Intelligent early warning method and system for abnormal health of international traveler

Abstract

The invention relates to the technical field of medical diagnosis, and discloses an international traveler health abnormity intelligent early warning method and system, and the method comprises the steps: collecting the multi-source health data of an international traveler, and carrying out the standardization processing of the multi-source health data, and obtaining the structured health data of the international traveler; extracting basic features of the structured health data based on a health anomaly knowledge base of international travelers, and performing multi-dimensional fusion on the basic features to obtain multi-dimensional feature vectors of the basic features; based on a preset health risk threshold, performing risk level assessment on the multi-dimensional feature vector to obtain a preliminary risk identifier of the international traveler; acquiring travel track information of the international traveler, and performing spatio-temporal context fusion analysis on the initial risk identifier and the travel track information to obtain a comprehensive risk assessment result of the international traveler; generating an early warning signal of the international traveler according to the comprehensive risk assessment result; according to the invention, the efficiency of intelligent early warning of abnormal health of international travelers can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical diagnosis, in particular to an intelligent health anomaly early warning method and system for international travelers. BACKGROUND

[0002] In the field of health anomaly early warning for international travelers, the existing technology has significant deficiencies in processing multi-source health data. Due to the scattered sources and various formats of international traveler health data, the existing solutions lack effective standardization processing mechanisms, neither fully removing noise interference in the original data nor realizing format unification of data from different sources, resulting in a large amount of health data existing in unstructured form. Such unstructured data is difficult to be effectively parsed to extract key health features, making the subsequent risk assessment lose a reliable data basis, directly affecting the accuracy of health anomaly identification, and failing to provide effective data support for early warning.

[0003] At the same time, the existing health anomaly early warning technology has a single evaluation dimension, generally relying only on health data for risk judgment, ignoring the influence of the spatio-temporal information contained in the travel trajectory of international travelers on health risks. The existing solutions do not consider the environmental risk factors of the areas where different travel nodes are located, nor do they combine the spatio-temporal decay effect and regional risk transmission characteristics for dynamic analysis, resulting in deviations between the risk assessment results and the actual health risk situation. Such single-dimensional evaluation mode is difficult to fully reflect the real health risk state of travelers, and the early warning efficiency is low, which cannot timely and accurately identify potential health anomalies, and thus cannot meet the demand for rapid health risk early warning in the international travel scenario. Therefore, how to improve the efficiency of intelligent health anomaly early warning for international travelers has become a problem to be solved. SUMMARY

[0004] The present application provides an intelligent health anomaly early warning method and system for international travelers to solve the problems raised in the background art.

[0005] To achieve the above-mentioned purpose, the present application provides an intelligent health anomaly early warning method for international travelers, comprising:

[0006] S1, collecting multi-source health data of international travelers, and performing standardization processing on the multi-source health data to obtain structured health data of the international travelers;

[0007] S2, based on the health anomaly knowledge base of the international travelers, extracting the basic features of the structured health data, and performing multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features;

[0008] S3, based on a preset health risk threshold, performing risk level evaluation on the multi-dimensional feature vector to obtain a preliminary risk identification of the international travelers;

[0009] S4, obtaining the travel trajectory information of the international traveler, and performing spatio-temporal context fusion analysis on the preliminary risk identification and the travel trajectory information to obtain a comprehensive risk assessment result of the international traveler;

[0010] S5, generating a warning signal of the international traveler according to the comprehensive risk assessment result.

[0011] In a preferred embodiment, the multi-source health data of the international traveler is collected, and the multi-source health data is standardized to obtain the structured health data of the international traveler, including:

[0012] Collecting original health data in the international traveler health data source;

[0013] Noise cleaning of the original health data to obtain clean health data of the original health data;

[0014] Uniform format processing of the clean health data to obtain the structured health data of the international traveler.

[0015] In a preferred embodiment, the health abnormality knowledge base of the international traveler is used to extract the basic features of the structured health data, and the basic features are multi-dimensionally fused to obtain a multi-dimensional feature vector of the basic features, including:

[0016] Based on the feature mapping relationship of the health abnormality knowledge base, the structured health data is feature mapped to obtain the basic features of the structured health data;

[0017] Normalizing the basic features to obtain standardized basic features of the basic features;

[0018] Weighted fusion of the standardized basic features to obtain the fusion features of the basic features;

[0019] Dimension reconstruction of the fusion features to obtain a multi-dimensional feature vector of the basic features.

[0020] In a preferred embodiment, the weighted fusion processing of the standardized basic features to obtain the fusion features of the basic features includes:

[0021] Based on a preset weight allocation strategy, the influence degree of the standardized basic features is evaluated to obtain an initial fusion weight of the standardized basic features;

[0022] The relative dispersion degree between the standardized basic features is calculated, wherein the calculation formula of the relative dispersion degree is as follows:

[0023] ;

[0024] In the formula, The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the One characteristic, This represents the mean of all features in the standardized basic features. This represents the standard deviation of all features in the standardized basic features;

[0025] Based on the relative degree of dispersion, the initial fusion weights are adaptively corrected to obtain the fusion weights of the standardized basic features;

[0026] Based on the fusion weights, the standardized basic features are weighted and fused to obtain a preliminary fusion result of the standardized basic features;

[0027] The numerical fluctuations in the preliminary fusion results are eliminated to obtain the fusion features of the basic features.

[0028] In a preferred embodiment, the calculation formula for the preliminary fusion result is as follows:

[0029] ;

[0030] In the formula, This indicates the preliminary fusion result. This represents the total number of features in the standardized basic features. The first standardized basic feature represents the The fusion weights of each feature The first standardized basic feature represents the The fusion weights of each feature This represents the preset dispersion adjustment coefficient. The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the One characteristic.

[0031] In a preferred embodiment, the step of assessing the risk level of the multidimensional feature vector based on a preset health risk threshold to obtain a preliminary risk identifier for the international traveler includes:

[0032] comparing the multi-dimensional feature vector item by item based on a preset health risk threshold to obtain a single-item risk comparison result of the multi-dimensional feature vector;

[0033] analyzing the single-item risk comparison result to obtain single-item risk information of the multi-dimensional feature vector;

[0034] identifying an abnormal feature from the single-item risk information to obtain a key risk feature of the multi-dimensional feature vector;

[0035] comprehensively judging the key risk feature based on a risk level determination rule in the health risk threshold to obtain a preliminary risk level of the multi-dimensional feature vector;

[0036] performing symbol mapping on the preliminary risk level based on a preset risk identifier to obtain a preliminary risk identification of the international traveler.

