Personalized security risk early warning method and system integrated with rapid distress call, and medium

Abstract

The invention provides a personalized safety risk early warning method and system integrated with rapid help calling and a medium, and relates to the technical field of safety early warning, and the method comprises the steps: collecting and obtaining an information platform data set through employing a crawler technology, and carrying out the safety risk recognition, and obtaining a risk region feature set and a risk region position set; calculating to obtain a real-time distance information set of the risk area position set and the current position information of the target user; performing multi-dimensional risk assessment on the risk area feature set and the real-time distance information set to obtain risk area hidden danger factors; and constructing a dynamic safety early warning strategy library, carrying out strategy analysis on the hidden danger factors of the risk area, determining a target safety risk early warning strategy, and carrying out dynamic safety early warning on the target user. The technical problems of incomplete safety risk identification and insufficient early warning accuracy caused by limited data sources in the prior art are solved, and the technical effects of improving the real-time performance and accuracy of risk early warning and the flexibility of emergency response are achieved.

Description

Technical Field

[0001] This application relates to the technical field of safety warning, and particularly to a personalized safety risk warning method, system and medium integrating rapid emergency call. Background Art

[0002] As an important tool for dealing with emergencies and reducing losses, the safety risk warning system is widely used in many fields such as public safety, transportation, and environmental monitoring. Existing safety risk warning systems mostly rely on fixed sensor networks or data collected manually. Although these methods can provide risk information to a certain extent, due to limited data sources, the scope and accuracy of safety risk identification are greatly restricted. For example, the coverage of the sensor network may not cover all potential risk areas, and manually collected data has problems such as poor timeliness and strong subjectivity. Secondly, traditional warning systems usually rely on preset dangerous situations and patterns, with lagging information updates and unable to accurately reflect the rapidly changing safety situation. For example, sudden social events, real-time geographical environment changes, etc. are not timely incorporated into risk assessment, thus reducing the effectiveness and response ability of the warning. Summary of the Invention

[0003] This application provides a personalized safety risk warning method, system and medium integrating rapid emergency call, solving the technical problem that in the prior art, due to limited data sources, safety risk identification is incomplete and warning accuracy is insufficient, and achieving the technical effects of improving the timeliness, accuracy of risk warning and the flexibility of emergency response.

[0004] In view of the above problems, on the one hand, this application provides a personalized safety risk warning method integrating rapid emergency call, and the method includes: collecting and obtaining an information platform data set by using web crawler technology, performing safety risk identification on the information platform data set to obtain a risk area feature set and a risk area location set; obtaining the current location information of the target user, and calculating to obtain a real-time distance information set of the risk area location set and the current location information; respectively performing multi-dimensional risk assessment on the risk area feature set and the real-time distance information set to obtain risk area hazard factors; constructing a dynamic safety warning strategy library, which integrates a personalized warning strategy set and a rapid emergency call strategy; performing strategy analysis on the risk area hazard factors based on the dynamic safety warning strategy library to determine a target safety risk warning strategy, and performing dynamic safety warning on the target user through the target safety risk warning strategy.

[0005] On the other hand, the present application also provides a personalized safety risk warning system integrating rapid emergency call. The system includes: a risk area identification module, which is used to collect and obtain an information platform data set by using web crawler technology, identify safety risks in the information platform data set, and obtain a risk area feature set and a risk area location set; a real-time distance acquisition module, which is used to obtain the current location information of the target user and calculate a real-time distance information set of the risk area location set and the current location information; a multi-dimensional risk assessment module, which is used to perform multi-dimensional risk assessment on the risk area feature set and the real-time distance information set respectively to obtain risk area hazard factors; a policy library construction module, which is used to construct a dynamic safety warning policy library, and the dynamic safety warning policy library integrates a personalized warning policy set and a rapid emergency call policy; a safety risk warning module, which is used to perform policy analysis on the risk area hazard factors based on the dynamic safety warning policy library, determine a target safety risk warning policy, and perform dynamic safety warning on the target user through the target safety risk warning policy.

[0006] In a third aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned personalized safety risk warning method integrating rapid emergency call are implemented.

[0007] One or more technical solutions provided in the present application have at least the following technical effects or advantages:

[0008] By widely collecting data from information platforms through web crawler technology, the limitations of traditional data sources can be broken through, and richer and more comprehensive risk-related information can be obtained. After the data is identified for safety risks, a risk area feature set and a location set are generated, providing accurate input for subsequent risk assessment. By monitoring the real-time location of the user and calculating the real-time distance information set from the risk area, the relevance between the user and potential risks is quantified. The real-time distance information set provides a dynamic spatio-temporal dimension for subsequent risk assessment, so as to adjust the warning policy in real time according to the change of the user's location, enhancing the timeliness and pertinence of the warning. Multi-dimensional risk assessment is performed on the risk area feature set and the real-time distance information set respectively, comprehensively considering the characteristics of the risk area and the distance between the user and the risk area, and generating risk area hazard factors. This multi-dimensional assessment method can more comprehensively and accurately reflect the actual risks faced by the user. Integrating a personalized warning policy set and a rapid emergency call policy to construct a dynamic safety warning policy library can provide customized warnings according to the preferences and needs of the user, and at the same time provide an instant rescue support plan for the user in case of emergency, significantly improving the practicality of the warning and the user experience. Through policy analysis of the risk hazard factors, the most suitable warning policy can be accurately selected and dynamic warning information can be pushed to the user in real time, not only improving the accuracy and personalization level of the warning, but also enhancing the emergency response efficiency through the rapid emergency call mechanism.

