0024] Extracting the maximum value of each rainfall time series associated feature vector in the sequence of rainfall time series associated feature vectors to obtain a rainfall time series significant information representation vector;
[0025] Activating the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector;
[0026] Using the characteristic values of each position in the rainfall time series significant guide weight vector as weight, weighting each rainfall time series associated characteristic vector in the sequence of rainfall time series associated characteristic vectors to obtain a sequence of significantly optimized rainfall time series associated characteristic vectors;
[0027] Inputting the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on the significant time series attenuation of node energy to obtain a rainfall time series propagation aggregate representation vector as a rainfall time series propagation aggregate representation feature;
[0028] Based on the rainfall time series propagation aggregation representation characteristics, a cumulative precipitation forecast value is obtained, and based on the cumulative precipitation forecast value, it is determined whether to generate an early warning prompt.
[0029] Optionally, the time queue of rainfall is input into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors, including: inputting the time queue of rainfall into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series associated feature vectors.
[0030] By adopting the above technical solution, by obtaining the time queue of rainfall, and using data analysis and processing technology based on deep learning to analyze the time series correlation characteristics and optimize the significance of the rainfall, the cumulative precipitation forecast value is automatically obtained based on the time series propagation aggregation characteristics of the rainfall in each local time period, and it is determined whether to generate an early warning prompt. In this way, using deep learning technology, it is possible to mine the deep correlations and patterns in time series data, including nonlinear relationships and complex interaction effects, and better simulate and predict the changing trend of rainfall, thereby improving the accuracy of rainfall prediction.
[0031] Other features and advantages of the present application will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] The above and other features, advantages and aspects of the embodiments of the present application will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale. In the drawings:
[0033] Figure 1 It is a block diagram of an intelligent early warning system based on simulation reasoning according to an exemplary embodiment.
[0034] Figure 2 The present invention is a flowchart of an intelligent early warning method based on simulation reasoning according to an exemplary embodiment.
[0035] Figure 3 is a schematic diagram of panoramic reasoning results according to an exemplary embodiment.
[0036] Figure 4 It is a block diagram of an electronic device according to an exemplary embodiment.
[0037] Figure 5 It is an application scenario diagram of an intelligent early warning system based on simulation reasoning according to an exemplary embodiment. DETAILED DESCRIPTION
[0038] The embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be construed as being limited to the embodiments described herein. Instead, these embodiments are provided to provide a more thorough and complete understanding of the present application. It should be understood that the drawings and embodiments of the present application are only for exemplary purposes and are not intended to limit the scope of protection of the present application.
[0039] It should be understood that the various steps described in the method implementation of the present application can be performed in different orders and / or performed in parallel. In addition, the method implementation may include additional steps and / or omit the steps shown. The scope of the present application is not limited in this respect.
[0040] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.
[0041] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0042] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0043] The names of the messages or information exchanged between multiple devices in the embodiments of the present application are only used for illustrative purposes and are not used to limit the scope of these messages or information.
[0044] The specific implementation of the present application is described in detail below with reference to the accompanying drawings.
[0045] Figure 1 FIG. 1 is a block diagram of an intelligent early warning system based on simulation reasoning according to an exemplary embodiment. Figure 1 As shown, the intelligent early warning system 100 includes:
[0046] A rainfall time series acquisition module 101 is used to obtain a time series of rainfall;
[0047] A rainfall time series correlation feature extraction module 102 is used to input the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series correlation feature vectors;
[0048] A rainfall time series correlation feature maximum value extraction module 103 is used to extract the maximum value of each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors to obtain a rainfall time series significant information representation vector;
[0049] A rainfall time series significant information activation processing module 104 is used to activate the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector;
[0050] A rainfall time series correlation significant optimization module 105 is used to weight each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors by taking the feature value of each position in the rainfall time series significant guide weight vector as a weight to obtain a sequence of significantly optimized rainfall time series correlation feature vectors;
[0051] A rainfall time series propagation aggregation module 106 is used to input the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on the significant time series attenuation of node energy to obtain a rainfall time series propagation aggregation representation vector as a rainfall time series propagation aggregation representation feature;
[0052] The cumulative precipitation prediction module 107 is used to obtain a cumulative precipitation prediction value based on the rainfall time series propagation aggregation representation characteristics, and determine whether to generate a warning prompt based on the cumulative precipitation prediction value.