[0037] In a preferred embodiment, the travel trajectory information of the international traveler is obtained, and the preliminary risk identification is analyzed in a spatio-temporal context with the travel trajectory information to obtain a comprehensive risk assessment result of the international traveler, including:

[0038] obtaining travel trajectory information of the international traveler and adding a time stamp to the travel trajectory information to obtain a travel node of the travel trajectory information;

[0039] performing relevance matching on the preliminary risk identification and the travel trajectory information to obtain an initial risk value of the travel trajectory information;

[0040] spatio-temporally aggregating environmental risk factors of a region where the travel node is located to obtain a regional risk coefficient of the travel node;

[0041] calculating a dynamic risk value of the travel node according to a preset spatio-temporal attenuation effect and regional risk transmission;

[0042] integrating and analyzing the dynamic risk value to obtain a comprehensive risk assessment result of the international traveler.

[0043] In a preferred embodiment, the calculation formula of the dynamic risk value is as follows:

[0044] ;

[0045] In the formula, denotes a dynamic risk value of the i-th travel node, denotes a dynamic risk value of the i-th travel node, denotes a dynamic risk value of the i-th travel node, denotes a dynamic risk value of the i-th travel node, denotes a natural constant, representing a preset time attenuation coefficient, representing a time interval of adjacent journey nodes in the journey nodes, representing a regional risk coefficient of the first journey node, representing a preset space conduction coefficient, representing a space correlation degree of adjacent journey nodes in the journey nodes.

[0046] In a preferred embodiment, the generating the early warning signal of the international traveler according to the comprehensive risk assessment result comprises:

[0047] comparing the comprehensive risk assessment result with preset early warning level threshold values in multiple levels to obtain a target early warning level of the comprehensive risk assessment result;

[0048] according to the target early warning level, calling a preset early warning content template to generate basic early warning information of the comprehensive risk assessment result;

[0049] combining key information of the comprehensive risk assessment result and the basic early warning information into complete early warning information of the comprehensive risk assessment result;

[0050] performing information encapsulation on the complete early warning information to obtain the early warning signal of the international traveler.

[0051] In order to solve the above problems, the present application further provides an intelligent early warning system for health abnormalities of international travelers, which comprises:

[0052] a data acquisition and preprocessing module for acquiring multi-source health data of international travelers and performing standardized processing on the multi-source health data to obtain structured health data of the international travelers;

[0053] a multi-dimensional feature fusion module for extracting basic features of the structured health data based on a preset health abnormality knowledge base and performing multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features;

[0054] a risk level evaluation module for performing risk level evaluation on the multi-dimensional feature vector based on a preset health risk threshold to obtain a preliminary risk identification of the international traveler;

[0055] a space-time fusion analysis module for obtaining journey trajectory information of the international traveler and performing space-time context fusion analysis on the preliminary risk identification and the journey trajectory information to obtain a comprehensive risk assessment result of the international traveler;

[0056] an early warning signal generation module for generating an early warning signal of the international traveler according to the comprehensive risk assessment result.

[0057] Compared with the prior art, the present application has the following beneficial effects:

[0058] 1. The present application obtains structured data by noise cleaning and format unification of multi-source health data of international travelers, eliminates the interference of unstructured data; then extracts basic features based on health anomaly knowledge base, forms a multi-dimensional feature vector through normalization, adaptive weighted fusion considering relative dispersion degree and dimension reconstruction, fully integrates key health features, avoids single feature deviation, provides reliable basis for risk assessment, and greatly improves the accuracy of preliminary risk identification.

[0059] 2. The present application adds time stamp to form travel nodes by combining travel trajectory information, integrates preliminary risk identification and regional risk coefficient, calculates dynamic risk value according to time-space attenuation effect and spatial conductivity, fully considers the influence of time-space dimension on health risk, avoids the limitation of single health data evaluation, is more suitable for the real risk state of travelers, and then generates accurate early warning signal, effectively improves the efficiency of health anomaly early warning. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 A flowchart of an international traveler health anomaly intelligent early warning method provided by an embodiment of the present application is shown in the figure;

[0061] Figure 2 A functional module diagram of an international traveler health anomaly intelligent early warning system provided by an embodiment of the present application is shown in the figure;

[0062] The implementation, functional characteristics and advantages of the present application will be further described with reference to the embodiments and the accompanying drawings. DETAILED DESCRIPTION

[0063] It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0064] The embodiment of the application provides an international traveler health anomaly intelligent early warning method. The execution subject of the international traveler health anomaly intelligent early warning method includes but is not limited to at least one of electronic devices such as a server, a terminal and the like which can be configured to execute the method provided by the embodiment of the application. In other words, the international traveler health anomaly intelligent early warning method can be executed by software or hardware installed in a terminal device or a server device. The server includes but is not limited to a single server, a server cluster, a cloud server or a cloud server cluster and the like. The server can be a stand-alone server, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content distribution networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms.

[0065] Referring to Figure 1 Fig. 1 is a flowchart of an international traveler health anomaly intelligent early warning method provided by an embodiment of the application. In the embodiment, the international traveler health anomaly intelligent early warning method includes the following steps.

[0066] S1, collecting multi-source health data of an international traveler, and performing standardized processing on the multi-source health data to obtain structured health data of the international traveler;

[0067] In the embodiment of the application, the collecting multi-source health data of an international traveler, and performing standardized processing on the multi-source health data to obtain structured health data of the international traveler includes the following steps.

[0068] Collecting original health data in an international traveler health data source;

[0069] Performing noise cleaning on the original health data to obtain clean health data of the original health data;

[0070] Performing format unification processing on the clean health data to obtain the structured health data of the international traveler.

[0071] Specifically, the whole process is to collect original health data in an international traveler health data source, perform noise cleaning on the original health data to obtain clean health data, and then perform format unification processing on the clean health data to obtain structured health data.

[0072] Further, when collecting the original health data, the health monitoring information system associated with the health declaration system, the health record system of the medical institution and the traffic travel platform negotiates the data docking permission and the security mechanism in advance, submits a data use application and explains the purpose as international traveler health risk assessment.

[0073] Further, after being reviewed and passed by each system operator, a stable data docking channel is established using the SSL encryption transmission protocol. After the channel is built, three data transmission tests are conducted to ensure that the data received is complete and has no transmission errors, and to avoid data being illegally obtained or tampered with during transmission.

[0074] Further, the unprocessed information containing basic identity information, physical symptom records, body temperature measurement values, vaccination history, and medical history, etc. is extracted within the preset range. During the extraction process, it is verified in real time whether each piece of information belongs to the preset range.