[0009] In summary, by integrating multi-source data, conducting multi-dimensional risk assessment, constructing a dynamic security warning strategy library, and performing personalized risk warning based on this strategy library, the present application can comprehensively and accurately identify the relationship between risk areas and target users, achieve personalized security risk warning through dynamic strategy analysis, and integrate a quick help-seeking strategy during the warning, improving the efficiency of dealing with risks. Thus, the real-time performance, accuracy of security risk warning, and the efficiency and personalization of emergency response are comprehensively improved, providing a more reliable and efficient security risk warning service for target users.

[0010] The above description is only an overview of the technical solution of the present application. In order to better understand the technical means of the present application, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features, and advantages of the present application more obvious and understandable, the following specifically illustrates the specific embodiments of the present application. Brief Description of the Drawings

[0011] Figure 1 It is a schematic flowchart of an integrated quick help-seeking personalized security risk warning method provided by an embodiment of the present application.

[0012] Figure 2 It is a schematic flowchart of obtaining a risk area location set in the integrated quick help-seeking personalized security risk warning method provided by an embodiment of the present application.

[0013] Figure 3 It is a schematic structural diagram of an integrated quick help-seeking personalized security risk warning system provided by an embodiment of the present application.

[0014] Description of the reference numerals: Risk area identification module 10, real-time distance acquisition module 20, multi-dimensional risk assessment module 30, strategy library construction module 40, security risk warning module 50. Detailed Embodiments

[0015] By providing an integrated quick help-seeking personalized security risk warning method, system, and medium, the embodiment of the present application solves the technical problem in the prior art that due to limited data sources, the security risk identification is not comprehensive and the warning accuracy is insufficient, and achieves the technical effect of improving the real-time performance, accuracy of risk warning, and flexibility of emergency response.

[0016] Embodiment 1, as Figure 1 shown, the embodiment of the present application provides an integrated quick help-seeking personalized security risk warning method, and the method includes:

[0017] Step S1: Use web crawler technology to collect and obtain an information platform data set, and perform security risk identification on the information platform data set to obtain a risk area feature set and a risk area location set.

[0018] Specifically, a crawler tool (such as Scrapy or BeautifulSoup in Python) is used to access the web pages of the information platform, obtain the data content in the web pages, and form an information platform dataset. Then, security risk identification is performed on this dataset, and natural language processing (NLP) techniques (such as a text classification model based on deep learning) are used to analyze the data and identify the security risk information therein. For example, it is judged whether an event belongs to a security risk through keyword matching (such as keywords related to security risks like "earthquake", "flood", etc.) or semantic analysis. The characteristics of the risk area (such as risk type, severity, etc.) and location information (such as the longitude and latitude or the region name of the event occurrence location) are extracted from the identified risk events, respectively forming a risk area characteristic set and a risk area location set.

[0019] Collecting data from external information platforms through crawler technology expands the scope of data sources and enables real-time acquisition of real-time information about potential security risks, thus improving the timeliness and accuracy of early warnings.

[0020] Step S2: Obtain the current location information of the target user, and calculate and obtain a real-time distance information set of the risk area location set and the current location information.

[0021] Specifically, the current location information of the target user (represented in the form of longitude and latitude) is obtained through the positioning function (such as GPS positioning, Beidou positioning, etc.) in the user's mobile device (such as a mobile phone) or network positioning technology (such as Wi-Fi positioning). Then, a geographic information system (GIS) tool or a simple geographic distance calculation formula (such as the Haversine formula) is used to calculate the distance between the current location information of the target user and each risk area location in the risk area location set obtained in step S1, and the calculated distance information is stored as a set to generate a real-time distance information set. As the user's location changes, the latest user location can be obtained in real time to update the distance information and improve the real-time performance of security risk early warnings.

[0022] Through distance calculation, a spatial connection between the risk area and the target user is established, providing an important distance factor basis for subsequent assessment of the impact of risks on the target user, which helps to more accurately judge the relevance of risks.

[0023] Step S3: Perform multi-dimensional risk assessments on the risk area characteristic set and the real-time distance information set respectively to obtain risk area hidden danger factors.

[0024] Specifically, the risk is comprehensively evaluated from multiple dimensions (such as the characteristics of the risk area, the distance from the user, etc.). Based on the information in the risk area feature set (such as risk type, level), a basic risk value is assigned to each risk area. For example, for a natural disaster risk area, if it is an earthquake-active area and located at a plate boundary (characteristic), a relatively high basic risk value will be given. According to the real-time distance information set, the distance weight is calculated. The closer the distance, the higher the weight, indicating that the risk is more urgent. Then, using weighted summation or other comprehensive evaluation algorithms (such as the analytic hierarchy process), the risk area feature set and the real-time distance information set are comprehensively calculated to obtain the hidden danger factor of the risk area. This hidden danger factor of the risk area is used to measure the potential danger degree of the risk area.