[0053] Based on this, the present application proposes an intelligent early warning system based on simulation reasoning, which obtains the time queue of rainfall and uses data analysis and processing technology based on deep learning to analyze the time series correlation characteristics and optimize the significance of the rainfall, so as to automatically obtain the cumulative precipitation forecast value based on the time series propagation aggregation characteristics of the rainfall in each local time period, and determine whether to generate an early warning prompt. In this way, using deep learning technology, it is possible to mine the deep correlations and patterns in time series data, including nonlinear relationships and complex interaction effects, and better simulate and predict the changing trend of rainfall, thereby improving the accuracy of rainfall prediction.
[0054] Specifically, in the technical solution of the present application, first, a time queue of rainfall is obtained. Then, considering that there are fluctuations or dynamic changes in the time series characteristics within a local time range in the time queue of rainfall, and considering that the 1D-CNN model has a good perception and capture ability for time series data. Based on this, in the technical solution of the present application, the time queue of rainfall is input into a rainfall sequence encoder based on the 1D-CNN model to capture and extract the rainfall time series change feature information within each local time segment, and obtain a sequence of rainfall time series associated feature vectors.
[0055] In one embodiment of the present application, the rainfall time series correlation feature extraction module is used to: input the rainfall time queue into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series correlation feature vectors.
[0056] Then, considering that each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors contains key time series feature information of rainfall in a local time period, therefore, in the technical solution of the present application, the maximum value of each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors is extracted to obtain a rainfall time series significant information representation vector. In other words, the maximum value can reflect the most significant feature information in the rainfall time series, and compared with the average value or other statistics, the maximum value can better highlight the key features in the time series.
[0057] Next, considering that the importance and influence of each value in the rainfall time series significant information representation vector on precipitation prediction are different. Therefore, in order to be able to focus more on key and significant values, in the technical solution of the present application, the rainfall time series significant information representation vector is input into the Sigmoid activation function to obtain the rainfall time series significant guidance weight vector. In this way, each value in the rainfall time series significant information representation vector can be converted into a corresponding weight through the Sigmoid activation function, so that the model can focus more on relatively important or significant rainfall time series correlation features, thereby helping the system to better focus on and utilize key information in the time series, and improve the accuracy of subsequent analysis and rainfall prediction.
[0058] In one embodiment of the present application, the rainfall time series significant information activation processing module is used to: input the rainfall time series significant information representation vector into a Sigmoid activation function to obtain the rainfall time series significant guidance weight vector.
[0059] Then, considering that in the sequence of rainfall time series correlation feature vectors, each vector represents the rainfall feature information at a certain time point, but the importance of these feature vectors is not exactly the same, therefore, it is necessary to distinguish them according to their relative significance in the entire time series, so as to focus on the important rainfall time series correlation features. Specifically, in the technical solution of the present application, the characteristic values of each position in the rainfall time series significant guide weight vector are used as weights, and each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors is weighted to obtain a sequence of significantly optimized rainfall time series correlation feature vectors. In this way, those more significant rainfall time series correlation features can be highlighted, and the influence of those less important features can be relatively reduced, so as to further optimize and refine the original rainfall time series correlation features, thereby improving the overall analysis and prediction performance.
[0060] In one embodiment of the present application, the rainfall time series propagation aggregation module includes: a significantly optimized rainfall energy significant descriptor calculation unit, which is used to calculate the energy significant descriptors of each significantly optimized rainfall time series associated feature vector in the sequence of the significantly optimized rainfall time series associated feature vector based on the mean and standard deviation of each significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall time series energy significant descriptors; a timestamp extraction unit, which is used to extract the timestamps of each significantly optimized rainfall time series associated feature vector in the sequence of the significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall timestamps; a significantly optimized rainfall feature vector acquisition unit, which is used to use the significantly optimized rainfall time series associated feature vector corresponding to the current time point in the sequence of the significantly optimized rainfall time series associated feature vector as the current node feature vector and use the significantly optimized rainfall time series associated feature vectors corresponding to other time points as the historical node feature vector to obtain the current significantly optimized rainfall time series associated feature vector. A significantly optimized rainfall energy time decay factor calculation unit is used to calculate the energy time decay factor of each historically significantly optimized rainfall time series associated feature vector in the sequence of the historically significantly optimized rainfall time series associated feature vector relative to the current significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall energy time decay factors; a historically significantly optimized rainfall time series propagation unit is used to use the sequence of the significantly optimized rainfall time series energy significant descriptors as positive adjustment factors and the sequence of the significantly optimized rainfall energy time decay factors as negative adjustment factors to perform time series propagation aggregation on the sequence of the historically significantly optimized rainfall time series associated feature vectors to obtain a historically significantly optimized rainfall time series propagation aggregation representation feature vector; a vector fusion unit is used to fuse the historically significantly optimized rainfall time series propagation aggregation representation feature vector and the current significantly optimized rainfall time series associated feature vector to obtain the rainfall time series propagation aggregation representation vector.
