Urban safety monitoring quantitative prediction method and system based on big data analysis

Abstract

The invention discloses an urban safety monitoring quantitative prediction method and system based on big data analysis, and the method comprises the steps: determining a plurality of monitoring indexes of urban safety development, collecting the multi-source heterogeneous historical data of each monitoring index, and carrying out the preprocessing; performing data coupling and feature extraction on the preprocessed multi-source heterogeneous historical data, and selecting an analysis model according to data features and data industry features; carrying out training and cross validation on the analysis model based on the multi-source heterogeneous historical data of each monitoring index, and generating an urban safety monitoring quantitative prediction model; quantitative prediction is carried out based on multi-source heterogeneous real-time data of each monitoring index by using an urban safety monitoring quantitative prediction model, and different analysis models are selected to adapt to various data quality and data industry characteristics, so that the accuracy of quantitative prediction of the monitoring indexes is improved; the operation situation of the monitoring point can be effectively and quantitatively predicted, and early warning can be carried out based on prediction.

Description

Technical Field

[0001] The present invention relates to the technical field of urban safety monitoring, and in particular to a method and system for quantitative prediction of urban safety monitoring based on big data analysis. Background Art

[0002] With the accelerating pace of urbanization, the continuous expansion of cities, and the increasing complexity of urban systems, various safety hazards and risk factors are intertwined and overlapping. Traditional safety monitoring methods are no longer able to meet the needs of modern urban safety management. Currently, urban safety monitoring involves collecting, analyzing, and evaluating various safety-related data to monitor real-time risks in various areas of urban safety. This approach primarily uses real-time sensor data to determine alarms, analyze these alarms, and issue early warnings. Analyzing historical data of monitoring indicators allows for long-term forecasts of these values. However, the accuracy of these forecasts is significantly affected by data quality, scale, and diversity, making it difficult to achieve accurate quantitative forecasts of monitoring indicators and effectively warn of future risks. Summary of the Invention

[0003] In response to the problems shown above, the present invention provides a quantitative prediction method and system for urban safety monitoring based on big data analysis to solve the problem mentioned in the background technology that by analyzing the historical data of monitoring indicators, the monitoring values ​​can be predicted for a period of time, but the accuracy of the prediction is greatly affected by the data quality, data scale, and data diversity. It is usually difficult to achieve accurate quantitative prediction of monitoring indicators to effectively warn of risks that may arise in the future.

[0004] A quantitative prediction method for urban safety monitoring based on big data analysis, characterized by comprising the following steps:

[0005] Determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator and perform pre-processing;

[0006] Perform data coupling and feature extraction on pre-processed multi-source heterogeneous historical data, and select analysis models based on data characteristics and data industry characteristics;

[0007] The analysis model is trained and cross-validated based on multi-source heterogeneous historical data for each monitoring indicator to generate a quantitative prediction model for urban safety monitoring;

[0008] The quantitative prediction model of urban safety monitoring is used to make quantitative predictions based on multi-source heterogeneous real-time data of each monitoring indicator.

[0009] Preferably, the determining of multiple monitoring indicators for urban safety development, collecting multi-source heterogeneous historical data for each monitoring indicator and pre-processing the data, includes:

[0010] Identify multiple security dimensions of urban safety development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item;

[0011] Integrate all monitoring indicators for each security dimension, identify multiple data sources for each monitoring indicator, and determine the data privacy of each data source;

[0012] Determine data access permissions based on data privacy, and use data access permissions to collect multi-source heterogeneous historical data for each monitoring indicator;

[0013] Clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

[0014] Preferably, the data coupling and feature extraction of the pre-processed multi-source heterogeneous historical data, and the selection of an analysis model based on the data features and data industry characteristics, include:

[0015] Perform spatial and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network;

[0016] Extracting statistical features and domain-specific features corresponding to the coupled data, and determining prediction attributes based on the statistical features and domain-specific features;

[0017] Determine the evaluation logic based on the predicted attributes, and determine the model analysis objectives based on the evaluation logic and data industry characteristics;

[0018] Determine the model analysis modal mechanism according to the model analysis objectives, and select the analysis model based on the model analysis modal mechanism.