[0025] Evaluating risks by integrating multiple factors avoids the one-sidedness of single-factor evaluation, can more accurately measure the potential danger degree of the risk area, and provides a more scientific basis for formulating subsequent early warning strategies.

[0026] Step S4: Construct a dynamic safety early warning strategy library, which integrates a personalized early warning strategy set and a quick help strategy.

[0027] Specifically, the dynamic safety early warning strategy library is a collection containing various personalized safety early warning strategies and quick help strategies. This collection is dynamic and can be updated and adjusted according to different risk situations. Among them, the personalized early warning strategy set is an early warning strategy customized according to the user's preferences, needs or usage scenarios; the quick help strategy is a strategy to quickly initiate rescue or help in case of an emergency.

[0028] Constructing a dynamic safety early warning strategy library requires considering multiple factors. First, collect various possible safety early warning strategies. For example, early warning strategies for natural disasters (such as reminding users to find a safe shelter during an earthquake warning), early warning strategies for social security incidents (such as reminding users to avoid abnormal event areas), etc. For the personalized early warning strategy set, it needs to be customized according to the user's custom information, such as the preferred notification method, the risk types of concern, etc. The quick help strategy needs to establish a connection interface with local rescue agencies and integrate functions such as one-key help and emergency contact notification to ensure quick initiation of rescue in high-risk situations. In terms of technical implementation, a database management system (such as MySQL) can be used to store these strategies, and the dynamic update and invocation of the strategies can be achieved through programming (such as Java or Python).

[0029] The personalized early warning strategy set in the dynamic safety early warning strategy library can meet the needs of different users, improve the pertinence of early warning, while the quick help strategy enhances the timeliness of dealing with risks, and overall improves the practicality of safety risk early warning.

[0030] Step S5: Based on the dynamic security warning policy library, conduct policy analysis on the risk area hazard factors, determine the target security risk warning policy, and conduct dynamic security warning for the target user through the target security risk warning policy.

[0031] Specifically, use the risk area hazard factors obtained in step S3 as input, and search and match in the dynamic security warning policy library constructed in step S4. For example, if the risk area hazard factors indicate a high earthquake risk and the target user is near the epicenter, then search in the policy library for a warning policy suitable for this situation, such as sending an emergency earthquake warning message to inform the user to immediately find a sturdy shelter. After determining the target security risk warning policy, send the warning information to the target user through the notification system of the mobile device (such as the notification bar or push notification). During the warning process, adjust the warning policy in real time according to the changes in the risk hazard factors, and update the warning information to ensure that the target user always receives the latest risk tips.

[0032] By matching the risk area hazard factors with the dynamic security warning policy library, the most suitable warning policy is determined, realizing personalized and dynamic security risk warning for the target user, and improving the accuracy and effectiveness of the warning.

[0033] Furthermore, as Figure 2 shown, step S1 includes:

[0034] Step S11: Clean and deduplicate the data set of the information platform according to the data application standard to obtain a standard information platform data set.

[0035] Step S12: Use natural language processing technology to conduct text analysis on the standard information platform data set, and extract and obtain an associated information risk characteristic set.

[0036] Step S13: Train and construct a security risk identification model, identify and mark the associated information risk characteristic set based on the security risk identification model, and output the risk area feature set.

[0037] Step S14: Conduct geographical location extraction and latitude and longitude conversion based on the risk area feature set to obtain the risk area location set.

[0038] Specifically, the data application standard is a series of specifications and guidelines set to enable the effective utilization of data. For example, the effective format of data in the information platform dataset, the requirements for data integrity, etc. The information platform dataset is inspected according to the data application standard, and data cleaning and deduplication processes are carried out to obtain the standard information platform dataset. For data cleaning, it can be checked whether the data format meets the requirements. For example, it can be checked whether the date is in the correct format (such as "YYYY-MM-DD"), and the data that does not meet the requirements is corrected or deleted. For deduplication, database query statements or specialized data processing tools (such as the Pandas library in Python) can be used. For example, in Pandas, the drop_duplicates function can be used to delete duplicate rows in the dataset. Through data cleaning and deduplication, the quality of the information platform dataset is improved, invalid data and redundant data are removed, making subsequent data processing and analysis more efficient and accurate, and providing high-quality input data for building an accurate security risk identification model.

[0039] The associated information risk feature set is a set of data related to risk features extracted from the standard information platform dataset through text analysis. Natural language processing tools and algorithms are used to perform text analysis on the standard information platform dataset. For example, open-source natural language processing toolkits such as NLTK (Natural Language Toolkit) or Stanford NLP can be used. First, lexical analysis is carried out to split the text into words, and then syntactic analysis is performed to determine the relationships between words. Then semantic understanding is carried out to identify the words and phrases related to risk features. For example, for a sentence like "The flood submerged a large area of farmland and the water level reached 3 meters", through natural language processing technology, information such as "flood" (risk type) and "3 meters" (water level, risk feature) can be identified, and these information are extracted to form the associated information risk feature set.