[0075] Further, if irrelevant information outside the range is found, it is immediately filtered. Information items missing within the preset range are marked “to be supplemented”. After subsequent confirmation with the system docking party whether there is any omission, the information meeting the requirements is retrieved from the databases of various systems and summarized to form the original health data.

[0076] Further, during noise cleaning, each record in the original health data is checked row by row to identify duplicate data, missing data, and abnormal data in detail.

[0077] Further, for duplicate data, the core identification fields are compared to determine whether the data is duplicated. The core identification fields include the traveler's unique identity information and the data collection time. If the two fields are completely identical, it is determined that the data is duplicated. When deleting the redundant duplicate entries, the record with more complete data fields is retained first.

[0078] Further, for data missing key information fields, the data provider is immediately contacted to explain the specific missing fields and record numbers, and to request verification and supplement of accurate information.

[0079] Further, for non-key information fields missing, the status of “information not provided” is marked at the corresponding field position. The marking time is recorded simultaneously to facilitate subsequent tracing.

[0080] Further, for abnormal data, the source party is contacted immediately to provide specific screenshots of the abnormal data and data source paths to help the source party quickly locate the problem data. If the source party cannot provide a reasonable explanation or correction basis, and no response is received within 48 hours, it is determined that the abnormal record cannot be confirmed and is directly excluded. After these operations, the information obtained is the clean health data.

[0081] Further, during format unification processing, the format differences of different source information in the clean health data are comprehensively analyzed, and the format conversion rules are uniformly planned. The date is strictly adjusted to the standard format of four-digit year-two-digit month-two-digit date.

[0082] Further, before converting the numerical units, the original unit label of each source data is confirmed, if the original unit is not labeled, contact the data provider to verify, avoid conversion error caused by unit misjudgment, and then all numerical units are converted into international standard units.

[0083] Further, when replacing the text description, the various types of text description are replaced with standard terms by comparing with the preset standard glossary, and for special expressions not covered in the standard glossary, the special expressions are first recorded and fed back to the glossary maintenance team, and a "to be standardized" label is temporarily used, and after updating the glossary, the special expressions are supplemented and corrected.

[0084] Further, the data is arranged according to the preset structured data table structure, and the preset structured data table fields include traveler identity information field, health index field and medical history record field, and each field has a clear filling specification.

[0085] Further, the data items corresponding to each record are checked field by field during the arrangement, and the data format is checked according to the specification to check whether the data format meets the requirements, if there is a format inconsistency, the format unification link is returned for reprocessing, until all data items are accurately filled in the fixed fields of the data table, and the information obtained after the processing is completed is the structured health data.

[0086] In general, the original health data in the international traveler health data source is collected by establishing a docking channel with the related system, and extracting the unprocessed information containing specific content according to the preset range.

[0087] In general, the clean health data obtained by noise cleaning of the original health data is obtained by checking and identifying repeated, missing and abnormal data, and the information obtained by deleting, supplementing, marking or rejecting is processed.

[0088] In general, the structured health data obtained by format unification processing of the clean health data is obtained by processing the format difference of information from different sources and arranging according to the preset table structure.

[0089] S2, based on the health abnormality knowledge base of the international traveler, extracting the basic features of the structured health data, and performing multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features;

[0090] In the embodiment of the application, based on the health abnormality knowledge base of the international traveler, extracting the basic features of the structured health data, and performing multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features, including:

[0091] Based on the feature mapping relationship of the health abnormality knowledge base, the structured health data is feature mapped to obtain the basic features of the structured health data;

[0092] normalizing the basic features to obtain standardized basic features of the basic features;

[0093] performing weighted fusion on the standardized basic features to obtain fusion features of the basic features;

[0094] performing dimension reconstruction on the fusion features to obtain a multi-dimensional feature vector of the basic features.

[0095] The performing weighted fusion on the standardized basic features to obtain fusion features of the basic features comprises:

[0096] based on a preset weight distribution strategy, performing influence degree evaluation on the standardized basic features to obtain initial fusion weights of the standardized basic features;

[0097] calculating relative dispersion degrees between the standardized basic features, wherein a calculation formula of the relative dispersion degrees is as follows:

[0098] ;

[0099] wherein, denotes a relative dispersion degree of an i-th feature of the standardized basic features, denotes an i-th feature of the standardized basic features, denotes a mean value of all features in the standardized basic features, denotes a standard deviation of all features in the standardized basic features; based on the relative dispersion degrees, adaptively correcting the initial fusion weights to obtain fusion weights of the standardized basic features;

[0100] based on the fusion weights, performing weighted fusion on the standardized basic features to obtain a preliminary fusion result of the standardized basic features;

[0101] eliminating numerical fluctuations in the preliminary fusion result to obtain fusion features of the basic features.

[0102] The calculation formula of the preliminary fusion result is as follows:

[0103]

[0104] ;

[0105] wherein, denotes the preliminary fusion result, denotes a total number of features in the standardized basic features, denotes an i-th feature of the standardized basic features, ​​a fusion weight of the feature, representing the standardized basic feature a fusion weight of the feature, representing a preset discrete degree adjustment coefficient, representing the standardized basic feature a relative discrete degree of the feature, representing the standardized basic feature a relative discrete degree of the feature, representing the standardized basic feature the feature.

[0106] Specifically, the whole process is to map the structured health data to the basic features based on the feature mapping relationship of the health abnormality knowledge base.

[0107] Specifically, the basic features are normalized to obtain the standardized basic features, the standardized basic features are fused to obtain the fused features, and finally the fused features are dimensionally reconstructed to obtain the multi-dimensional feature vector of the basic features.

[0108] Further, when the feature mapping is performed based on the feature mapping relationship of the health abnormality knowledge base, a one-to-one mapping rule between each field of the structured health data and the basic features has been established in advance in the health abnormality knowledge base.

[0109] Further, when operating, each field and its corresponding specific data in the structured health data are extracted one by one, and then each field is compared with the mapping rule in the knowledge base one by one to determine the basic feature name matched by the field, and the specific data of the field is taken as the attribute value of the matched basic feature.

[0110] Further, the basic feature names and attribute values corresponding to all fields are sorted into an ordered feature set, and this set is the basic features of the structured health data.

[0111] Further, when the basic features are normalized, the historical value range of each basic feature is first retrieved from the health abnormality knowledge base.

[0112] Further, for each basic feature, the difference between its current attribute value and the minimum value of the historical value range is first calculated, and then this difference is divided by the difference between the maximum value and the minimum value of the historical value range to obtain a value between zero and one.

[0113] Further, the current attribute value of each basic feature is replaced by the value between zero and one, and the set formed by integrating all the basic features with normalized values is the standardized basic features of the basic features.