[0019] Preferably, the training and cross-validation of the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring includes:

[0020] Adjust the analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model;

[0021] Set the input and output layer parameters of the analysis model based on the main and auxiliary tasks, and train the analysis model using multi-source heterogeneous historical data for each monitoring indicator;

[0022] The trained analysis model is cross-validated using a stratified sampling method and a rolling time window, and the eligibility of the multi-dimensional validation indicators of the analysis model is determined based on the validation results;

[0023] Based on the eligibility of multi-dimensional verification indicators, the model is optimized to generate a quantitative prediction model for urban safety monitoring.

[0024] Preferably, the method of using the urban safety monitoring quantitative prediction model to perform quantitative prediction based on multi-source heterogeneous real-time data of each monitoring indicator includes:

[0025] Input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters;

[0026] Determine the reasonable state range of each monitoring indicator through the standard parameters of urban safety development, and determine the abnormal state based on the state change parameters and the reasonable state range of each monitoring indicator;

[0027] Determine the effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator based on the effect indicator parameters;

[0028] Determine safety risk responses based on the quantitative output of each monitoring indicator and issue early warnings based on safety risk responses.

[0029] A city safety monitoring quantitative prediction system based on big data analysis, the system comprising:

[0030] The acquisition module is used to determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator, and perform pre-processing;

[0031] The selection module is used to perform data coupling and feature extraction on pre-processed multi-source heterogeneous historical data, and select analysis models based on data features and data industry characteristics;

[0032] A generation module is used to train and cross-validate the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring;

[0033] The quantitative prediction module is used to make quantitative predictions based on multi-source heterogeneous real-time data of each monitoring indicator using the urban safety monitoring quantitative prediction model.

[0034] Preferably, the acquisition module includes:

[0035] The first determination submodule is used to determine multiple security dimensions of urban security development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item;

[0036] The second determination submodule is used to integrate all monitoring indicators of each security dimension, determine the multiple data sources of each monitoring indicator, and determine the data privacy of each data source;

[0037] The collection submodule is used to determine data retrieval permissions based on data privacy and collect multi-source heterogeneous historical data for each monitoring indicator through data retrieval permissions;

[0038] The preprocessing submodule is used to clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

[0039] Preferably, the selection module includes:

[0040] The processing submodule is used to perform spatial and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling processing on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network;

[0041] The third determination submodule is used to extract statistical features and domain-specific features corresponding to the coupled data, and determine the prediction attributes based on the statistical features and domain-specific features;

[0042] The fourth determination submodule is used to determine the evaluation logic according to the prediction attributes and determine the model analysis target based on the evaluation logic and data industry characteristics;

[0043] The selection submodule is used to determine the model analysis modal mechanism according to the model analysis target and select the analysis model based on the model analysis modal mechanism.

[0044] Preferably, the generating module includes:

[0045] Define submodules to adjust analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model;

[0046] The training submodule is used to set the input and output layer parameters of the analysis model based on the main task and auxiliary tasks, and train the analysis model with multi-source heterogeneous historical data of each monitoring indicator;

[0047] The validation submodule is used to cross-validate the trained analysis model through stratified sampling and rolling time windows, and determine the eligibility of the multi-dimensional validation indicators of the analysis model based on the validation results;

[0048] The generation submodule is used to optimize the model based on the eligibility of multi-dimensional verification indicators to generate a quantitative prediction model for urban safety monitoring.

[0049] Preferably, the quantitative prediction module includes:

[0050] The acquisition submodule is used to input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters;

[0051] The fifth determination submodule is used to determine the reasonable state range of each monitoring indicator through the urban safety development standard parameters, and determine the abnormal state according to the state change parameters and the reasonable state range of each monitoring indicator;

[0052] a sixth determination submodule, configured to determine effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator according to the effect indicator parameters;

[0053] The judgment submodule is used to judge the safety risk response based on the quantitative output of each monitoring indicator and to issue an early warning based on the safety risk response.

[0054] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures particularly pointed out in the written description and the accompanying drawings.