[0040] Using machine learning algorithms (such as decision tree algorithms or neural network algorithms), a security risk identification model is trained based on the labeled training dataset (such as the text of known risk events). The to-be-identified associated information risk feature set is input into the trained security risk identification model, and the model makes identification marks according to the patterns learned before, and outputs the risk area feature set. For example, for a new piece of information, if it is identified as being about earthquake risk, risk area features such as the magnitude range and focal depth of the earthquake can be marked. Through the security risk identification model constructed by training, it can automatically make identification marks on the associated information risk feature set, improving the efficiency and accuracy of security risk identification, and reducing the workload and errors of manual identification.

[0041] Extract the geographical location and convert the longitude and latitude from the risk area feature set to obtain the risk area location set. For geographical location extraction, information extraction techniques can be used to find geographical location-related words or phrases, such as city names and street names, from the risk area feature set. Then, use the geocoding service to convert the extracted geographical location names into longitude and latitude coordinates, thereby obtaining the risk area location set. Representing the geographical location of the risk area in the form of longitude and latitude coordinates accurately facilitates subsequent operations such as calculating the distance to the target user, providing precise geographical location data for the entire security risk warning.

[0042] Furthermore, step S13 includes:

[0043] Step S131: Obtain a historical risk area data set, which includes historical information risk characteristic data and corresponding risk area feature data.

[0044] Step S132: Identify the risk types for the historical risk area data set to obtain a historical risk area type sample set.

[0045] Step S133: Evaluate and identify the risk levels for the historical risk area data set to obtain a historical risk area level sample set.

[0046] Step S134: Based on the historical risk area type sample set and the historical risk area level sample set, perform model training and merging to construct a security risk identification model.

[0047] Specifically, obtain data related to historical risk areas from existing data repositories (such as the enterprise's internal risk data warehouse or a public risk data statistics platform). For example, relevant data on flood disasters can be obtained from the historical disaster data records of the meteorological department, including the location where the flood occurred, the water level height at that time, etc. as historical information risk characteristic data, and at the same time, obtain the topographical features and population distribution of the area as risk area feature data, thereby forming a historical risk area data set.

[0048] Identify the risk types for the historical risk area data set through manual annotation or by using existing classification rules. For example, for data on earthquake risk areas, mark it as the "earthquake" type; for data on flood risk areas, mark it as the "flood" type, etc. Database management tools (such as MySQL) can be used to store and manage these marked sample data to form a historical risk area type sample set.

[0049] According to the pre-set risk level assessment criteria, conduct risk level assessment and identification on the historical risk area dataset, and divide and mark the risk levels corresponding to each historical risk area data. For example, determine the risk level of each sample based on some indicators in the historical data (such as the intensity and impact range of disasters), and divide the risks into three levels: high, medium, and low. Exemplarily, for flood disasters, if the water level exceeds a certain height and the affected area is large, it is marked as a high risk level; if the water level and the affected area are at a medium level, it is marked as a medium risk level. Data processing software (such as the Pandas library in Python) can be used to assist in the marking operation, and the marked sample data is stored as a historical risk area level sample set.

[0050] Select a machine learning algorithm, such as a decision tree algorithm, a support vector machine (SVM) algorithm, or a neural network algorithm. Combine the historical risk area type sample set and the historical risk area level sample set as training data and input it into the algorithm, and use the TensorFlow or PyTorch framework to train the algorithm model. During the training process, the algorithm will learn the characteristics and laws of risks based on the risk type markings and risk level markings in the sample set, thereby generating a security risk identification model. For example, the decision tree algorithm will construct a decision tree and learn how to distinguish different risk types and judge the risk level by continuously dividing the training data. By combining the two sample sets for model training, the constructed security risk identification model can accurately identify both the risk type and the risk level, improving the comprehensive performance of the model and making it more comprehensive and accurate in security risk identification.

[0051] Further, step S134 includes:

[0052] Step S134-1: Conduct model supervised training based on the historical risk area type sample set to obtain a semi-model for risk area type identification.

[0053] Step S134-2: Conduct model supervised training on the historical risk area level sample set to obtain a semi-model for risk area level identification.

[0054] Step S134-3: Combine and fuse the risk area type identification semi-model and the risk area level identification semi-model to generate the security risk identification model.

[0055] Specifically, during the model training process, the historical risk area type sample set and the historical risk area level sample set are used as training data for model supervised training respectively, to obtain a semi-model for risk area type recognition and a semi-model for risk area level recognition. Taking the semi-model for risk area type recognition as an example, first, the historical risk area type sample set is divided into a training set and a test set. The training set is used to train the model, and the test set is used to evaluate the performance of the model. During the training process, the input features in the sample set (such as historical information risk characteristic data and corresponding risk area characteristic data) are input into the Support Vector Machine (SVM) algorithm. The algorithm continuously adjusts the parameters of the model according to the output of the marked risk type, to minimize the error between the predicted risk type and the actual risk type. For example, when a set of feature data about an earthquake is input, the model gradually learns the corresponding relationship between these features and the "earthquake" risk type. After multiple iterative trainings, a semi-model for risk area type recognition is obtained.

[0056] For the semi-model for risk area level recognition, a Multi-Layer Perceptron (MLP) neural network is selected for training. The historical risk area level sample set is divided into a training set and a test set. During the training process, the input features related to the risk level in the sample set (such as the intensity and influence range of disasters) are input into the MLP neural network. The neural network continuously adjusts the connection weights between neurons according to the output of the marked risk level. For example, when a set of data about the intensity and influence range of a flood disaster is input, the neural network learns the corresponding relationship between these features and high, medium, and low risk levels. After multiple rounds of training, a semi-model for risk area level recognition is obtained.