[0114] Further, when the standardized basic features are fused by weighting, the health anomaly knowledge base has set a fixed weight for each standardized basic feature according to the importance of the health anomaly risk assessment of each basic feature.

[0115] Further, during operation, the normalized values of each standardized basic feature are extracted one by one, and then each value is multiplied by the corresponding fixed weight to obtain the weighted calculation result of each standardized basic feature.

[0116] Further, the weighted calculation results of all standardized basic features are added to obtain a comprehensive quantitative value, which is the fusion feature of the basic feature.

[0117] Further, when the fusion feature is dimensionally reconstructed, the dimension division standard of the multi-dimensional feature vector is pre-set, and the health risk assessment dimension is usually divided into four fixed dimensions of "physiological indicator dimension", "symptom performance dimension", "immune state dimension" and "basic health dimension", each dimension corresponds to a specific type of feature contribution value in the fusion feature.

[0118] Further, the sub-feature values related to each dimension are extracted from the comprehensive quantitative value of the fusion feature.

[0119] Further, according to the fixed order of "physiological indicator dimension-symptom performance dimension-immune state dimension-basic health dimension", the sub-feature values of the four dimensions are arranged in sequence to form an ordered sub-feature value sequence, which is a multi-dimensional feature vector of the basic feature.

[0120] Specifically, the whole process is to determine the initial fusion weight of the standardized basic feature based on the pre-set weight allocation strategy, and calculate the relative dispersion degree of these features.

[0121] Specifically, the initial fusion weight is corrected based on the relative dispersion degree to obtain the fusion weight, and then the preliminary fusion result is calculated by combining the formula, and finally the numerical fluctuation of the preliminary fusion result is eliminated to obtain the fusion feature of the basic feature.

[0122] Further, when determining the initial fusion weight based on the pre-set weight allocation strategy, the strategy is pre-set by health experts according to the importance of different standardized basic features such as body temperature feature, respiratory symptom feature, immune protection feature, etc. The initial weight value of each standardized basic feature is pre-set.

[0123] Further, for each standardized basic feature, the influence degree of the feature on the subsequent fusion result is evaluated according to the initial weight rule of the corresponding feature in the strategy, so as to determine the initial weight value of each feature, and these values together constitute the initial fusion weight of the standardized basic feature.

[0124] Furthermore, when calculating the relative dispersion among standardized basic features, the values ​​of all standardized basic features are first collected, and these values ​​are added together and then divided by the total number of features to obtain the mean of all features.

[0125] Furthermore, calculate the square of the difference between each value and the mean, add all the squared values ​​together, divide by the total number of features, and then take the square root of the result to obtain the standard deviation of all features.

[0126] Furthermore, for the first The first standardized basic feature is obtained by dividing the absolute difference between the value of that feature and the mean by the standard deviation. The relative dispersion of each feature.

[0127] Furthermore, when obtaining the fusion weights by correcting the initial fusion weights based on the relative dispersion, the relative dispersion of each standardized basic feature is observed. If the relative dispersion of a certain feature is large, it indicates that the feature has stronger differences in the whole and has a more significant role in distinguishing the fusion results. Therefore, the initial fusion weight corresponding to the feature is appropriately increased according to the preset ratio.

[0128] Furthermore, if the relative dispersion of a certain feature is small, it indicates that its difference is weak and its distinguishing effect on the fusion result is not obvious. Therefore, the corresponding initial fusion weight is appropriately reduced according to the preset ratio. Through such adjustment, the fusion weight of each standardized basic feature after correction is obtained.

[0129] Furthermore, when calculating the preliminary fusion result based on the fusion weights and formulas, it is necessary to first calculate the fusion result for each standardized basic feature. Item, of which It is the fusion weight of this feature. It is a fixed dispersion adjustment coefficient pre-set by health experts based on historical fusion effects and expected impacts on dispersion. It represents the relative dispersion of this feature.

[0130] Furthermore, calculate the sum of all features. The sum of the terms, here It is the first The fusion weights of each feature It is the first The relative dispersion of each feature; then each feature's Divide by this sum to get the percentage coefficient.

[0131] Furthermore, the proportion coefficient of each feature is compared with its own value. Multiply and sum. The sum of the corresponding values ​​extracted from the normalized basic features is the preliminary fusion result of the normalized basic features. wherein the parameters are obtained by statistically participating in the weighted fusion of the standardized basic features.

[0132] Further, and are the fusion weights determined after the pre-set weight strategy and the relative dispersion degree are corrected; is a fixed dispersion degree adjustment coefficient pre-set by an expert; and are the relative dispersion degrees obtained by calculating the mean value, the standard deviation, and then dividing the absolute difference between the feature value and the mean value by the standard deviation; is the value extracted from the normalized standardized basic features.

[0133] Further, the significance of this formula lies in comprehensively considering the fusion weight and the relative dispersion degree of each feature, so that the feature with a large relative dispersion degree can obtain a more reasonable weight proportion in the fusion result, thereby more accurately integrating the information of each standardized basic feature.

[0134] Further, from the trend, when the relative dispersion degree of a certain feature increases, and the contribution of the feature to increases; when increases, the influence of the relative dispersion degree on will be more intense; when the fusion weight increases, the contribution of the feature to also increases.

[0135] Further, when eliminating the numerical fluctuations in the preliminary fusion result to obtain the fusion features of the basic features, a sliding average method is adopted, a fixed length window is set, and the preliminary fusion result is arranged in order.

[0136] Further, the average values are calculated by sequentially selecting the values in the window, and the original values in the window are replaced with these average values, so that the numerical changes are more smooth, thereby eliminating the fluctuations in the preliminary fusion result. The values obtained after such processing are the fusion features of the basic features.

[0137] In general, the feature mapping relationship based on the health abnormal knowledge base is used to map the structured health data to obtain the basic features, which is a process of determining the corresponding basic features and attribute values of each field by comparing the mapping rules of the knowledge base and arranging them into a set.

[0138] In general, the normalization processing of the basic features to obtain the standardized basic features is a process of retrieving the historical value range, calculating the normalized values, and replacing the original attribute values to form a set.

[0139] In general, the weighting fusion of the standardized basic features to obtain the fusion features is a process of multiplying the normalized values of each feature by a fixed weight and then summing to obtain a comprehensive quantitative value.

[0140] In general, the dimension reconstruction of the fusion features to obtain the multi-dimensional feature vector of the basic features is a process of disassembling the sub-feature values according to the preset dimensions and arranging them in a fixed order to form a sequence.

[0141] In general, the initial fusion weight of the standardized basic features based on the preset weight distribution strategy is a process of presetting an initial weight value by an expert according to the importance of the feature to the health risk and determining the initial weight value.

