A method, equipment, medium, and product for optimizing flood disaster early warning decision-making.

Abstract

This application discloses a method, device, medium, and product for optimizing flood disaster early warning decision-making, relating to the field of environmental disaster management and prediction. The method includes: determining a decision feature matrix based on a multi-dimensional decision-making context; determining deterministic values ​​based on the decision feature matrix and dynamically calibrating psychological parameters; improving the Gini index based on the weights of the decision feature matrix and the probability of flood events, and constructing a split decision tree model based on the improved Gini index; determining key split points on the probability of flood events based on the trained split decision tree model; determining an adaptive decision threshold based on the probability of flood events and the key split points; determining a framing effect compensation factor; adjusting the adaptive decision threshold using the framing effect compensation factor; and determining the flood disaster early warning decision based on the adjusted threshold. This application can improve the accuracy and robustness of flood early warning decision-making.

Description

Technical Field

[0001] This application relates to the field of environmental disaster management and prediction, and in particular to a method, equipment, medium and product for optimizing flood disaster early warning decision-making. Background Technology

[0002] Flood early warning decision technology is a key link in the modern flood control and disaster reduction system. It achieves accurate prediction, timely early warning and scientific decision-making of floods by comprehensively using a variety of technical means. The existing flood early warning decision technology mainly relies on the following two types of methods: (1) Traditional economic decision-making models: such as the cost-loss ratio model, which determines the decision threshold by calculating the ratio of early warning cost to potential loss. However, this method has significant defects: such as the lack of psychological factors, and the failure to quantify the irrational behavioral characteristics of decision-makers in risk probability perception (such as the distortion of probability weight function) and loss aversion tendency; the limitations of static thresholds, which use fixed threshold parameters and cannot reflect the change in risk attitude when the probability of flood changes dynamically. (2) Single model application: commonly used models include behavioral models such as prospect theory, which can characterize the risk preferences of decision-makers, but are not coupled with flood probability prediction models, resulting in a lack of real-time hydrological data support for decision results; machine learning classification models represented by decision tree models, although they can handle classification decision problems, do not introduce psychological parameters and are not sensitive enough to decision bias caused by framing effects.

[0003] It can be seen that the existing technical problems are mainly concentrated in the following four aspects: (1) Lack of psychological behavior quantification: Traditional models cannot quantify the dynamic changes in decision-makers' risk attitudes in different flood probability ranges (such as risk aversion-seeking reversal in the 25%-50% probability range), resulting in excessive deviation in the setting of warning thresholds; (2) Insufficient model synergy: In the existing technology, the models run independently, causing the connection between psychological parameters and hydrological prediction data to break, resulting in cross-model data transmission errors; (3) Defects in dynamic adaptability: Static threshold methods (such as fixing r=0.5) are difficult to adapt to real-time changes in flood probability, and the prediction accuracy decreases in cross-scenario testing; (4) Framing effect interference: Single models do not have a built-in framing effect correction mechanism. When the same decision problem is expressed in the form of "loss" or "gain", the probability threshold for triggering risk attitude change is very different. Summary of the Invention

[0004] The purpose of this application is to provide a method, equipment, medium, and product for optimizing flood disaster early warning decision-making, which can improve the accuracy and robustness of flood early warning decisions, while reducing the risk of misallocation of emergency resources due to decision-making bias.

[0005] To achieve the above objectives, this application provides the following solution:

[0006] Firstly, this application provides a flood disaster early warning decision optimization method, which includes:

[0007] Based on a flood warning decision-making experiment, a multi-dimensional decision-making scenario combining probability information and cost-benefit analysis is constructed. A decision feature matrix is ​​then determined based on this multi-dimensional decision-making scenario. The multi-dimensional decision-making scenario uses rainfall forecasts deeply integrated with the probability of flood events, the actual losses caused by floods, and the cost of issuing flood warnings as core parameters. The decision feature matrix includes core parameters and binary decision labels under different decision-making scenarios.