[0061] Further, considering that the different significantly optimized rainfall time series associated features in the sequence of the significantly optimized rainfall time series associated feature vectors have different dynamic change relationships in the entire time series process, that is, as time goes by, the earlier time point has less impact on the final result, while the closer time point has greater impact. Therefore, in order to better integrate the information in the entire time domain based on the characteristics of each local time point, in the technical solution of the present application, the sequence of the significantly optimized rainfall time series associated feature vectors is input into the feature forward propagation network based on the significant time series decay of node energy to obtain the rainfall time series propagation aggregation representation vector. It is worth mentioning that the feature forward propagation network based on the significant time series decay of node energy aggregates and generates the propagation aggregation representation vector by focusing on the significance and time series decay characteristics of node energy. Specifically, first, based on the expected value and standard deviation of each significantly optimized rainfall time series associated feature vector, the energy significant descriptor of each feature is obtained to quantify the importance and stability of the feature vector at each time point. Then, the timestamps of each significantly optimized rainfall time series associated feature vector are extracted to ensure that the time order of each feature vector is retained and the time series change trend of each rainfall feature is understood. Next, the node feature vector at the current time point in the sequence of the significantly optimized rainfall time series associated feature vector is regarded as the current significantly optimized rainfall time series associated feature vector, and the others are regarded as the historical significantly optimized rainfall time series associated feature vectors to distinguish the current and historical node feature vectors. Subsequently, the energy time decay factor of the historical node feature vector relative to the current node feature vector is calculated to quantify the degree of attenuation of the influence of rainfall characteristics over time. Next, the energy significant descriptor is used as a positive adjustment factor and the energy time decay factor is used as a negative adjustment factor to perform weighted aggregation on each historically significantly optimized rainfall time series associated feature vector to enhance the features that are significantly related and critical to the current node features in the historical node features, while reasonably balancing the importance of early and recent feature information to obtain the historically significantly optimized rainfall time series propagation aggregation representation feature vector. Finally, the historically significantly optimized rainfall time series propagation aggregation representation feature vector and the current significantly optimized rainfall time series associated feature vector are fused to obtain the rainfall time series propagation aggregation representation vector that can fully reflect the rainfall time series evolution process.
[0062] In one embodiment of the present application, the significantly optimized rainfall energy significant descriptor calculation unit is used to: respectively calculate the mean and standard deviation of the significantly optimized rainfall time series association feature vector to obtain the significantly optimized rainfall time series association feature mean and the significantly optimized rainfall time series association feature standard deviation; calculate the positional difference between the significantly optimized rainfall time series association feature vector and the significantly optimized rainfall time series association feature mean to obtain a significantly optimized rainfall time series association difference vector; calculate the fourth power of each eigenvalue in the significantly optimized rainfall time series association difference vector to obtain a significantly optimized rainfall time series association modulation difference vector; calculate the expected value of the significantly optimized rainfall time series association modulation difference vector to obtain a significantly optimized rainfall time series association expected value; divide the significantly optimized rainfall time series association expected value by the fourth power of the significantly optimized rainfall time series association feature standard deviation to obtain the significantly optimized rainfall time series energy significant descriptor.
[0063] Further, in one embodiment of the present application, the significantly optimized rainfall energy time decay factor calculation unit is used to: respectively subtract the significantly optimized rainfall timestamp corresponding to the current significantly optimized rainfall time series associated characteristic vector from the significantly optimized rainfall timestamp corresponding to each historical significantly optimized rainfall time series associated characteristic vector, and then round down to obtain a sequence of historical significantly optimized rainfall time span values; calculate the square of the sequence of historical significantly optimized rainfall time span values divided by the time attenuation inverse scaling parameter by position to obtain a sequence of historical significantly optimized rainfall time span square modulation values, and then multiply the sequence of historical significantly optimized rainfall time span square modulation values by the attenuation rate positive scaling parameter by position to obtain a sequence of significantly optimized rainfall time series span energy decay coefficients; use each significantly optimized rainfall time series span energy decay coefficient in the sequence of significantly optimized rainfall time series span energy decay coefficients as an exponential power, and calculate the natural exponential function value with the natural constant e as the base to obtain the sequence of significantly optimized rainfall energy time decay factors.
[0064] Furthermore, in one embodiment of the present application, the historical significant optimized rainfall timing propagation unit is used to: divide the sequence of the significantly optimized rainfall timing energy significant descriptors and the sequence of the significantly optimized rainfall energy time attenuation factors by position to obtain a sequence of significantly optimized rainfall adjustment factors, and calculate the positionally weighted sum of the sequence of the significantly optimized rainfall adjustment factors and the sequence of the historical significantly optimized rainfall timing associated feature vectors to obtain the historical significantly optimized rainfall timing propagation aggregation representation feature vector.