[0142] In general, the relative dispersion degree between the standardized basic features is calculated by calculating the mean, standard deviation, and then dividing the absolute difference between the feature value and the mean by the standard deviation.

[0143] In general, the fusion weight is obtained by correcting the initial fusion weight based on the relative dispersion degree, which is a process of adjusting the initial weight according to the size of the relative dispersion degree of the feature according to a preset proportion.

[0144] In general, the preliminary fusion result is calculated based on the fusion weight and the formula, which is a process of calculating the proportion coefficient of each item and multiplying it by the feature value to sum up, while reflecting the meaning and trend of the formula.

[0145] In general, the fusion feature of the basic feature is obtained by eliminating the numerical fluctuations in the preliminary fusion result, which is a process of replacing the original value with a moving average to smooth the change.

[0146] S3, based on a preset health risk threshold, risk level assessment is performed on the multi-dimensional feature vector to obtain a preliminary risk identification of the international traveler;

[0147] In the embodiments of the present application, the preliminary risk identification of the international traveler is obtained by performing risk level assessment on the multi-dimensional feature vector based on a preset health risk threshold, which includes:

[0148] Based on the preset health risk threshold, the multi-dimensional feature vector is compared item by item to obtain a single risk comparison result of the multi-dimensional feature vector;

[0149] Analyzing the single risk comparison result to obtain single risk information of the multi-dimensional feature vector;

[0150] Identifying abnormal features from the single risk information to obtain key risk features of the multi-dimensional feature vector;

[0151] comprehensively analyze the key risk features based on a risk level determination rule in the health risk threshold to obtain a preliminary risk level of the multi-dimensional feature vector;

[0152] symbolically map the preliminary risk level based on a preset risk identifier to obtain a preliminary risk identification of the international traveler.

[0153] Specifically, the entire process is to compare the multi-dimensional feature vector item by item based on the preset health risk threshold to obtain a single-item risk comparison result, analyze the single-item risk comparison result to obtain single-item risk information, and then identify abnormal features from the single-item risk information to obtain key risk features.

[0154] Specifically, the preliminary risk level is obtained by comprehensively analyzing the key risk features based on a risk level determination rule in the health risk threshold, and finally the preliminary risk identification of the international traveler is obtained by symbolically mapping the preliminary risk level based on a preset risk identifier.

[0155] Further, when comparing the multi-dimensional feature vector item by item based on the preset health risk threshold, the preset health risk threshold includes a standard value corresponding to each feature item in the multi-dimensional feature vector.

[0156] Further, the numerical value of each feature item is extracted from the multi-dimensional feature vector one by one, and then the numerical value of each feature item is compared with the corresponding standard value in the health risk threshold, and then it is recorded whether the numerical value of each feature item reaches the standard value, exceeds the standard value or does not reach the standard value.

[0157] Further, after all the feature items are compared, these recorded contents constitute the single-item risk comparison result of the multi-dimensional feature vector.

[0158] Further, when analyzing the single-item risk comparison result, for each record in the single-item risk comparison result, first identify the feature item name corresponding to the record, and then describe in detail the comparison between the numerical value of the feature item and the corresponding standard value, that is, specifically describe whether the standard value is reached, exceeded or not reached.

[0159] Further, these information is sorted according to each feature item, and the sorted content containing the feature item name and the corresponding comparison condition is the single-item risk information of the multi-dimensional feature vector.

[0160] Further, when identifying abnormal features from the single-item risk information, first filter out the feature items with comparison conditions of reaching the standard value or exceeding the standard value from the single-item risk information.

[0161] Further, special markers are added to these screened-out feature items to distinguish them from other non-compliant feature items in the single-item risk information. After adding the markers, these feature items are the key risk features of the multi-dimensional feature vector.

[0162] Further, when comprehensively analyzing the key risk features based on the risk level determination rules in the health risk threshold, the risk level determination rules in the health risk threshold include both the correspondence between the number of key risk features and different risk levels and the correspondence between the severity of key risk features and different risk levels. The total number of key risk features is first counted, and then the severity of each key risk feature is evaluated, that is, the magnitude of the value of each key risk feature exceeding the corresponding standard value is determined.

[0163] Further, by comparing the total number of key risk features and the severity of each key risk feature with the risk level determination rules in the health risk threshold, the matching risk level is determined. This determined risk level is the preliminary risk level of the multi-dimensional feature vector.

[0164] Further, when symbol mapping the preliminary risk level based on the preset risk identifier, the preset risk identifier is a specific set of symbols. Each symbol in the set forms a unique correspondence with a different preliminary risk level.

[0165] Further, the preliminary risk level of the multi-dimensional feature vector obtained before is confirmed to be which one, and the symbol uniquely corresponding to the preliminary risk level is found from the preset risk identifier. This found symbol is the preliminary risk identifier of the international traveler.

[0166] In summary, the process of obtaining the single-item risk comparison result by comparing the multi-dimensional feature vector with the preset health risk threshold is to compare the feature item values with the corresponding standard values and record the results.

[0167] In summary, the process of obtaining the single-item risk information by analyzing the single-item risk comparison result is to determine the corresponding feature item names and comparison conditions of each record and organize the information.

[0168] In summary, the process of identifying key risk features from single-item risk information is to screen out compliant or over-standard feature items and add marker symbols to form features.

[0169] In summary, the process of obtaining the preliminary risk level by comprehensively analyzing the key risk features based on the risk level determination rules in the health risk threshold is to count the number, evaluate the severity, and determine the level by comparing the rules.

[0170] In general, the preliminary risk level is symbolically mapped based on the preset risk identifier to obtain a preliminary risk identification, which is a process of finding a unique corresponding symbol of the preliminary risk level to form an identification.

[0171] S4, obtaining the travel trajectory information of the international traveler, and performing spatio-temporal context fusion analysis on the preliminary risk identification and the travel trajectory information to obtain a comprehensive risk assessment result of the international traveler;

[0172] In the embodiment of the application, the travel trajectory information of the international traveler is obtained, and spatio-temporal context fusion analysis is performed on the preliminary risk identification and the travel trajectory information to obtain a comprehensive risk assessment result of the international traveler, including:

[0173] The travel trajectory information of the international traveler is obtained, and a timestamp is added to the travel trajectory information to obtain a travel node of the travel trajectory information.

[0174] The preliminary risk identification and the travel trajectory information are associated and matched to obtain an initial risk value of the travel trajectory information.

[0175] The environmental risk factors of the region where the travel node is located are spatio-temporally aggregated to obtain a regional risk coefficient of the travel node.

[0176] According to the preset spatio-temporal attenuation effect and regional risk transmission, a dynamic risk value of the travel node is calculated.