[0008] Based on the decision feature matrix, the certainty equivalent value is determined, and psychological parameters are dynamically calibrated to obtain the weight of the probability of flood events occurring; the certainty equivalent value represents the quantified risk attitude value of decision-makers when making different decisions in the face of uncertain disaster forecasts;

[0009] The Gini index is improved based on the decision feature matrix and the weights of the probability of flood events, and a split decision tree model is constructed based on the improved Gini index.

[0010] The key split points in the probability of flood events are determined based on the trained split decision tree model; these key split points are used to reflect the changes in the behavior of decision-makers in different probability intervals.

[0011] Determine the adaptive decision threshold based on the probability of flood events and the corresponding critical split points;

[0012] Based on the flood recap report, semantic feature vectors were determined using a word vector model; and framing effect compensation factors were determined based on the semantic feature vectors.

[0013] The adaptive decision threshold is adjusted using a framing effect compensation factor to obtain the adjusted threshold.

[0014] Flood disaster early warning decisions are made based on the adjusted thresholds.

[0015] Optionally, the step of determining deterministic equivalents based on the decision feature matrix and dynamically calibrating psychological parameters to obtain the weights of the probability of flood events occurring specifically includes:

[0016] Using formula Determine the value of certainty ;in, The weighting function is the probability of a flood event occurring. P represents the probability of a flood event, and δ∈(0.28, 1] is the probability perception distortion coefficient. For the cumulative prospect theory value function, , Considering the economic utility of the consequences of decision-making, and the actual losses incurred by a flood or the cost of issuing a flood warning, ∈(0,1) represents the risk attitude coefficient. This is the loss aversion coefficient;

[0017] The weights of the probability of flood events are obtained by fitting the deterministic value of the optimization algorithm.

[0018] Optionally, the step of improving the Gini index based on the decision feature matrix and the weights of the probability of flood events, and constructing a split decision tree model based on the improved Gini index, specifically includes:

[0019] Using formula Determine the improved Gini index ;

[0020] in, Let be the weight of the probability of a flood event occurring at the i-th node. For the category of split points, Let be the binary decision label for the flood event at the i-th node. These are tree nodes in a split decision tree model.

[0021] Optionally, determining the adaptive decision threshold based on the probability of flood events and the corresponding critical split points specifically includes:

[0022] Using formula Construct a risk attitude-probability mapping matrix;

[0023] in, For adaptive decision threshold, and As a key split point, This represents the probability of a flood event occurring.

[0024] Optionally, the step of determining semantic feature vectors based on a word vector model according to the flood recap report, and determining a framing effect compensation factor based on the semantic feature vectors, specifically includes:

[0025] Using formula Determine the framing effect compensation factor ;

[0026] in, and It is a semantic feature vector, including the keyword density of flood loss and the keyword density of flood gain.

[0027] Optionally, adjusting the adaptive decision threshold using the framing effect compensation factor to obtain the adjusted threshold specifically includes:

[0028] Using formula Determine the adjusted threshold ;

[0029] in, For adaptive decision threshold, This is a framing effect compensation factor.

[0030] Optionally, the step of determining flood disaster early warning decisions based on the adjusted threshold specifically includes:

[0031] The real-time cost-loss ratio is determined based on the actual losses caused by the flood and the cost of issuing flood warnings.

[0032] Determine if the real-time cost loss ratio is less than or equal to the adjusted threshold; if yes, issue an alert; otherwise, do not issue an alert.

[0033] Based on the judgment results, output decision instructions with confidence ratings and risk attitude evolution curves.

[0034] Secondly, this application provides a computer device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the flood disaster early warning decision optimization method.

[0035] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned flood disaster early warning decision optimization method.

[0036] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned flood disaster early warning decision optimization method.