[0057] The semi-model for risk area type recognition and the semi-model for risk area level recognition are merged and fused into a complete model, namely the safety risk recognition model, which can simultaneously identify the risk area type and the risk area level. Multiple methods can be used for merging and fusion. A simple method is to perform weighted combination on the outputs of the two semi-models. For example, let the output of the semi-model for risk area type recognition be y1 (representing the risk type), and the output of the semi-model for risk area level recognition be y2 (representing the risk level). Then the output y of the final safety risk recognition model is y = w1y1 + w2y2, where w1 and w2 are weights determined according to experience or further experiments. More complex fusion methods can also be used, such as constructing a new neural network structure, embedding the two semi-models as sub-networks into it, and simultaneously training the entire network through the backpropagation algorithm, so that the two semi-models work together under the new structure.

[0058] Furthermore, step S134-3 includes:

[0059] Step 1: Merge and fuse the semi-model for risk area type recognition and the semi-model for risk area level recognition to construct an initial risk recognition model.

[0060] Step 2: Use the mean squared error function to perform loss evaluation calculations on the initial risk identification model to obtain model loss data.

[0061] Step 3: Based on the model loss data, perform feedback optimization and update on the initial risk identification model to generate the security risk identification model.

[0062] Specifically, merge and fuse the semi - models for risk area type identification and risk area level identification. As mentioned before, methods such as weighted combination or constructing a new neural network structure can be used for merging and fusing to construct the initial risk identification model.

[0063] Input the test data set (including sample data with marked risk types and risk levels) into the initial risk identification model to obtain the predicted output. Then, calculate the model loss data according to the formula of the mean squared error function. For example, if there are 100 test samples, input these 100 samples into the model in sequence to get the predicted results, and then compare with the true results for calculation to obtain a specific mean squared error value as the model loss data. Quantify the prediction error of the initial risk identification model through the mean squared error function, intuitively understand the current performance status of the model, and provide a clear goal and basis for subsequent model optimization.

[0064] Perform feedback optimization and update on the initial risk identification model according to the model loss data, such as adjusting weights, learning rates, etc., to reduce the prediction error and improve the model performance. For the initial risk identification model with a neural network structure, the backpropagation algorithm can be used for feedback optimization and update. According to the calculated model loss data, the backpropagation algorithm will calculate the contribution degree of each parameter to the loss, and then adjust the parameters of the model (such as weights and biases in the neural network) according to a certain learning rate. For example, if it is found that the weight of a certain neuron contributes greatly to the model loss, appropriately adjust the value of this weight so that the model can reduce the error in the next prediction. For other machine learning algorithms, use corresponding optimization algorithms, such as the gradient descent method, to adjust the model parameters according to the model loss data. After multiple iterations, generate the security risk identification model. Through feedback optimization and update, the accuracy of the model can be effectively improved, enabling the security risk identification model to more accurately identify the risk area type and risk area level, reduce the prediction error, and improve the reliability of the model in practical applications.

[0065] Further, step S3 includes:

[0066] Step S31: Assign weights to each risk type in the risk area feature set to obtain risk type weight assignment information.

[0067] Step S32: Based on the risk type weight distribution information, perform weighted calculations on each risk level in the risk area feature set to determine the risk area hidden danger parameter set.

[0068] Step S33: According to the real-time distance information set, determine the area hidden danger enhancement influence coefficient set.

[0069] Step S34: Based on the area hidden danger enhancement influence coefficient set, perform addition supplementation and summation calculations on the risk area hidden danger parameter set to obtain the risk area hidden danger factor.

[0070] Specifically, analyze the risk types (such as fires, natural disasters, etc.) included in the risk area feature set. According to expert experience or historical data analysis, assign weights to each risk type. The weight values usually range from 0 to 1, and the larger the value, the more important the risk type. Store the risk type and its corresponding weight value as the risk type weight distribution information. Exemplarily, the expert evaluation method can be used, that is, invite experts in relevant fields to determine the weights based on experience and knowledge; data analysis methods such as principal component analysis can also be used to determine the weights from a large amount of historical data, so as to obtain the risk type weight distribution information. The risk type weight distribution information provides an important basis for subsequent risk assessments, can highlight the importance of different risk types in the overall risk, and make the risk assessment more scientific and reasonable.

[0071] Convert the risk level to a numerical value (such as high = 3, medium = 2, low = 1). According to the risk type weight distribution information, perform weighted calculations on the risk level. Exemplarily, the calculation formula is: hidden danger parameter = risk level value × risk type weight. Store the calculated hidden danger parameter as the risk area hidden danger parameter set. Through weighted calculations, the impacts of both the risk type and the risk level are comprehensively considered, and the obtained risk area hidden danger parameter set can more accurately reflect the hidden danger degree of the risk area, providing important data support for comprehensively evaluating the risk.