[0177] The dynamic risk value is integrated and analyzed to obtain a comprehensive risk assessment result of the international traveler.

[0178] The calculation formula of the dynamic risk value is as follows:

[0179] ;

[0180] In the formula, denotes the dynamic risk value of the i-th travel node, denotes the dynamic risk value of the i-th travel node, denotes the dynamic risk value of the i-th travel node, denotes a natural constant, denotes a preset time attenuation coefficient, denotes a time interval between adjacent travel nodes in the travel node, denotes the regional risk coefficient of the i-th travel node, denotes a preset space transmission coefficient, denotes a space correlation degree between adjacent travel nodes in the travel node. ​​

[0181] Specifically, the whole process is to obtain international traveler itinerary information and generate itinerary nodes, associate the preliminary risk identification with the itinerary information to obtain an initial risk value, and perform spatio-temporal aggregation on the regional environmental risk factors of the itinerary nodes to obtain a regional risk coefficient.

[0182] Specifically, the dynamic risk value of the itinerary node is calculated by combining the spatio-temporal decay effect and the regional risk transmission, and the related formula is explained, and finally the comprehensive risk assessment result of the international traveler is obtained by integrating and analyzing the dynamic risk value.

[0183] Further, when obtaining the itinerary information and generating the itinerary nodes, data docking is established with the flight operation platform, the high-speed rail passenger transport platform, the long-distance passenger transport platform, and the hotel registration system to extract the itinerary-related information such as the departure location, the arrival location, the transportation mode type, the transportation class number, the hotel address, and the stay duration at each location within the travel period of the traveler.

[0184] Further, these information is integrated to form complete itinerary information; then, for each key trajectory point in the itinerary information, such as the departure airport or station, the transfer airport or station, the arrival airport or station, and the hotel address, specific time information accurate to the year, month, day, hour, and minute is added, each key trajectory point with specific time information is defined as an independent data unit, and the set formed by all independent data units constitutes the itinerary node of the itinerary information.

[0185] Further, when associating the preliminary risk identification with the itinerary information to obtain an initial risk value, the three health risk levels of "high risk", "medium risk", and "low risk" preset by the preliminary risk identification are first determined.

[0186] Further, all the passing locations such as cities, regions, and specific stay places in the itinerary information are sorted one by one, and through the risk region division database of the disease control center, it is confirmed whether these passing locations belong to the risk associated region matching the risk level of the preliminary risk identification.

[0187] Further, according to the matching result of the risk level of the preliminary risk identification and the risk association of the passing location, a fixed value is assigned, if the preliminary risk identification is high risk and the passing location is a high-risk associated region, the initial risk value is assigned a fixed high value, if the preliminary risk identification is medium risk and the passing location is a medium-risk associated region, the initial risk value is assigned a fixed medium value, if the preliminary risk identification is low risk or the passing location has no corresponding risk association, the initial risk value is assigned a fixed low value, and the fixed value is the initial risk value of the itinerary information.

[0188] Further, when the environmental risk factors of the area where the travel node is located are aggregated in space-time to obtain the area risk coefficient, the specific type of the environmental risk factors of the area where the travel node is located is first determined, covering the recent incidence rate of infectious diseases, temperature, humidity, air quality index related to high incidence of respiratory diseases, and drinking water safety sampling results, food safety qualified rate and other factors directly related to health risks in the area.

[0189] Further, by connecting the data platform of the CDC in the area, the real-time monitoring system of the meteorological department, and the public information platform of the market supervision department sampling information, the corresponding environmental risk factor data of each travel node during the stay time period is obtained.

[0190] Further, the obtained environmental risk factor data is summarized according to the stay time dimension, and the average value of each factor in the stay time period is calculated.

[0191] Further, the average values of various factors in the same area are comprehensively evaluated according to the spatial dimension, and the comprehensive results are converted into specific quantitative values according to the importance of each factor on health risks. The quantitative value is the area risk coefficient of the travel node.

[0192] Further, when calculating the dynamic risk value of the travel node in combination with the preset space-time decay effect and the area risk transmission, the preset space-time decay effect is that the influence of the risk of the previous travel node on the current travel node will decrease by a fixed proportion as the time interval increases. The preset area risk transmission is that the health risks of adjacent areas are transmitted to each other.

[0193] Further, if the area risk coefficient of the area where the previous node is located is high, and the geographical distance between the area where the current node is located and the previous area is within the preset close distance range, the risk of the previous area will be transmitted to the current area by a fixed proportion.

[0194] Further, when calculating the dynamic risk value of the first travel node, the initial risk value is taken as the basis, and the area risk coefficient of the node is combined to obtain the basic risk value.

[0195] Further, when calculating the dynamic risk value of the subsequent travel node, the dynamic risk value of the previous node calculated in the previous stage is taken as the basis, and the time stamps of the and travel nodes are extracted to calculate the time interval between adjacent travel nodes.

[0196] Further, in combination with the fixed time decay coefficient obtained from the statistical risk decay rule in the historical travel data and the mathematical inherent fixed value natural constant, the value of the risk of the previous node after time decay is obtained.

[0197] Furthermore, the attenuated value is added to the regional risk coefficient of the current node.

[0198] Furthermore, based on historical data on risk transmission cases in adjacent areas, a fixed spatial transmission coefficient is determined after analyzing the relationship between spatial distance and transmission intensity. This coefficient is then obtained through a geographic information system. and the Spatial factors such as geographical distance and transportation connectivity of the regions where each travel node is located are used to assign spatial correlation to adjacent travel nodes, and the calculation includes the spatial transmission coefficient and the spatial correlation.

[0199] Furthermore, multiply the sum of the previous steps by this part to obtain the dynamic risk value of the current travel node.

[0200] Furthermore, from the perspective of the formula, this calculation can integrate the impact of time decay and spatial transmission on risk, and the final dynamic risk value can comprehensively reflect the dynamic risk level of the current node under the combined effect of time and space factors.

[0201] Furthermore, judging from the trend of the formula, as the time interval between adjacent travel nodes increases, the dynamic risk value tends to decrease, reflecting that the impact of the risk of the previous node on the current node weakens over time.

[0202] Furthermore, as the regional risk coefficient of the current node increases, or the spatial correlation between adjacent travel nodes increases, the dynamic risk value will tend to increase, reflecting that the higher the environmental risk of the current node itself and the higher the spatial correlation between adjacent regions, the higher the dynamic risk value.

[0203] Furthermore, when integrating and analyzing dynamic risk values ​​to obtain a comprehensive risk assessment result for international travelers, dynamic risk values ​​of all itinerary nodes are first collected and arranged in chronological order of the itinerary occurrence to form a dynamic risk value sequence.