[0055] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0057] Figure 1 This is a workflow diagram of a quantitative prediction method for urban safety monitoring based on big data analysis provided by the present invention;

[0058] Figure 2 Another workflow diagram of the urban safety monitoring quantitative prediction method based on big data analysis provided by the present invention;

[0059] Figure 3 This is a schematic diagram of the structure of a city safety monitoring and quantitative prediction system based on big data analysis provided by the present invention;

[0060] Figure 4 This is a structural schematic diagram of a quantitative prediction module in a city safety monitoring quantitative prediction system based on big data analysis provided by the present invention. DETAILED DESCRIPTION

[0061] Exemplary embodiments will be described in detail herein, with examples illustrated in the accompanying drawings. In the following description, when referring to the drawings, identical numerals in different figures represent identical or similar elements, unless otherwise indicated. The embodiments described in the following exemplary embodiments are not intended to represent all possible embodiments consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present disclosure, as detailed in the appended claims.

[0062] At present, with the accelerated advancement of urbanization, the continuous expansion of urban scale, the increasing complexity of urban systems, and the interweaving and superposition of various safety hazards and risk factors, traditional safety monitoring methods can no longer meet the needs of modern urban safety management. At present, urban safety monitoring is carried out by collecting, analyzing and evaluating various types of urban safety-related data, monitoring the real-time risks in the operation of various fields of urban safety, and mainly judging the alarms based on the real-time monitoring values ​​of the sensors, and then analyzing the alarms and issuing early warnings. By analyzing the historical data of the monitoring indicators, the monitoring values ​​can be predicted for a period of time, but the accuracy of the prediction is greatly affected by the data quality, data scale, and data diversity. It is usually difficult to achieve accurate quantitative prediction of the monitoring indicators to effectively warn of risks that may arise in the future. In order to solve the above problems, this embodiment discloses a quantitative prediction method for urban safety monitoring based on big data analysis.

[0063] A quantitative prediction method for urban safety monitoring based on big data analysis, such as Figure 1 As shown, the following steps are included:

[0064] Step S101: Determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator, and perform pre-processing;

[0065] Step S102: perform data coupling and feature extraction on the pre-processed multi-source heterogeneous historical data, and select an analysis model based on the data features and data industry characteristics;

[0066] Step S103: training and cross-validating the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring;

[0067] Step S104: Utilize the urban safety monitoring quantitative prediction model to perform quantitative prediction based on multi-source heterogeneous real-time data of each monitoring indicator.

[0068] In this embodiment, urban safety development includes: public safety, infrastructure safety, health safety, social safety, and economic safety;

[0069] In this embodiment, the monitoring indicators are represented as mapping monitoring indicators of various safety items. For example, the mapping monitoring indicators of infrastructure safety include: the proportion of dangerous and old buildings, the safety rating of bridges and tunnels, the aging rate of underground pipelines, and the failure rate of public transportation facilities.

[0070] The working principle of the above technical solution is as follows: determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator and pre-process it; perform data coupling and feature extraction on the pre-processed multi-source heterogeneous historical data, and select an analysis model based on data characteristics and data industry characteristics; train and cross-validate the analysis model based on the multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring; use the quantitative prediction model for urban safety monitoring to perform quantitative prediction based on the multi-source heterogeneous real-time data of each monitoring indicator.

[0071] The beneficial effects of the above technical solution are: by selecting different analysis models to adapt to various data quality and data industry characteristics, the accuracy of quantitative prediction of monitoring indicators is improved, and the operating status of the monitoring points can be effectively quantitatively predicted and early warning can be given based on the prediction. It solves the problem mentioned in the existing technology that by analyzing the historical data of the monitoring indicators, the monitoring values ​​can be predicted for a period of time, but the accuracy of the prediction is greatly affected by the data quality, data scale, and data diversity. It is usually difficult to achieve accurate quantitative prediction of monitoring indicators to effectively warn of possible risks in the future.