[0037] According to the specific embodiments provided in this application, this application has the following technical effects:

[0038] This application provides a method, equipment, medium, and product for optimizing flood disaster early warning decision-making. It constructs a multi-dimensional decision-making scenario combining probabilistic information and cost-benefit analysis; determines a decision feature matrix based on the multi-dimensional scenario; determines deterministic and other values ​​based on the decision feature matrix; and dynamically calibrates psychological parameters to obtain the weights of the probability of flood events occurring. This improves the Gini index. This application implements a coupling algorithm between psychological parameters and the splitting criterion (Gini index), overcoming the dimensionality curse of traditional statistical methods in heterogeneous data processing. It also establishes a joint calibration method for probability weight functions and value functions to solve the multi-scale matching problem between psychological parameters and hydrological data. Based on the probability of flood events and the corresponding key splitting points, an adaptive decision threshold is determined, i.e., a two-dimensional mapping matrix of risk attitude and flood event probability is constructed to achieve adaptive adjustment of the threshold under different early warning levels. This application utilizes a framing effect compensation factor to adjust the adaptive decision-making threshold, i.e., to set up a framing effect correction mechanism to reduce decision-making bias caused by framing effects. It combines human behavior simulation from psychology with machine learning techniques from computer science to quantify decision-makers' risk preferences and judgment biases in uncertain environments, thereby optimizing the automated decision-making process in flood warning systems. This application is mainly applied to early warning and risk analysis of hydrological and meteorological disasters, especially for efficient assessment of flood risks in dynamic environments, improving the accuracy and reliability of disaster response decisions, while reducing the risk of misallocation of emergency resources due to decision-making biases. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a flowchart illustrating a flood disaster early warning decision optimization method according to an embodiment of this application;

[0041] Figure 2 This is a schematic diagram illustrating the principle of a flood disaster early warning decision optimization method in one embodiment of this application. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] In one exemplary embodiment, such as Figure 1 and Figure 2 As shown, a flood disaster early warning decision optimization method is provided, which includes the following steps S101 to S108. Wherein:

[0045] S101, Based on the flood warning decision-making experiment, a multi-dimensional decision-making scenario combining probability information and cost-benefit analysis is constructed; and a decision feature matrix is ​​determined based on the multi-dimensional decision-making scenario; the multi-dimensional decision-making scenario uses the probability of flood event occurrence P∈[0,1] of deep fusion of rainfall forecasts, the actual loss L caused by flood occurrence (including direct and indirect losses), and the cost C of issuing a flood warning (including the cost of issuing a warning and the cost of incorrect decisions) as core parameters; the decision feature matrix includes core parameters and binary decision labels under different decision-making scenarios; wherein, the binary decision label Y∈{0,1}, 1 indicates that among the 168 decision-makers participating in the experiment, a warning was issued to the public, and 0 indicates that among the 168 decision-makers participating in the experiment, a warning was not issued to the public;

[0046] The probability of a flood event, P, is dynamically discretized into four risk sub-intervals (low probability of flood (0%-25%), medium probability of flood (25%-50%), high probability of flood (50%-75%), and high probability of flood (75%-100%)), and the cost-loss ratio C / L is calculated.

[0047] The output decision feature matrix X=[P, C / L, Y] serves as the basis for subsequent psychological parameter estimation and model training.

[0048] S102, Based on the decision feature matrix, determine the certainty equivalent value and perform dynamic calibration of psychological parameters to obtain the weight of the probability of flood event occurrence; the certainty equivalent value represents the quantified risk attitude value of decision-makers when making different decisions in the face of uncertain disaster forecasts;

[0049] The purpose of S102 is to quantify decision-makers' risk preferences and behavioral biases based on historical data (decision characteristic matrix); S102 specifically includes:

[0050] (1) Given a fixed probability p of a flood event, provide an explanation of whether the decision-maker should "issue an alert" or "not issue an alert," along with the costs and losses associated with each decision. Then, ask the decision-maker to choose which of the following:

[0051] 1. Accept a fixed payment of M amount to avoid this decision;

[0052] 2. Or bear the risk of costs or losses arising from the decision-making process;

[0053] 3. Adjust M according to the decision-maker's choice until they have no difference between a certain amount and the corresponding risk. This certain amount M is the certainty equivalent value CE of the risk prospect when the probability p of the flood event occurs.