[0072] Set weights according to the distance. The closer the distance, the higher the weight. For example, when the distance is less than 1 km, the weight = 1.5; when the distance is between 1 km and 5 km, the weight = 1.2; when the distance is greater than 5 km, the weight = 1.0. According to the real-time distance information set, calculate the enhancement influence coefficient of each risk area. Store the calculated influence coefficient as the area hidden danger enhancement influence coefficient set. By setting the distance weight, the spatio-temporal correlation between the risk hidden danger and the user's location is quantified, enhancing the dynamics of the risk assessment.

[0073] Multiply each parameter in the risk area hidden danger parameter set by the corresponding enhancement influence coefficient to obtain the adjusted hidden danger parameter, and then sum up the adjusted hidden danger parameters to obtain the risk area hidden danger factor. Store the calculated risk area hidden danger factor as the final risk assessment result. The risk area hidden danger factor can dynamically reflect the risk level faced by users and provide an accurate quantitative basis for the subsequent selection of early warning strategies.

[0074] The above steps comprehensively consider the characteristics, risk levels, and distance factors of the risk area. The obtained risk area hidden danger factor can comprehensively and accurately measure the comprehensive hidden danger degree of the risk area, providing a scientific basis for the formulation of subsequent safety early warning strategies.

[0075] Furthermore, step S5 includes:

[0076] Step S51: Set the high-risk hidden danger factor threshold. If the risk area hidden danger factor set does not reach the high-risk hidden danger factor threshold, trigger the personalized early warning strategy set in the dynamic safety early warning strategy library.

[0077] Step S52: Based on the personalized early warning strategy set, perform dynamic strategy matching on the risk area hidden danger factor to determine the personalized safety risk early warning strategy.

[0078] Specifically, the high-risk hidden danger factor threshold is a preset numerical standard used to measure whether the risk area hidden danger factor reaches the high-risk level. According to the risk assessment standard or historical data, set the high-risk hidden danger factor threshold, compare the risk area hidden danger factor set with the high-risk hidden danger factor threshold. If the hidden danger factor does not reach the high-risk threshold, trigger the personalized early warning strategy set from the dynamic safety early warning strategy library. These personalized early warning strategy sets are collections of early warning strategies customized according to user preferences and scenario requirements, and can provide corresponding early warning measures for specific risk situations.

[0079] According to the numerical range or characteristics of the hidden danger factors in the risk area, select the matching strategy from the personalized early warning strategy set, and adjust the early warning strategy in real time according to the changes of the hidden danger factors to ensure the timeliness and accuracy of the early warning information. The matched strategy is used as the final personalized safety risk early warning strategy. Multiple matching methods can be adopted. For example, rule-based matching. Exemplarily, if the hidden danger factors in the risk area are mainly related to natural disaster risks and the risk level is low, then select the strategy related to low natural disaster risks from the personalized early warning strategy set, such as sending general reminder information about natural disaster prevention (such as text message reminders to pay attention to weather changes and make disaster prevention preparations, etc.); model-based matching methods can also be used. Exemplarily, construct a strategy matching model, take the hidden danger factors in the risk area as input, and the model outputs the most matching personalized safety risk early warning strategy. By dynamically matching the strategy to determine the personalized safety risk early warning strategy, the accuracy and effectiveness of the early warning can be improved, making the early warning measures more in line with the actual risk situation, so as to better guide users to prevent and respond to risks.

[0080] Further, step S5 further includes:

[0081] Step S53: When the hidden danger factor set in the risk area reaches the high-risk hidden danger factor threshold, trigger the rapid call for help strategy in the dynamic safety early warning strategy library, and output the rapid call for help strategy as the target safety risk early warning strategy.

[0082] Specifically, when the value of the hidden danger factor set in the risk area reaches the pre-set high-risk hidden danger factor threshold, automatically retrieve and trigger the rapid call for help strategy from the dynamic safety early warning strategy library, such as sending an emergency distress signal to relevant rescue departments and notifying surrounding emergency resources. By promptly activating the rapid call for help strategy in high-risk situations, ensure a rapid response to high-risk situations and minimize potential losses and hazards to the greatest extent. Output the rapid call for help strategy as the target safety risk early warning strategy, so that the entire early warning system has clear and effective countermeasures when facing high-risk situations, improving the overall risk response ability.

[0083] In summary, the personalized safety risk early warning method integrating rapid call for help provided by the embodiments of the present application has the following technical effects:

[0084] In the embodiments of the present application, first, web crawler technology is used to collect and clean data from information platforms to broaden the data sources. Natural language processing technology is utilized to extract risk characteristics, and a security risk identification model is combined to mark risk areas, obtaining a risk area feature set and a location set, providing more diverse real-time information support for subsequent risk identification. Then, by calculating the real-time distance information between the target user and the risk areas, combined with a multi-dimensional risk assessment method, a comprehensive analysis of the hidden danger factors in the risk areas is carried out, and the hidden danger factors are accurately quantified through weight assignment and weighted calculation, significantly improving the accuracy and timeliness of early warning. Secondly, a dynamic security early warning strategy library is constructed, integrating personalized early warning strategies and rapid SOS strategies, and the early warning plan is dynamically adjusted according to the level of hidden danger factors. When the hidden danger factors in the risk areas do not reach the high-risk threshold, the personalized early warning strategy is triggered to push appropriate security reminders to users; while when the hidden danger factors reach the high-risk threshold, the rapid SOS strategy is triggered to automatically activate the emergency call for help and rescue response mechanism to ensure that users can get timely help in extreme situations. The introduction of the dynamic security early warning strategy library and the personalized early warning strategy set not only enhances the flexibility and adaptability of early warning, but also provides customized security guarantees for users. The integrated rapid SOS strategy provides instant rescue support for users in emergencies, further improving the emergency response efficiency. Overall, the embodiments of the present application comprehensively improve the real-time performance, accuracy of security risk early warning, as well as the efficiency and personalization of emergency response, providing a more reliable and efficient security risk early warning service for target users.