[0072] In this embodiment, after determining multiple monitoring indicators for urban safety development, the following are also included:

[0073] Obtain safety processing related information for each monitoring indicator, and determine the identification parameters and early warning parameters for each monitoring indicator based on the safety processing related information;

[0074] Determine the dynamic strategy for quantitative safety assessment of each monitoring indicator based on the judgment parameters and early warning parameters of each monitoring indicator;

[0075] Adjust the monitoring data weight of each monitoring indicator based on the dynamic strategy of quantitative security assessment, and build a security resilience evaluation system for each monitoring indicator based on the adjustment results;

[0076] Determine the dynamic change parameters of the first reference indicator within the safe and controllable range and the dynamic change parameters of the second reference indicator outside the safe and controllable range based on the safety resilience evaluation system of each monitoring indicator;

[0077] Determine the basic control equation of each monitoring indicator according to the dynamic change parameter of the first reference indicator and the dynamic change parameter of the second reference indicator;

[0078] Determine the safety evaluation model for each monitoring indicator based on the basic control equation, and evaluate the multiple data sources for each monitoring indicator according to the safety evaluation model;

[0079] Determine risk data sources and non-risk data sources based on the evaluation results, obtain statistical historical data of risk data sources, and classify the statistical historical data using a quantitative grading model to obtain a hierarchical structure of the statistical historical data;

[0080] Determine the potential safety hazard factors of each risk data source based on the hierarchical structure, and determine the random probability of the key assessment of each risk data source based on the potential safety hazard factors;

[0081] Determine the necessity of data collection for each risk data source based on the random probability of key assessment of each risk data source;

[0082] Based on the necessity of data collection for each risk data source, the data retrieval source for each monitoring indicator is determined, and the data retrieval source is used as a reference sample to collect multi-source heterogeneous historical data for each monitoring indicator.

[0083] The beneficial effects of the above technical solution are: by conducting in-depth security assessment principle mining for each monitoring indicator and then screening out risk data sources, it is possible to accurately screen out data sources with high security risks based on the security threat nature of each monitoring indicator and then conduct targeted data analysis on them, thereby avoiding omissions in the investigation of high security risk data sources and improving practicality. Furthermore, by evaluating the necessity of data collection for each risk data source, it is possible to more intuitively determine the necessary data sources for collection, further ensuring the high quality and high price system of the data.

[0084] In one embodiment, determining multiple monitoring indicators for urban safety development, collecting multi-source heterogeneous historical data for each monitoring indicator and pre-processing the data may include:

[0085] Identify multiple security dimensions of urban safety development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item;

[0086] Integrate all monitoring indicators for each security dimension, identify multiple data sources for each monitoring indicator, and determine the data privacy of each data source;

[0087] Determine data access permissions based on data privacy, and use data access permissions to collect multi-source heterogeneous historical data for each monitoring indicator;

[0088] Clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

[0089] The beneficial effects of the above technical solution are: by determining data retrieval permissions and collecting multi-source heterogeneous historical data of each monitoring indicator through data retrieval permissions, the integrity of data traceability can be guaranteed. Furthermore, by performing data preprocessing, the data integrity and high precision can be guaranteed, laying the foundation for subsequent model training and improving stability and practicality.

[0090] In one embodiment, Figure 2 As shown, the pre-processed multi-source heterogeneous historical data is coupled and features are extracted, and an analysis model is selected based on the data features and data industry characteristics, including:

[0091] Step S201: perform spatial alignment and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling processing on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network;

[0092] Step S202: extracting statistical features and domain-specific features corresponding to the coupled data, and determining prediction attributes based on the statistical features and domain-specific features;

[0093] Step S203: Determine the evaluation logic based on the predicted attributes, and determine the model analysis target based on the evaluation logic and data industry characteristics;

[0094] Step S204: determining a model analysis modal mechanism according to a model analysis objective, and selecting an analysis model based on the model analysis modal mechanism.

[0095] The beneficial effect of the above technical solution is that by determining the prediction attributes and then determining the model analysis modal mechanism, the appropriate analysis model can be accurately selected based on the industry characteristics and data features of the data, ensuring the accuracy and precision of the prediction results.