[0054] (2) Calling the cumulative prospect theory value function:

[0055] ;

[0056] (3) Update the weighting function for the probability of flood events:

[0057] ;

[0058] Where δ∈(0.28,1] is the probability-perceived distortion coefficient. Let be the weighting function for the probability of a flood event, and δ∈(0.28, 1] be the probability-perceived distortion coefficient. For the cumulative prospect theory value function, Considering the economic utility of the consequences of decision-making, and the actual losses incurred by a flood or the cost of issuing a flood warning, ∈(0,1) represents the risk attitude coefficient. This is the loss aversion coefficient;

[0059] (4) Calculate the definite equivalent value Certainty and other values This represents a quantitative value of the risk attitude of decision-makers when facing uncertain disaster forecasts and taking different decisions. It is estimated by fitting the label Y using an optimization algorithm (i.e., gradient descent). The value of makes determinism equivalent. and actual data collected through experiments The closest.

[0060] ;

[0061] S103, improve the Gini index based on the decision feature matrix and the weights of the probability of flood events, and construct a split decision tree model based on the improved Gini index; the split decision tree model can reflect the psychological bias of decision-makers;

[0062] S103 specifically includes:

[0063] (1) Train a split decision tree model using weighted samples to predict the decision behavior label Y∈{0,1}. The probability weights of the weighted samples are used to... Adjusting the weighting of different samples in the split decision tree model allows the splitting process to reflect the decision-maker's psychological biases.

[0064] (2) When evaluating the split point, calculate the improved Gini index (the formula for the Gini index in the traditional splitting model is...). ),in Category of split points The proportion used to measure the impurity of nodes).

[0065] ;

[0066] in, Let be the weight of the probability of a flood event occurring at the i-th node. For the category of split points, Let be the binary decision label for the flood event at the i-th node. These are tree nodes in a split decision tree model.

[0067] S104, Based on the trained split decision tree model, determine the key split points in the probability of flood events occurring; the key split points are used to reflect the changes in the behavior of decision-makers in different probability intervals.

[0068] Using a trained split decision tree model, key split points on the probability P of a flood event are identified. These key split points serve as a threshold r, reflecting changes in decision-makers' behavior across different probability intervals (e.g., low risk or high risk).

[0069] S105, based on the probability of flood events and the corresponding key split points, determines the adaptive decision threshold, realizes the dynamic transition of the threshold in the 25%-50% probability range, and eliminates the static threshold deviation;

[0070] S105 specifically includes:

[0071] Using formula Construct a risk attitude-probability mapping matrix;

[0072] in, For adaptive decision threshold, and As a key split point, This represents the probability of a flood event occurring.

[0073] As a specific example, =0.81, which is the low-risk threshold, indicating that decision-makers are highly sensitive to whether to issue a flood warning in low-risk situations. =0.53, which is the high-risk threshold, indicates that decision-makers exhibit greater risk aversion in high-risk situations and are more inclined to take proactive preventative measures to minimize potential catastrophic losses.

[0074] S106. Based on the flood recap report, semantic feature vectors are determined using a word vector model; and a framing effect compensation factor is determined based on the semantic feature vectors; the decision bias caused by framing effect is reduced to within ±5% using the framing effect compensation factor.

[0075] As a specific implementation, based on the text describing the decision problem, semantic feature vectors S∈ℝ are extracted using a word vector model. 6 (Including keyword density for flood loss and keyword density for flood benefit (early warning effect))

[0076] Using formula Determine the framing effect compensation factor ;

[0077] in, and It is a semantic feature vector, including the keyword density of flood loss and the keyword density of flood gain.

[0078] S107, the adaptive decision threshold is adjusted using the framing effect compensation factor to obtain the adjusted threshold;

[0079] Using formula Determine the adjusted threshold ;

[0080] S108, determine flood disaster early warning decisions based on the adjusted threshold.