[0085] Embodiment 2, as Figure 3 shown, based on the same inventive concept as the foregoing Embodiment 1, the embodiments of the present application provide a personalized security risk early warning system integrating rapid SOS, and the system includes:

[0086] A risk area identification module 10, configured to collect and obtain an information platform data set by using web crawler technology, and perform security risk identification on the information platform data set to obtain a risk area feature set and a risk area location set.

[0087] A real-time distance acquisition module 20, configured to acquire the current location information of a target user, and calculate and obtain a real-time distance information set of the risk area location set and the current location information.

[0088] A multi-dimensional risk assessment module 30, configured to perform multi-dimensional risk assessment on the risk area feature set and the real-time distance information set respectively to obtain risk area hidden danger factors.

[0089] A strategy library construction module 40, configured to construct a dynamic security early warning strategy library, and the dynamic security early warning strategy library integrates a personalized early warning strategy set and a rapid SOS strategy.

[0090] The security risk warning module 50 is used to perform policy analysis on the hidden danger factors in the risk area based on the dynamic security warning policy library, determine the target security risk warning policy, and perform dynamic security warning on the target user through the target security risk warning policy.

[0091] Further, the risk area identification module 10 in the embodiment of the present application is further used to perform the following steps:

[0092] Perform data cleaning and deduplication processing on the information platform data set according to the data application standard to obtain a standard information platform data set; use natural language processing technology to perform text analysis on the standard information platform data set, and extract and obtain an associated information risk characteristic set; train and construct a security risk identification model, and perform identification and marking on the associated information risk characteristic set based on the security risk identification model to output the risk area feature set; perform geographical location extraction and longitude and latitude conversion based on the risk area feature set to obtain the risk area location set.

[0093] Further, the risk area identification module 10 in the embodiment of the present application is further used to perform the following steps:

[0094] Obtain a historical risk area data set, where the historical risk area data set includes historical information risk characteristic data and corresponding risk area feature data; perform risk type identification on the historical risk area data set to obtain a historical risk area type sample set; perform risk level assessment and identification on the historical risk area data set to obtain a historical risk area level sample set; perform model training and merging based on the historical risk area type sample set and the historical risk area level sample set to construct a security risk identification model.

[0095] Further, the risk area identification module 10 in the embodiment of the present application is further used to perform the following steps:

[0096] Perform model supervised training based on the historical risk area type sample set to obtain a risk area type identification semi-model; perform model supervised training on the historical risk area level sample set to obtain a risk area level identification semi-model; merge and fuse the risk area type identification semi-model and the risk area level identification semi-model to generate the security risk identification model.

[0097] Further, the risk area identification module 10 in the embodiment of the present application is further used to perform the following steps:

[0098] Merge and integrate the risk area type recognition semi-model and the risk area level recognition semi-model to construct an initial risk recognition model; use the mean square error function to calculate the loss evaluation of the initial risk recognition model to obtain model loss data; based on the model loss data, perform feedback optimization and update on the initial risk recognition model to generate the security risk recognition model.

[0099] Further, the multi-dimensional risk assessment module 30 in the embodiment of the present application is further configured to perform the following steps:

[0100] Assign weights to each risk type in the risk area feature set to obtain risk type weight assignment information; perform weighted calculation on each risk level in the risk area feature set based on the risk type weight assignment information to determine the risk area hidden danger parameter set; according to the real-time distance information set, determine the area hidden danger strengthening influence coefficient set; based on the area hidden danger strengthening influence coefficient set, perform addition supplement and summation calculation on the risk area hidden danger parameter set to obtain the risk area hidden danger factor.

[0101] Further, the security risk warning module 50 in the embodiment of the present application is further configured to perform the following steps:

[0102] Set a high-risk hidden danger factor threshold. If the risk area hidden danger factor set does not reach the high-risk hidden danger factor threshold, trigger the personalized warning strategy set in the dynamic security warning strategy library; perform dynamic strategy matching on the risk area hidden danger factor based on the personalized warning strategy set to determine the personalized security risk warning strategy.

[0103] Further, the security risk warning module 50 in the embodiment of the present application is further configured to perform the following steps:

[0104] When the risk area hidden danger factor set reaches the high-risk hidden danger factor threshold, trigger the rapid call for help strategy in the dynamic security warning strategy library, and output the rapid call for help strategy as the target security risk warning strategy.

[0105] Through the foregoing detailed description of the personalized security risk warning method integrating rapid call for help in this specification, those skilled in the art can clearly know the personalized security risk warning system integrating rapid call for help in this embodiment. For the system disclosed in Embodiment 2, since it corresponds to the method disclosed in Embodiment 1, it has corresponding functional modules and beneficial effects. For the relevant parts, refer to the description in the method part.