[0096] In one embodiment, the training and cross-validation of the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring includes:

[0097] Adjust the analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model;

[0098] Set the input and output layer parameters of the analysis model based on the main and auxiliary tasks, and train the analysis model using multi-source heterogeneous historical data for each monitoring indicator;

[0099] The trained analysis model is cross-validated using a stratified sampling method and a rolling time window, and the eligibility of the multi-dimensional validation indicators of the analysis model is determined based on the validation results;

[0100] Based on the eligibility of multi-dimensional verification indicators, the model is optimized to generate a quantitative prediction model for urban safety monitoring.

[0101] The beneficial effects of the above technical solution are: by setting the input layer and output layer parameters of the analysis model for model training, the model can be quickly trained and converged according to the training data, thereby improving the training efficiency and success. Furthermore, by evaluating the model performance and qualification based on the qualification of multi-dimensional verification indicators, the trained model can be comprehensively qualified and optimized, thereby ensuring the working effect and stability of the model.

[0102] In one embodiment, the method of using the urban safety monitoring quantitative prediction model to perform quantitative prediction based on multi-source heterogeneous real-time data of each monitoring indicator includes:

[0103] Input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters;

[0104] Determine the reasonable state range of each monitoring indicator through the standard parameters of urban safety development, and determine the abnormal state based on the state change parameters and the reasonable state range of each monitoring indicator;

[0105] Determine the effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator based on the effect indicator parameters;

[0106] Determine safety risk responses based on the quantitative output of each monitoring indicator and issue early warnings based on safety risk responses.

[0107] The beneficial effects of the above technical solution are: by determining the safety risk effect indicators based on the state change parameters and then making quantitative predictions, accurate predictions can be made based on the core state change parameters of various monitoring indicators, ensuring the rationality and reliability of the prediction results.

[0108] In one embodiment, this embodiment also discloses a city safety monitoring quantitative prediction system based on big data analysis, such as Figure 3 As shown, the system includes:

[0109] The acquisition module 301 is used to determine multiple monitoring indicators of urban safety development, collect multi-source heterogeneous historical data of each monitoring indicator and perform pre-processing;

[0110] Selection module 302 is used to perform data coupling and feature extraction on the pre-processed multi-source heterogeneous historical data, and select an analysis model based on the data features and data industry characteristics;

[0111] A generation module 303 is used to train and cross-validate the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring;

[0112] The quantitative prediction module 304 is used to perform quantitative prediction based on multi-source heterogeneous real-time data of each monitoring indicator using the urban safety monitoring quantitative prediction model.

[0113] The working principle and beneficial effects of the above technical solution have been explained in the method embodiment and will not be repeated here.

[0114] In one embodiment, the acquisition module includes:

[0115] The first determination submodule is used to determine multiple security dimensions of urban security development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item;

[0116] The second determination submodule is used to integrate all monitoring indicators of each security dimension, determine the multiple data sources of each monitoring indicator, and determine the data privacy of each data source;

[0117] The collection submodule is used to determine data retrieval permissions based on data privacy and collect multi-source heterogeneous historical data for each monitoring indicator through data retrieval permissions;

[0118] The preprocessing submodule is used to clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

[0119] In one embodiment, the selection module includes:

[0120] The processing submodule is used to perform spatial and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling processing on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network;

[0121] The third determination submodule is used to extract statistical features and domain-specific features corresponding to the coupled data, and determine the prediction attributes based on the statistical features and domain-specific features;

[0122] The fourth determination submodule is used to determine the evaluation logic according to the prediction attributes and determine the model analysis target based on the evaluation logic and data industry characteristics;

[0123] The selection submodule is used to determine the model analysis modal mechanism according to the model analysis target and select the analysis model based on the model analysis modal mechanism.

[0124] In one embodiment, the generating module includes:

[0125] Define submodules to adjust analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model;

[0126] The training submodule is used to set the input and output layer parameters of the analysis model based on the main task and auxiliary tasks, and train the analysis model with multi-source heterogeneous historical data of each monitoring indicator;

[0127] The validation submodule is used to cross-validate the trained analysis model through stratified sampling and rolling time windows, and determine the eligibility of the multi-dimensional validation indicators of the analysis model based on the validation results;

[0128] The generation submodule is used to optimize the model based on the eligibility of multi-dimensional verification indicators to generate a quantitative prediction model for urban safety monitoring.