[0204] Furthermore, the number of nodes in the sequence that fall within the preset high-risk, medium-risk, and low-risk value ranges is counted, and the proportion of nodes at each risk level to the total number of nodes is calculated.

[0205] Furthermore, we can simultaneously observe the changing trend of dynamic risk values ​​to determine whether there are situations where the risk values ​​of three or more consecutive nodes increase, two or more consecutive nodes maintain high risk values, or the risk value of a certain node suddenly exceeds the high risk range.

[0206] Further, the proportion of the number of node risk levels and the risk change trend are comprehensively considered, if the proportion of high-risk nodes exceeds fifty percent or there is a continuous high-risk node trend, it is determined that the comprehensive risk is high, if the proportion of medium-risk nodes exceeds sixty percent and there is no continuous high-risk node, it is determined that the comprehensive risk is medium, if the proportion of low-risk nodes exceeds eighty percent and there is no high-risk node, it is determined that the comprehensive risk is low.

[0207] Further, the journey information of the specific high-risk node and the risk value are combined to form a text description, and the content including the risk level and the text description is the comprehensive risk assessment result of the international traveler.

[0208] Overall, obtaining the journey trajectory information of the international traveler and generating the journey node are a process of extracting information through multi-platform data docking and adding timestamps to the key trajectory points to form a collection of independent data units.

[0209] Overall, the correlation matching of the preliminary risk identification and the journey trajectory information obtains the initial risk value, which is a process of determining the risk degree, confirming the risk correlation of the passing place, and assigning a fixed value.

[0210] Overall, the spatio-temporal aggregation of the environmental risk factors of the region where the journey node is located obtains the regional risk coefficient, which is a process of determining the factor type, obtaining data, and converting to quantitative values according to the spatio-temporal dimension.

[0211] Overall, the calculation of the dynamic risk value of the journey node combined with the spatio-temporal decay effect and the regional risk transmission and the explanation of the related formula are a process of calculating the dynamic risk value according to the preset effect and transmission combined with each parameter and explaining the significance and trend of the formula.

[0212] Overall, the integration analysis of the dynamic risk value obtains the comprehensive risk assessment result of the international traveler, which is a process of collecting dynamic risk values, calculating risk level proportions, observing change trends, and forming a text description of the comprehensive risk level.

[0213] S5, according to the comprehensive risk assessment result, a warning signal of the international traveler is generated.

[0214] In the embodiment of the application, according to the comprehensive risk assessment result, the warning signal of the international traveler is generated, which includes:

[0215] The comprehensive risk assessment result is compared with a preset warning level threshold in multiple levels to obtain a target warning level of the comprehensive risk assessment result;

[0216] According to the target warning level, a preset warning content template is called to generate basic warning information of the comprehensive risk assessment result;

[0217] combining the key information of the comprehensive risk assessment result and the basic early warning information into complete early warning information of the comprehensive risk assessment result;

[0218] performing information encapsulation on the complete early warning information to obtain the early warning signal of the international traveler.

[0219] Specifically, the entire process is to perform multi-level comparison between the comprehensive risk assessment result and preset early warning level thresholds to obtain a target early warning level of the comprehensive risk assessment result.

[0220] Specifically, according to the target early warning level, a preset early warning content template is called to generate basic early warning information of the comprehensive risk assessment result, the key information of the comprehensive risk assessment result is combined with the basic early warning information to form complete early warning information of the comprehensive risk assessment result, and finally, information encapsulation is performed on the complete early warning information to obtain the early warning signal of the international traveler.

[0221] Further, when the comprehensive risk assessment result is compared with the preset early warning level thresholds, the preset early warning level thresholds have been divided into clear multi-level ranges in advance, and each range corresponds to a unique early warning level name.

[0222] Further, a specific value representing the overall risk degree is obtained from the comprehensive risk assessment result, then comparison is started from the threshold range of the highest level of early warning level, it is judged whether the specific value falls within the threshold range of the current comparison, if the value is in the range, it is determined that the early warning level corresponding to the range is the target early warning level of the comprehensive risk assessment result, if the value is not in the range, comparison is continued with the threshold range of the next level of early warning level.

[0223] Further, the comparison action is repeated until the threshold range to which the value belongs is found, and the corresponding early warning level is determined, and the determined early warning level is the target early warning level of the comprehensive risk assessment result.

[0224] Further, when the target early warning level is called to call the preset early warning content template, the preset early warning content template is set according to different early warning levels, each early warning level corresponds to a unique early warning content template, and the template contains fixed early warning prompt content matched with the risk level.

[0225] Further, according to the determined target early warning level, the corresponding early warning content template of the target early warning level is searched in the preset early warning content template library, after the corresponding template is found, all early warning prompt contents in the template are directly extracted, and these extracted early warning prompt contents are the basic early warning information of the comprehensive risk assessment result.

[0226] Further, when combining the key information of the comprehensive risk assessment result and the basic warning information, the key information is extracted from the comprehensive risk assessment result first, which includes the travel node risk situation of the international traveler, the main health risk associated factors, the risk duration estimation and other core contents.

[0227] Further, according to the preset combination order, the extracted key information is sequentially added to the specified position of the basic warning information, wherein the key information is placed in the front part of the basic warning information for clearly presenting the source and specific situation of the risk, and the core prompt content of the basic warning information is placed in the rear part for clearly presenting the risk prompt and response suggestion, and the complete information formed by the combination of the two in this way is the complete warning information of the comprehensive risk assessment result.

[0228] Further, when information packaging is performed on the complete warning information, a fixed structure of information packaging is first determined, which includes a warning signal identifier, a complete warning information main body and an information generation time, wherein the warning signal identifier is a special mark for distinguishing other types of signals, and the information generation time is the specific time when the packaging operation is performed.

[0229] Further, the complete warning information formed before is filled into the complete warning information main body part in the packaging structure, and the corresponding warning signal identifier and information generation time are supplemented, so that the three parts of content are orderly integrated together according to the fixed structure to form a data package with complete structure and comprehensive information, which is the warning signal of the international traveler.

[0230] In general, the target warning level of the comprehensive risk assessment result is obtained by comparing the comprehensive risk assessment result with the preset warning level threshold in multiple levels, which is a process of obtaining the specific value of the assessment result and comparing the threshold range from high level to low level to determine the level.

[0231] In general, the basic warning information of the comprehensive risk assessment result is generated by calling the preset warning content template according to the target warning level, which is a process of finding the corresponding template and extracting the fixed warning prompt content in the template.