[0081] S108 specifically includes:

[0082] S1, determine the real-time cost-loss ratio based on the actual losses caused by the flood and the cost of issuing flood warnings. ;

[0083] S2, determine whether the real-time cost loss ratio is less than or equal to the adjusted threshold; if yes, issue an alert; if no, do not issue an alert.

[0084] ;

[0085] S3, based on the judgment result It outputs decision instructions with confidence ratings and risk attitude evolution curves, thereby demonstrating the impact of flood forecast probability, psychological interpretation, and framing effect on decision-making.

[0086] Compared with the prior art, this application has the following advantages:

[0087] (1) Improved accuracy of dynamic quantification of risk attitude: Psychological parameters ( The coupled design of risk neutrality and splitting criteria can accurately capture the risk attitude reversal phenomenon in the 25%-50% probability range of flood events. Experimental data shows that the decision accuracy in this range has increased from 68% in the traditional model to 82%. =0.81) and risk aversion threshold ( The dynamic transition mechanism of Gini index (=0.53) reduces the cross-scene prediction error to 8.2% (the error of the traditional static threshold model is 21.5%); (2) Breakthrough in model synergy: using the improved Gini index ( ) and probability weight function ( The embedded fusion of psychological parameters and hydrological data enables real-time interaction between psychological parameters and hydrological data, reducing the data transfer error between models from 12.7% to 3.4%; and in the Ramos case test, the area under the ROC curve (AUC) of the joint model reached 0.89, which is significantly better than the single CPT model (AUC=0.76) or DT model (AUC=0.81); (3) Enhanced frame effect anti-interference ability: based on the semantic parsing compensation factor β and the dynamic threshold function ( The product correction of ) eliminates the decision bias caused by different problem statements, and reduces the trigger difference between the loss frame (40% threshold) and the gain frame (75% threshold) to within ±5%; and in the test set containing 145 groups of semantic interference, the false alarm rate decreased from 22.3% to 9.1%; (4) Real-time decision efficiency optimization: according to the segmented threshold function ( The linear computational characteristics and pre-trained semantic vectors of the model can reduce the single decision computation time from 58ms in the traditional CPT model to 12ms, meeting the real-time response requirements of flood warning (<50ms); and the memory usage is reduced by 63% (from 2.1GB to 0.78GB), adapting to the deployment requirements of edge computing devices.

[0088] The following specific embodiments illustrate the beneficial effects of this application:

[0089] Example 1: Core Parameter Calibration Experiment

[0090] The flood decision-making simulation experiment (including 5 probabilistic scenarios × 3 loss frames) involved 168 emergency management personnel. The specific implementation steps are as follows:

[0091] (1) Behavioral data collection

[0092] The decision simulation interface was used to collect the early warning selection records of the subjects under different probabilities of flood event occurrence (P=10%, 25%, 40%, 60%, 80%) (a total of 2520 decision data); the response time, economic parameters (C / L∈[0.2, 1.8]) and problem statement framework (loss / gain) of each decision were recorded.

[0093] (2) Psychological parameter fitting

[0094] The maximum likelihood estimation method was used to solve for the parameters of the value function, and the optimal parameters were obtained: α=0.88±0.03, γ=0.65±0.05, λ=2.25 (95% confidence interval).

[0095] (3) Training of split decision tree model

[0096] Input feature matrix X=[P, C / L, frame type], label Y=actual decision; use improved Gini index (embedding w(p)) to split nodes and generate a tree structure with a depth of 5; obtain key threshold r1=0.81 (splitting criterion when P<25%).

[0097] (4) Verification results

[0098] 4.1 Risk attitude reversal interval test: In the range of P=25%-50%, the risk aversion rate of decision-makers dropped sharply from 82% to 37% (χ²=29.7, p<0.001).

[0099] 4.2 Model fit: McFadden's R² = 0.63, which is significantly better than the traditional logit model (R² = 0.41).