[0129] In one embodiment, Figure 4 As shown, the quantitative prediction module 304 includes:

[0130] The acquisition submodule 3041 is used to input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters;

[0131] The fifth determining submodule 3042 is configured to determine a reasonable state range for each monitoring indicator using the city safety development standard parameter, and determine an abnormal state based on the state change parameter and the reasonable state range for each monitoring indicator;

[0132] The sixth determining submodule 3043 is used to determine effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator according to the effect indicator parameters;

[0133] The judgment submodule 3044 is used to judge the safety risk response based on the quantitative output of each monitoring indicator and issue an early warning based on the safety risk response.

[0134] Those skilled in the art should understand that the first and second in the present invention simply refer to different application stages.

[0135] Other embodiments of the present disclosure will readily occur to those skilled in the art after considering the specification and practicing the disclosure herein. This application is intended to cover any variations, uses, or adaptations of the present disclosure that follow from the general principles of the present disclosure and include common knowledge or customary techniques in the art not disclosed herein. The description and examples are to be considered as exemplary only, with the true scope and spirit of the present disclosure being indicated by the following claims.

[0136] It should be understood that the present disclosure is not limited to the exact structures that have been described above and shown in the drawings, and that various modifications and changes can be made without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A quantitative prediction method for urban safety monitoring based on big data analysis, characterized in that: The following steps are involved: Determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator and perform pre-processing; Perform data coupling and feature extraction on pre-processed multi-source heterogeneous historical data, and select analysis models based on data characteristics and data industry characteristics; The analysis model is trained and cross-validated based on multi-source heterogeneous historical data for each monitoring indicator to generate a quantitative prediction model for urban safety monitoring; The quantitative prediction model of urban safety monitoring is used to make quantitative predictions based on multi-source heterogeneous real-time data of each monitoring indicator.

2. The urban safety monitoring quantitative prediction method based on big data analysis according to claim 1 is characterized in that: The aforementioned process of determining multiple monitoring indicators for urban safety development, collecting multi-source heterogeneous historical data for each monitoring indicator and pre-processing the data includes: Identify multiple security dimensions of urban safety development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item; Integrate all monitoring indicators for each security dimension, identify multiple data sources for each monitoring indicator, and determine the data privacy of each data source; Determine data access permissions based on data privacy, and use data access permissions to collect multi-source heterogeneous historical data for each monitoring indicator; Clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

3. The urban safety monitoring quantitative prediction method based on big data analysis according to claim 1 is characterized in that: The aforementioned data coupling and feature extraction of pre-processed multi-source heterogeneous historical data, and selection of analysis models based on data features and data industry characteristics, include: Perform spatial and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network; Extracting statistical features and domain-specific features corresponding to the coupled data, and determining prediction attributes based on the statistical features and domain-specific features; Determine the evaluation logic based on the predicted attributes, and determine the model analysis objectives based on the evaluation logic and data industry characteristics; Determine the model analysis modal mechanism according to the model analysis objectives, and select the analysis model based on the model analysis modal mechanism.

4. The method for quantitative prediction of urban safety monitoring based on big data analysis according to claim 1 is characterized in that: The analysis model is trained and cross-validated based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring, including: Adjust the analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model; Set the input and output layer parameters of the analysis model based on the main and auxiliary tasks, and train the analysis model using multi-source heterogeneous historical data for each monitoring indicator; The trained analysis model is cross-validated using a stratified sampling method and a rolling time window, and the eligibility of the multi-dimensional validation indicators of the analysis model is determined based on the validation results; Based on the eligibility of multi-dimensional verification indicators, the model is optimized to generate a quantitative prediction model for urban safety monitoring.