[0065] Specifically, the sequence of the significantly optimized rainfall time series associated feature vectors is input into a feature forward propagation network based on the significant time series attenuation of node energy, and processed using the following forward propagation formula to obtain the rainfall time series propagation aggregation representation vector; wherein the forward propagation formula is:
[0066] X={x1,x2,…x k-1 , x k}
[0067]
[0068] Wherein, X is the sequence of the significantly optimized rainfall time series correlation feature vector, x k-1 and x k are respectively the (k-1)th and kth significantly optimized rainfall time series associated feature vectors in the sequence of significantly optimized rainfall time series associated feature vectors, k represents the number of feature vectors in the sequence of significantly optimized rainfall time series associated feature vectors, is the eigenvalue of each position in the ith significantly optimized rainfall time series associated feature vector in the sequence of the significantly optimized rainfall time series associated feature vector, E(·) is the calculated expected value, μ i and σ i 2 are the mean and variance of the ith significantly optimized rainfall time series associated feature vector, L is the number of eigenvalues in each significantly optimized rainfall time series associated feature vector, is the energy-significant descriptor of the kth significantly optimized rainfall temporal correlation feature vector, is the energy-significant descriptor of the i-th significantly optimized rainfall temporal correlation feature vector, t k and t i The distribution represents the timestamps of the kth and ith significantly optimized rainfall time series associated feature vectors, is a rounding operation, δ is the time decay inverse scaling parameter, β is the decay rate positive scaling parameter, θ and γ are weight hyperparameters, exp(·) represents the exponential function value with the natural constant e as the base, and f c is the rainfall time series propagation aggregation representation vector.
[0069] Then, the rainfall time series propagation aggregation representation vector is input into the decoder-based cumulative precipitation predictor to obtain the cumulative precipitation prediction decoded value, and based on the comparison between the cumulative precipitation prediction decoded value and the preset threshold, it is determined whether to generate an early warning prompt. That is, the rainfall time series propagation aggregation representation feature obtained by performing time series propagation aggregation using the sequence of significantly optimized rainfall time series associated feature vectors is decoded to automatically obtain the cumulative precipitation prediction value and determine whether to generate an early warning prompt. In this way, using deep learning technology, it is possible to mine deep correlations and patterns in time series data, including nonlinear relationships and complex interaction effects, and to better simulate and predict the changing trend of rainfall, thereby improving the accuracy of rainfall prediction.
[0070] In one embodiment of the present application, the cumulative precipitation prediction module includes: a decoding unit, used to input the rainfall time series propagation aggregation representation vector into a decoder-based cumulative precipitation predictor to obtain a cumulative precipitation prediction decoded value; an early warning determination unit, used to determine whether to generate the early warning prompt based on a comparison between the cumulative precipitation prediction decoded value and a preset threshold.
[0071] Preferably, inputting the rainfall time series propagation aggregation representation vector into a decoder-based cumulative precipitation predictor to obtain a cumulative precipitation prediction decoded value comprises:
[0072] Subtract the zero norm of the rainfall time series propagation aggregation representation vector from the length of the rainfall time series propagation aggregation representation vector to obtain a rainfall time series propagation aggregation isolated representation value;
[0073] Calculate the logarithm of the sum of the square of the rainfall time series propagation aggregated isolated representation value and the rainfall time series propagation aggregated isolated representation value with base 2 to obtain the rainfall time series propagation aggregated information order value;
[0074] Calculate a power function with each eigenvalue of the rainfall time series propagation aggregation representation vector as the base, the difference of the rainfall time series propagation aggregation isolated representation value minus one as the exponential, and perform point multiplication with the rainfall time series propagation aggregation information order value to obtain the rainfall time series propagation aggregation leading vector;
[0075] After performing a dot multiplication on the rainfall time series propagation aggregation representation vector and the rainfall time series propagation aggregation isolated representation value minus one, the dot multiplication is performed on the inverse of the rainfall time series propagation aggregation isolated representation value to obtain a rainfall time series propagation aggregation field constraint vector;
[0076] Calculate an exponential function with a natural constant as the base and each eigenvalue of the rainfall time series propagation aggregation field constraint vector as the exponent to obtain a rainfall time series propagation aggregation field bias vector;
[0077] Performing dot addition of the rainfall time series propagation aggregation leading vector and the rainfall time series propagation aggregation field bias vector to obtain an optimized rainfall time series propagation aggregation representation vector; and
[0078] The optimized rainfall time series propagation aggregation representation vector is input into a decoder-based cumulative precipitation predictor to obtain a cumulative precipitation prediction decoded value.