[0100] Example 2: Cross-Scenario Validation (Ramos Case)

[0101] The specific implementation steps for connecting to the real-time hydrological monitoring network of the Tagus River basin in Spain and acquiring historical flood event records (145 early warning decision data from 2010 to 2020) are as follows:

[0102] (1) Data migration and adaptation

[0103] The original cost data (C ∈ [2.3, 5.1] million euros) and loss data (L ∈ [8.7, 32.4] million euros) were converted using exchange rates and adjusted for purchasing power parity; a cross-cultural framing effect lexicon was established: the Spanish expression "dañospotenciales" (potential loss) was mapped to... vector;

[0104] (2) Dynamic threshold generation

[0105] Calculate T(P) based on the real-time probability of a flood event, P(t):

[0106] When P = 35%, T(P) = 0.81 + (0.63 - 0.81) × (35 - 25) / 25 = 0.738;

[0107] The semantic parser detected that the warning text contained "evitar pérdidas" (avoid loss), triggering... =1.15; (3) Decision instruction output

[0108] Calculate the final threshold: T _final =0.738 × 1.15 = 0.849;

[0109] When C / L = 0.79 (C = 42,000, L = 53,000), 0.79 < 0.849 triggers an early warning.

[0110] (4) Verification results

[0111] Early warning accuracy: 76.4% (68.2% for traditional models), with accuracy particularly improved to 81.3% in the P=30%-45% range;

[0112] Response time: The average time from meteorological data input to decision output is 23ms, which meets the requirements for real-time early warning;

[0113] Example 3: Frame Effect Correction Verification

[0114] The same decision problem is described using both a loss framework ("potential loss of L yuan") and a benefit framework ("potential savings of L yuan"); the specific implementation process is as follows:

[0115] (1) The semantic parser detected the difference in keyword frequency.

[0116] Loss framework: "loss" appears 3.2 times per sentence, "disaster" appears 1.8 times per sentence;

[0117] Benefit framework: "Savings" appears 2.7 times per sentence, and "Safety" appears 2.1 times per sentence;

[0118] (2) Calculate the compensation factor

[0119] In practical applications, normalization is used to limit... ∈[0.9, 1.1]);

[0120] Adjusting the threshold: When P = 40%, the original threshold T(P) = 0.72, then... =0.72 × 1.1 = 0.792;

[0121] (3) Verification results

[0122] The C / L threshold for triggering an alert under the loss framework increased from 0.72 to 0.792, and the difference between it and the threshold under the gain framework decreased from 35% to 8.7%.

[0123] In double-blind testing, decision-makers' consistency in decision-making across different frameworks increased from 61% to 89%.

[0124] Based on the above embodiments, the present application has been verified in terms of parameter calibration accuracy (α error < ±0.05), real-time response capability (<50ms), and anti-interference performance. Technological advancements, such as those leading to a correction effectiveness rate of >87%, have provided a practical and intelligent solution for flood early warning decision-making.

[0125] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal. The computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is connected to the system bus via the I / O interfaces. The processor of the computer device provides computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The I / O interfaces of the computer device are used for exchanging information between the processor and external devices. The communication interface of the computer device is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a flood disaster early warning decision optimization method.

[0126] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0127] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0128] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above-described method embodiments.

[0129] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0130] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0131] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0132] In this application, all actions to acquire signals, information, or data are carried out in compliance with the relevant data protection laws and policies of the country where the location is situated, and with the authorization granted by the owner of the relevant device.