5. The urban safety monitoring quantitative prediction method based on big data analysis according to claim 1 is characterized in that: The quantitative prediction model for urban safety monitoring is used to perform quantitative prediction based on multi-source heterogeneous real-time data of each monitoring indicator, including: Input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters; Determine the reasonable state range of each monitoring indicator through the standard parameters of urban safety development, and determine the abnormal state based on the state change parameters and the reasonable state range of each monitoring indicator; Determine the effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator based on the effect indicator parameters; Determine safety risk responses based on the quantitative output of each monitoring indicator and issue early warnings based on safety risk responses.

6. A city safety monitoring quantitative prediction system based on big data analysis, characterized in that: The system includes: The acquisition module is used to determine multiple monitoring indicators for urban safety development, collect multi-source heterogeneous historical data for each monitoring indicator, and perform pre-processing; The selection module is used to perform data coupling and feature extraction on pre-processed multi-source heterogeneous historical data, and select analysis models based on data features and data industry characteristics; A generation module is used to train and cross-validate the analysis model based on multi-source heterogeneous historical data of each monitoring indicator to generate a quantitative prediction model for urban safety monitoring; The quantitative prediction module is used to make quantitative predictions based on multi-source heterogeneous real-time data of each monitoring indicator using the urban safety monitoring quantitative prediction model.

7. The urban safety monitoring quantitative prediction system based on big data analysis according to claim 6 is characterized in that: The acquisition module includes: The first determination submodule is used to determine multiple security dimensions of urban security development, obtain multiple statistical items for each security dimension, and determine monitoring indicators for each statistical item; The second determination submodule is used to integrate all monitoring indicators of each security dimension, determine the multiple data sources of each monitoring indicator, and determine the data privacy of each data source; The collection submodule is used to determine data retrieval permissions based on data privacy and collect multi-source heterogeneous historical data for each monitoring indicator through data retrieval permissions; The preprocessing submodule is used to clean multi-source heterogeneous historical data, remove duplicate, missing and invalid data, and perform data format conversion and standardization preprocessing.

8. The urban safety monitoring quantitative prediction system based on big data analysis according to claim 6 is characterized in that: The selection module includes: The processing submodule is used to perform spatial and temporal alignment on the pre-processed multi-source heterogeneous historical data, and perform data coupling processing on the processed multi-source heterogeneous historical data through a heterogeneous graph neural network; The third determination submodule is used to extract statistical features and domain-specific features corresponding to the coupled data, and determine the prediction attributes based on the statistical features and domain-specific features; The fourth determination submodule is used to determine the evaluation logic according to the prediction attributes and determine the model analysis target based on the evaluation logic and data industry characteristics; The selection submodule is used to determine the model analysis modal mechanism according to the model analysis target and select the analysis model based on the model analysis modal mechanism.

9. The urban safety monitoring quantitative prediction system based on big data analysis according to claim 6 is characterized in that: The generation module includes: Define submodules to adjust analysis model parameters based on the data quality of multi-source heterogeneous historical data and define the main and auxiliary tasks of the analysis model; The training submodule is used to set the input and output layer parameters of the analysis model based on the main task and auxiliary tasks, and train the analysis model with multi-source heterogeneous historical data of each monitoring indicator; The validation submodule is used to cross-validate the trained analysis model through stratified sampling and rolling time windows, and determine the eligibility of the multi-dimensional validation indicators of the analysis model based on the validation results; The generation submodule is used to optimize the model based on the eligibility of multi-dimensional verification indicators to generate a quantitative prediction model for urban safety monitoring.

10. The urban safety monitoring quantitative prediction system based on big data analysis according to claim 6 is characterized in that: The quantitative prediction module includes: The acquisition submodule is used to input the multi-source heterogeneous real-time data of each monitoring indicator into the output layer of the urban safety monitoring quantitative prediction model to obtain the state change parameters; The fifth determination submodule is used to determine the reasonable state interval of each monitoring indicator through the urban safety development standard parameters, and determine the abnormal state according to the state change parameters and the reasonable state interval of each monitoring indicator; a sixth determination submodule, configured to determine effect indicator parameters under abnormal conditions, and determine the quantitative output of each monitoring indicator according to the effect indicator parameters; The judgment submodule is used to judge the safety risk response based on the quantitative output of each monitoring indicator and to issue an early warning based on the safety risk response.

Images