[0079] Specifically, the optimization process of the rainfall time series propagation aggregation representation vector, for example, denoted as V, is expressed as:
[0080]
[0081] n=S-||V||0
[0082] Among them, V ⊙(n-1) To calculate the power function with each eigenvalue of the rainfall time series propagation aggregation representation vector as the base, the difference of the rainfall time series propagation aggregation isolated representation value minus one is the exponential, V is the rainfall time series propagation aggregation representation vector, ||·||0 is the 0 norm of the vector, S is the length of the rainfall time series propagation aggregation representation vector, n is the rainfall time series propagation aggregation isolated representation value, log is the logarithm with base 2, exp[·] is the exponential operation of the vector, the exponential operation of the vector represents the natural exponential function value with the eigenvalues of each position in the vector raised to the power, V′ is the optimized rainfall time series propagation aggregation representation vector, ⊙ is the point multiplication by position, It is added by location point.
[0083] Here, the sequence of significantly optimized rainfall time series correlation feature vectors represents the local time domain one-dimensional time series correlation feature of rainfall based on the extreme value guided enhancement feature of the local time domain time series distribution significance. Therefore, when performing the decreasing forward propagation based on the energy significance of the time series nodes, the time series propagation aggregation deviation caused by the multiple reinforcement of the time series feature distribution in the local time domain and the imbalance of the decreasing energy significance is taken into account. It is expected to further improve the full time domain propagation aggregation regression comprehensibility of the rainfall time series propagation aggregation feature vector, thereby improving the accuracy of the cumulative precipitation prediction decoding value obtained by the cumulative precipitation predictor based on the decoder input.
[0084] Therefore, for the high-dimensional feature manifold of the rainfall time series propagation aggregation representation vector, the feature value of the feature set is used as the vector field representation of the aggregation dimension, and the superposition value of the vector field of the rainfall time series propagation aggregation representation vector at the isolated zero position is used as the order information to fix the local position of the feature value of its feature set, and add the bias of the reversibility of the feature regression distribution field of the rainfall time series propagation aggregation representation vector as a reward to achieve the mapping target tracking of the regression distribution of the rainfall time series propagation aggregation representation vector for the feature value position, so that the feature set of the rainfall time series propagation aggregation representation vector perceives the mapping migration to the aggregation distribution, thereby improving the accuracy of the cumulative precipitation prediction decoding value obtained by the rainfall time series propagation aggregation feature vector input based on the decoder by improving the full time domain propagation aggregation regression comprehensibility of the rainfall time series propagation aggregation representation vector. In this way, using deep learning technology, it is possible to mine the deep correlation and pattern in time series data, including nonlinear relationships and complex interaction effects, and can better simulate and predict the changing trend of rainfall, thereby improving the accuracy of rainfall prediction.
[0085] In summary, the above scheme is adopted to obtain the time queue of rainfall, and adopt the data analysis and processing technology based on deep learning to analyze the time series correlation characteristics and optimize the significance of the rainfall, so as to automatically obtain the cumulative precipitation forecast value based on the time series propagation aggregation characteristics of the rainfall in each local time period, and determine whether to generate an early warning prompt. In this way, using deep learning technology, it is possible to mine the deep correlations and patterns in time series data, including nonlinear relationships and complex interaction effects, and better simulate and predict the changing trend of rainfall, thereby improving the accuracy of rainfall prediction.
[0086] Figure 2 is a flow chart of an intelligent early warning method based on simulation reasoning according to an exemplary embodiment. Figure 2 As shown, the method includes:
[0087] Step S201, obtaining a time queue of rainfall;
[0088] Step S202, inputting the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors;
[0089] Step S203, extracting the maximum value of each rainfall time series associated feature vector in the sequence of rainfall time series associated feature vectors to obtain a rainfall time series significant information representation vector;
[0090] Step S204, activating the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector;
[0091] Step S205, using the characteristic value of each position in the rainfall time series significant guide weight vector as a weight, weighting each rainfall time series associated characteristic vector in the sequence of rainfall time series associated characteristic vectors to obtain a sequence of significantly optimized rainfall time series associated characteristic vectors;
[0092] Step S206, inputting the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on significant time series attenuation of node energy to obtain a rainfall time series propagation aggregate representation vector as a rainfall time series propagation aggregate representation feature;
[0093] Step S207: obtaining a cumulative precipitation forecast value based on the rainfall time series propagation aggregation representation characteristics, and determining whether to generate a warning prompt based on the cumulative precipitation forecast value.