[0133] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0134] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for optimizing flood disaster early warning decision-making, characterized in that, The flood disaster early warning decision optimization method includes: Based on a flood warning decision-making experiment, a multi-dimensional decision-making scenario combining probability information and cost-benefit analysis is constructed. A decision feature matrix is ​​then determined based on this multi-dimensional decision-making scenario. The multi-dimensional decision-making scenario uses rainfall forecasts deeply integrated with the probability of flood events, the actual losses caused by floods, and the cost of issuing flood warnings as core parameters. The decision feature matrix includes core parameters and binary decision labels under different decision-making scenarios. Based on the decision feature matrix, the certainty equivalent value is determined, and psychological parameters are dynamically calibrated to obtain the weight of the probability of flood events occurring; the certainty equivalent value represents the quantified risk attitude value of decision-makers when making different decisions in the face of uncertain disaster forecasts; The Gini index is improved based on the decision feature matrix and the weights of the probability of flood events, and a split decision tree model is constructed based on the improved Gini index. The key split points in the probability of flood events are determined based on the trained split decision tree model; these key split points are used to reflect the changes in the behavior of decision-makers in different probability intervals. Determine the adaptive decision threshold based on the probability of flood events and the corresponding critical split points; Based on the flood recap report, semantic feature vectors were determined using a word vector model; and framing effect compensation factors were determined based on the semantic feature vectors. The adaptive decision threshold is adjusted using a framing effect compensation factor to obtain the adjusted threshold. Flood disaster early warning decisions are made based on the adjusted thresholds; The process of determining deterministic values ​​based on the decision feature matrix and dynamically calibrating psychological parameters to obtain the weights of the probability of flood events occurring specifically includes: Using formula Determining certainty and equivalent value ;in, The weighting function is the probability of a flood event occurring. P represents the probability of a flood event, and δ∈(0.28, 1] is the probability perception distortion coefficient. For the cumulative prospect theory value function, , Considering the economic utility of the consequences of decision-making, and the actual losses caused by flooding or the cost of issuing flood warnings, ∈(0,1) represents the risk attitude coefficient. This is the loss aversion coefficient; The weights of the probability of flood events are obtained by fitting the deterministic value of the optimization algorithm.

2. The flood disaster early warning decision optimization method according to claim 1, characterized in that, The improvement of the Gini index based on the decision feature matrix and the weights of the probability of flood events, and the construction of a split decision tree model based on the improved Gini index, specifically includes: Using formula Determine the improved Gini index ; in, Let be the weight of the probability of a flood event occurring at the i-th node. For the category of split points, Let be the binary decision label for the flood event at the i-th node. These are tree nodes in a split decision tree model.

3. The flood disaster early warning decision optimization method according to claim 1, characterized in that, The process of determining the adaptive decision threshold based on the probability of flood events and the corresponding critical split points specifically includes: Using formula Construct a risk attitude-probability mapping matrix; in, For adaptive decision threshold, and As a key split point, This represents the probability of a flood event occurring.

4. The flood disaster early warning decision optimization method according to claim 1, characterized in that, Based on the flood recap report, semantic feature vectors are determined using a word vector model. And the framing effect compensation factor is determined based on the semantic feature vector, specifically including: Using formula Determine the framing effect compensation factor ; in, and It is a semantic feature vector, including the keyword density of flood loss and the keyword density of flood gain.

5. The flood disaster early warning decision optimization method according to claim 1, characterized in that, The process of adjusting the adaptive decision threshold using a framing effect compensation factor to obtain the adjusted threshold specifically includes: Using formula Determine the adjusted threshold ; in, For adaptive decision threshold, This is a framing effect compensation factor.

6. The flood disaster early warning decision optimization method according to claim 1, characterized in that, The process of determining flood disaster early warning decisions based on the adjusted threshold specifically includes: The real-time cost-loss ratio is determined based on the actual losses caused by the flood and the cost of issuing flood warnings. Determine if the real-time cost loss ratio is less than or equal to the adjusted threshold; if yes, issue an alert; otherwise, do not issue an alert. Based on the judgment results, output decision instructions with confidence ratings and risk attitude evolution curves.

7. A computer device, comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that the processor executes the computer program to implement the flood disaster early warning decision optimization method according to any one of claims 1-6.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the flood disaster early warning decision optimization method as described in any one of claims 1-6.

9. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the flood disaster early warning decision optimization method as described in any one of claims 1-6.

Images