[0106] Embodiment 3. Based on the same inventive concept as the personalized safety risk warning method integrating rapid help calling in Embodiment 1 described above, the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the above-described embodiment of the personalized safety risk warning method integrating rapid help calling, and can achieve the same technical effects. To avoid repetition, it will not be elaborated here.

[0107] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A personalized safety risk early warning method with integrated quick rescue call, characterized in that: The method comprises: Using crawler technology to collect and obtain information platform data sets, identify security risks on the information platform data sets, and obtain risk area feature sets and risk area location sets; Acquire the current location information of the target user, and calculate and obtain a real-time distance information set between the risk area location set and the current location information; Performing multi-dimensional risk assessment on the risk area feature set and the real-time distance information set respectively to obtain a risk area hidden danger factor; Building a dynamic safety warning strategy library, wherein the dynamic safety warning strategy library integrates a personalized warning strategy set and a quick rescue strategy; Based on the dynamic security warning strategy library, a strategy analysis is performed on the hidden danger factors in the risk area to determine a target security risk warning strategy, and a dynamic security warning is performed on the target user through the target security risk warning strategy.

2. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 1, characterized in that: The step of obtaining the risk area feature set and the risk area location set includes: Performing data cleaning and deduplication processing on the information platform data set according to the data application standard to obtain a standard information platform data set; Using natural language processing technology to perform text analysis on the standard information platform data set to extract a set of related information risk characteristics; Training and constructing a security risk identification model, identifying and marking the risk characteristic set of the associated information based on the security risk identification model, and outputting the risk area characteristic set; Geographic location extraction and longitude and latitude conversion are performed based on the risk area feature set to obtain the risk area location set.

3. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 2, characterized in that: The training constructs a security risk identification model, including: Obtaining a historical risk area data set, wherein the historical risk area data set includes historical information risk characteristic data and corresponding risk area characteristic data; Identifying the risk type of the historical risk area data set to obtain a sample set of historical risk area types; Performing risk level assessment and identification on the historical risk area data set to obtain a historical risk area level sample set; Model training is performed based on the historical risk area type sample set and the historical risk area level sample set to construct a safety risk identification model.

4. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 3, characterized in that: The construction of the security risk identification model includes: Performing model supervision training based on the historical risk area type sample set to obtain a risk area type identification semi-model; Performing model supervision training on the historical risk area level sample set to obtain a risk area level identification semi-model; The risk area type identification half model and the risk area level identification half model are merged and fused to generate the safety risk identification model.

5. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 4, characterized in that: The generating the security risk identification model comprises: The risk area type identification half model and the risk area level identification half model are merged to construct an initial risk identification model; Using a mean square error function to perform loss assessment calculation on the initial risk identification model to obtain model loss data; The initial risk identification model is feedback optimized and updated based on the model loss data to generate the safety risk identification model.

6. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 1, characterized in that: The risk area hidden danger factor obtained includes: Performing weight allocation on each risk type in the risk area feature set to obtain risk type weight allocation information; Performing weighted calculation on each risk level in the risk area feature set based on the risk type weight allocation information to determine a risk area hidden danger parameter set; Determining a regional hidden danger enhancement impact coefficient set according to the real-time distance information set; The risk area hidden danger parameter set is supplemented and summed up based on the area hidden danger reinforcement influence coefficient set to obtain the risk area hidden danger factor.

7. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 1, characterized in that: The determination of the target security risk early warning strategy includes: Setting a high-risk hidden danger factor threshold, if the risk area hidden danger factor set does not reach the high-risk hidden danger factor threshold, triggering the personalized warning strategy set in the dynamic safety warning strategy library; Based on the personalized warning strategy set, the risk area hidden danger factors are dynamically matched to determine the personalized safety risk warning strategy.

8. The personalized safety risk early warning method with integrated rapid rescue call as claimed in claim 7, characterized in that: The determination of the target security risk early warning strategy includes: When the risk area hidden danger factor set reaches the high-risk hidden danger factor threshold, the quick rescue strategy in the dynamic safety warning strategy library is triggered, and the quick rescue strategy is output as the target safety risk warning strategy.

9. A personalized safety risk early warning system with integrated quick rescue call, characterized by: The system is used to execute the personalized safety risk early warning method with integrated rapid rescue call according to any one of claims 1 to 8, and the system comprises: A risk area identification module is used to collect and obtain an information platform data set using crawler technology, identify security risks on the information platform data set, and obtain a risk area feature set and a risk area location set; A real-time distance acquisition module is used to acquire the current location information of the target user and calculate a real-time distance information set between the risk area location set and the current location information; A multi-dimensional risk assessment module, used to perform multi-dimensional risk assessment on the risk area feature set and the real-time distance information set respectively to obtain a risk area hidden danger factor; A strategy library construction module is used to construct a dynamic safety warning strategy library, wherein the dynamic safety warning strategy library integrates a personalized warning strategy set and a quick rescue strategy; The security risk warning module is used to perform a strategy analysis on the hidden danger factors in the risk area based on the dynamic security warning strategy library, determine the target security risk warning strategy, and perform a dynamic security warning on the target user through the target security risk warning strategy.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the personalized safety risk early warning method with integrated rapid rescue call as described in any one of claims 1-8 are implemented.

Images