[0094] In one embodiment of the present application, the time queue of rainfall is input into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors, including: inputting the time queue of rainfall into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series associated feature vectors.
[0095] Among them, the panoramic reasoning results of simulation reasoning refer to Figure 3 As shown in the figure, the x-axis represents time, each curve represents a monitoring item, and the y-axis represents the value of the monitoring item. The x-axis is the time period, the y-axis is the different monitoring items, and the z-axis is the normalized processing result of the monitoring value. The z-axis is the scale factor to ensure that all monitoring values can be displayed in the same figure.
[0096] Assume that all monitoring items are X = (x1, x2, x3...xn), the predicted time is divided into m segments, T = (1, 2, 3, ..., m) xij represents the predicted value of monitoring item xi at time j. Sk(xi) represents the time series predicted value of xi from time 1 to time k.
[0097] The execution process of this algorithm is as follows:
[0098] 1) Based on the calculation priority from high to low, the mechanism formula and non-mechanism model services are sorted. Each mechanism formula and non-mechanism model service can calculate the predicted value of one or thousands of monitoring items. The set of prediction calculation rules is expressed as RULES.
[0099] 2) Perform the following calculations for each moment:
[0100] a) Generate constraints LIMITS(X) for this round based on inference rules;
[0101] b) Predict the predicted values of all monitoring items at the current moment based on the prediction calculation rule set.
[0102] For rule k in RULES
[0103] xi k = rule_k(X k-1 , S k-1 (xi)), that is, the predicted value of xi at the current moment is calculated based on the historical time series prediction value of xi and the values of all monitoring items at the previous moment; this calculation includes the trend information of xi itself and the surrounding environment information.
[0104] If k If it does not exist, update xi k The value of
[0105] If k If it already exists, then determine xi according to the constraint rule k , whether the value is valid; if valid, add a new xi k If it is invalid, the calculation result will be ignored.
[0106] Point rule: An early warning that targets the predicted value of a single monitoring item at a certain moment.
[0107] Line rules: 1) Early warning based on the predicted value of a single monitoring item within a time period: for example, if the water level amplitude from t0 to t8 exceeds 2 meters, that is, max(xi)-min(xi)>2; 2) and early warning based on the predicted value of all monitoring items at a single moment; for example, at ti, rainfall intensity*1 hour>flow rate through the gate, where rainfall intensity and flow rate through the gate are two monitoring items.
[0108] Surface rules: Use the aggregated calculation results of time periods as comparison items to establish early warning rules between different monitoring items: for example, the cumulative rainfall from t0 to t8 > the cumulative flow through the gate, where rainfall and flow through the gate are monitoring items, and cumulative rainfall and cumulative flow through the gate are aggregated calculation results. The point, line, and surface early warning rules greatly enhance the flexibility and practicality of the early warning.
[0109] Reference below Figure 4 , which shows a schematic diagram of the structure of an electronic device 600 suitable for implementing the embodiment of the present application. The terminal device in the embodiment of the present application may include but is not limited to mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), vehicle-mounted terminals (such as vehicle-mounted navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 4 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.
[0110] like Figure 4 As shown, the electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 are also stored. The processing device 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.
[0111] Typically, the following devices may be connected to the I / O interface 605: an input device 606 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 607 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 608 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 609. The communication device 609 may allow the electronic device 600 to communicate with other devices wirelessly or by wire to exchange data. Although Figure 4 The electronic device 600 is shown with various devices, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.
[0112] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602. When the computer program is executed by the processing device 601, the above-mentioned functions defined in the method of the embodiment of the present application are executed.
[0113] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.
[0114] In some embodiments, the client and the server may communicate using any currently known or future developed network protocol such as HTTP (HyperText Transfer Protocol), and may be interconnected with any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include a local area network ("LAN"), a wide area network ("WAN"), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or future developed network.
[0115] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.
[0116] Computer program code for performing the operations of the present application may be written in one or more programming languages or a combination thereof, including, but not limited to, object-oriented programming languages, such as Java, Smalltalk, C++, and conventional procedural programming languages, such as "C" or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0117] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.
[0118] The modules involved in the embodiments described in this application can be implemented by software or hardware. The name of the module does not limit the module itself in some cases. For example, the test parameter acquisition module can also be described as a "module for acquiring device test parameters corresponding to the target device".
[0119] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.
[0120] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0121] Figure 5 FIG. 1 is an application scenario diagram of an intelligent early warning system based on simulation reasoning according to an exemplary embodiment. Figure 5 As shown, in this application scenario, first, the time queue of rainfall is obtained (for example, Figure 5 Then, the obtained rainfall time queue is input to a server (for example, Figure 5 In S) shown in , the server is capable of processing the rainfall time queue based on an intelligent early warning algorithm based on simulation reasoning to obtain a cumulative precipitation forecast value, and based on the cumulative precipitation forecast value, determines whether to generate an early warning prompt.
[0122] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.
[0123] In addition, although each operation is described in a specific order, this should not be construed as requiring these operations to be performed in the specific order shown or to be performed in a sequential order. Under certain circumstances, multitasking and parallel processing may be advantageous. Similarly, although some specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the application. Some features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination mode.
[0124] Although the subject matter has been described in language specific to structural features and / or method logic actions, it should be understood that the subject matter defined is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely exemplary forms of implementation. With respect to the apparatus in the above-mentioned embodiments, the specific manner in which each module performs the operation has been described in detail in the embodiments related to the method, and will not be described in detail here.
Claims
1. An intelligent early warning system based on simulation reasoning, characterized in that: include: The rainfall time series acquisition module is used to obtain the time queue of rainfall; A rainfall time series correlation feature extraction module is used to input the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series correlation feature vectors; A rainfall time series correlation feature maximum value extraction module is used to extract the maximum value of each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors to obtain a rainfall time series significant information representation vector; A rainfall time series significant information activation processing module is used to activate the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector; A rainfall time series correlation significant optimization module is used to weight each rainfall time series correlation feature vector in the sequence of rainfall time series correlation feature vectors by taking the feature value of each position in the rainfall time series significant guide weight vector as a weight to obtain a sequence of significantly optimized rainfall time series correlation feature vectors; A rainfall time series propagation aggregation module is used to input the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on the significant time series attenuation of node energy to obtain a rainfall time series propagation aggregation representation vector as a rainfall time series propagation aggregation representation feature; The cumulative precipitation prediction module is used to obtain a cumulative precipitation prediction value based on the rainfall time series propagation aggregation representation characteristics, and determine whether to generate an early warning prompt based on the cumulative precipitation prediction value.
2. The intelligent early warning system based on simulation reasoning according to claim 1 is characterized in that: The rainfall time series correlation feature extraction module is used to: input the rainfall time queue into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series correlation feature vectors.
3. The intelligent early warning system based on simulation reasoning according to claim 2 is characterized in that: The rainfall time series significant information activation processing module is used to: input the rainfall time series significant information representation vector into a Sigmoid activation function to obtain the rainfall time series significant guidance weight vector.
4. The intelligent early warning system based on simulation reasoning according to claim 3 is characterized in that: The rainfall time series propagation aggregation module includes: A significantly optimized rainfall energy significant descriptor calculation unit is used to calculate the energy significant descriptor of each significantly optimized rainfall time series associated feature vector in the sequence of significantly optimized rainfall time series associated feature vectors based on the mean and standard deviation of each significantly optimized rainfall time series associated feature vector to obtain a sequence of significantly optimized rainfall time series energy significant descriptors; A timestamp extraction unit, used for extracting the timestamp of each significantly optimized rainfall time series associated feature vector in the sequence of significantly optimized rainfall time series associated feature vectors to obtain a sequence of significantly optimized rainfall timestamps; A significantly optimized rainfall feature vector acquisition unit is used to use the significantly optimized rainfall time series association feature vector corresponding to the current time point in the sequence of the significantly optimized rainfall time series association feature vector as the current node feature vector and use the significantly optimized rainfall time series association feature vectors corresponding to other time points as the historical node feature vectors to obtain a sequence of the current significantly optimized rainfall time series association feature vector and the historical significantly optimized rainfall time series association feature vector; A significantly optimized rainfall energy time decay factor calculation unit is used to calculate the energy time decay factor of each historical significantly optimized rainfall time series associated characteristic vector in the sequence of the historical significantly optimized rainfall time series associated characteristic vectors relative to the current significantly optimized rainfall time series associated characteristic vector to obtain a sequence of significantly optimized rainfall energy time decay factors; A historically significant optimized rainfall time series propagation unit is used to perform time series propagation aggregation on the sequence of historically significant optimized rainfall time series associated feature vectors using the sequence of the significantly optimized rainfall time series energy significant descriptors as positive adjustment factors and the sequence of the significantly optimized rainfall energy time decay factors as negative adjustment factors to obtain a historically significant optimized rainfall time series propagation aggregation representation feature vector; A vector fusion unit is used to fuse the historical significantly optimized rainfall time series propagation aggregation representation feature vector and the current significantly optimized rainfall time series association feature vector to obtain the rainfall time series propagation aggregation representation vector.
5. The intelligent early warning system based on simulation reasoning according to claim 4 is characterized in that: The significantly optimized rainfall energy significant descriptor calculation unit is used to: Respectively calculating the mean and standard deviation of the significantly optimized rainfall time series correlation feature vector to obtain the significantly optimized rainfall time series correlation feature mean and the significantly optimized rainfall time series correlation feature standard deviation; Calculate the positional difference between the significantly optimized rainfall time series correlation feature vector and the significantly optimized rainfall time series correlation feature mean to obtain a significantly optimized rainfall time series correlation difference vector; Calculating the fourth power of each eigenvalue in the significantly optimized rainfall time series correlation difference vector to obtain a significantly optimized rainfall time series correlation modulation difference vector; Calculating the expected value of the significantly optimized rainfall time series associated modulation difference vector to obtain the significantly optimized rainfall time series associated expected value; The significantly optimized rainfall time series associated expected value is divided by the fourth power of the significantly optimized rainfall time series associated characteristic standard deviation to obtain the significantly optimized rainfall time series energy significant descriptor.
6. The intelligent early warning system based on simulation reasoning according to claim 5 is characterized in that: The significantly optimized rainfall energy time attenuation factor calculation unit is used to: The significantly optimized rainfall timestamp corresponding to the current significantly optimized rainfall time series associated feature vector is subtracted from the significantly optimized rainfall timestamp corresponding to each of the historical significantly optimized rainfall time series associated feature vectors, and then rounded down to obtain a sequence of historical significantly optimized rainfall time span values; After calculating the square of the positional division of the sequence of the historically significant optimized rainfall time span values and the time attenuation inverse scaling parameter to obtain a sequence of the historically significant optimized rainfall time span square modulation values, the sequence of the historically significant optimized rainfall time span square modulation values and the attenuation rate positive scaling parameter are positionally multiplied to obtain a sequence of the energy attenuation coefficients of the significantly optimized rainfall time series span; Taking each significantly optimized rainfall time series span energy attenuation coefficient in the sequence of significantly optimized rainfall time series span energy attenuation coefficients as an exponential power, the natural exponential function value with the natural constant e as the base is calculated to obtain the sequence of significantly optimized rainfall energy time attenuation factors.
7. The intelligent early warning system based on simulation reasoning according to claim 6 is characterized in that: The historically significant optimized rainfall time series propagation unit is used to: divide the sequence of the significantly optimized rainfall time series energy significant descriptors by position and the sequence of the significantly optimized rainfall energy time attenuation factors by position to obtain a sequence of significantly optimized rainfall adjustment factors, and calculate the positionally weighted sum of the sequence of the significantly optimized rainfall adjustment factors and the sequence of the historically significant optimized rainfall time series associated feature vectors to obtain the historically significant optimized rainfall time series propagation aggregation representation feature vector.
8. The intelligent early warning system based on simulation reasoning according to claim 7 is characterized in that: The cumulative precipitation prediction module comprises: A decoding unit, used for inputting the rainfall time series propagation aggregation representation vector into a decoder-based cumulative precipitation predictor to obtain a cumulative precipitation prediction decoded value; The early warning prompt determination unit is used to determine whether to generate the early warning prompt based on the comparison between the cumulative precipitation prediction decoded value and a preset threshold.
9. An intelligent early warning method based on simulation reasoning, characterized in that: include: Get the time queue of rainfall; Inputting the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors; Extracting the maximum value of each rainfall time series associated feature vector in the sequence of rainfall time series associated feature vectors to obtain a rainfall time series significant information representation vector; Activating the rainfall time series significant information representation vector to obtain a rainfall time series significant guidance weight vector; Using the characteristic values of each position in the rainfall time series significant guide weight vector as weight, weighting each rainfall time series associated characteristic vector in the sequence of rainfall time series associated characteristic vectors to obtain a sequence of significantly optimized rainfall time series associated characteristic vectors; Inputting the sequence of the significantly optimized rainfall time series associated feature vectors into a feature forward propagation network based on the significant time series attenuation of node energy to obtain a rainfall time series propagation aggregate representation vector as a rainfall time series propagation aggregate representation feature; Based on the rainfall time series propagation aggregation representation characteristics, a cumulative precipitation forecast value is obtained, and based on the cumulative precipitation forecast value, it is determined whether to generate an early warning prompt.
10. The intelligent early warning method based on simulation reasoning according to claim 9 is characterized in that: Inputting the rainfall time queue into a rainfall sequence encoder to obtain a sequence of rainfall time series associated feature vectors, including: inputting the rainfall time queue into a rainfall sequence encoder based on a 1D-CNN model to obtain a sequence of rainfall time series associated feature vectors.