[0232] In general, the key information of the comprehensive risk assessment result and the basic warning information are combined into the complete warning information of the comprehensive risk assessment result, which is a process of extracting the key information and integrating it with the basic warning information according to the preset order.

[0233] In general, the complete warning information is information packaged to obtain the warning signal of the international traveler, which is a process of filling information according to the fixed structure and integrating to form a data package.

[0234] As shown in Figure 2 , it is a functional module diagram of an international traveler health abnormality intelligent warning system provided by an embodiment of the application.

[0235] The international traveler health anomaly intelligent early warning system 100 can be installed in an electronic device. According to the functions implemented, the international traveler health anomaly intelligent early warning system 100 can include a data acquisition and preprocessing module 101, a multi-dimensional feature fusion module 102, a risk level evaluation module 103, a space-time fusion analysis module 104, and an early warning signal generation module 105. The modules of the present application can also be referred to as units, which refer to a series of computer program segments that can be executed by an electronic device processor and can complete a fixed function, which are stored in the memory of the electronic device.

[0236] In the present embodiment, the functions of each module / unit are as follows:

[0237] The data acquisition and preprocessing module 101 is configured to acquire multi-source health data of an international traveler, and to perform standardized processing on the multi-source health data to obtain structured health data of the international traveler.

[0238] The multi-dimensional feature fusion module 102 is configured to extract basic features of the structured health data based on a pre-set health anomaly knowledge base, and to perform multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features.

[0239] The risk level evaluation module 103 is configured to perform risk level evaluation on the multi-dimensional feature vector based on a pre-set health risk threshold to obtain a preliminary risk identification of the international traveler.

[0240] The space-time fusion analysis module 104 is configured to obtain travel trajectory information of the international traveler, and to perform space-time context fusion analysis on the preliminary risk identification and the travel trajectory information to obtain a comprehensive risk evaluation result of the international traveler.

[0241] The early warning signal generation module 105 is configured to generate an early warning signal of the international traveler according to the comprehensive risk evaluation result.

[0242] In several embodiments provided by the present application, it should be understood that the disclosed method and system can be implemented in other ways. For example, the system embodiments described above are merely illustrative. For example, the division of the modules is only a logical functional division, and other division methods can be used in actual implementation.

[0243] The modules described as separate components may or may not be physically separate, and the components displayed as modules may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0244] In addition, each functional module in various embodiments of the application can be integrated in one processing unit, or each unit can exist physically independently, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of hardware plus software functional modules.

[0245] It is obvious for those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and the present application can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application.

[0246] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Among them, artificial intelligence is to use digital computers or machine controlled by digital computers to simulate, extend and expand human intelligence, perceive environment, obtain knowledge and use knowledge to obtain the best results.

[0247] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A method for intelligent early warning of health abnormalities in international travelers, characterized in that, The method includes: S1. Collect multi-source health data of international travelers and standardize the multi-source health data to obtain structured health data of the international travelers; S2. Based on the knowledge base of health abnormalities of international travelers, extract the basic features of the structured health data, and perform multi-dimensional fusion on the basic features to obtain the multi-dimensional feature vector of the basic features; S3. Based on a preset health risk threshold, assess the risk level of the multidimensional feature vector to obtain a preliminary risk identifier for the international traveler; S4. Obtain the travel trajectory information of the international traveler, and perform spatiotemporal context fusion analysis on the preliminary risk identifier and the travel trajectory information to obtain the comprehensive risk assessment result of the international traveler; S5. Based on the comprehensive risk assessment results, generate a warning signal for the international traveler.

2. The intelligent early warning method for health abnormalities of international travelers as described in claim 1, characterized in that, The process of collecting multi-source health data from international travelers and standardizing the multi-source health data to obtain structured health data of the international travelers includes: Collect raw health data from international traveler health data sources; The original health data is cleaned of noise to obtain cleaned health data from the original health data. The clean and health data is processed to standardize its format, resulting in structured health data for the international traveler.

3. The intelligent early warning method for health abnormalities of international travelers as described in claim 1, characterized in that, The method, based on the knowledge base of health anomalies of international travelers, extracts the basic features of the structured health data and performs multi-dimensional fusion on the basic features to obtain a multi-dimensional feature vector of the basic features, including: Based on the feature mapping relationship of the health anomaly knowledge base, feature mapping is performed on the structured health data to obtain the basic features of the structured health data; The basic features are normalized to obtain the standardized basic features of the basic features; The standardized basic features are weighted and fused to obtain the fused features of the basic features; The fusion features are reconstructed in dimensions to obtain the multidimensional feature vector of the basic features.

4. The intelligent early warning method for health abnormalities of international travelers as described in claim 3, characterized in that, The weighted fusion processing of the standardized basic features to obtain the fused features of the basic features includes: Based on a preset weight allocation strategy, the influence of the standardized basic features is evaluated to obtain the initial fusion weights of the standardized basic features. Calculate the relative dispersion among the standardized basic features, wherein the formula for calculating the relative dispersion is as follows: ； In the formula, The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the One characteristic, This represents the mean of all features in the standardized basic features. This represents the standard deviation of all features in the standardized basic features; Based on the relative degree of dispersion, the initial fusion weights are adaptively corrected to obtain the fusion weights of the standardized basic features; Based on the fusion weights, the standardized basic features are weighted and fused to obtain a preliminary fusion result of the standardized basic features; The numerical fluctuations in the preliminary fusion results are eliminated to obtain the fusion features of the basic features.

5. The intelligent early warning method for health abnormalities of international travelers as described in claim 4, characterized in that, The calculation formula for the preliminary fusion result is as follows: ； In the formula, This indicates the preliminary fusion result. This represents the total number of features in the standardized basic features. The first standardized basic feature represents the The fusion weights of each feature The first standardized basic feature represents the The fusion weights of each feature This represents the preset dispersion adjustment coefficient. The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the The relative dispersion of each feature The first standardized basic feature represents the One characteristic.

6. The intelligent early warning method for health abnormalities of international travelers as described in claim 1, characterized in that, The process of assessing the risk level of the multidimensional feature vector based on a preset health risk threshold to obtain a preliminary risk identifier for the international traveler includes: Based on a preset health risk threshold, the multidimensional feature vector is compared item by item to obtain the individual risk comparison result of the multidimensional feature vector; The individual risk information of the multidimensional feature vector is obtained by analyzing the individual risk comparison results. The individual risk information is identified by anomaly feature identification to obtain the key risk features of the multidimensional feature vector; Based on the risk level determination rules in the health risk threshold, the key risk features are comprehensively evaluated to obtain the preliminary risk level of the multidimensional feature vector; Based on a preset risk identifier, the preliminary risk level is symbolically mapped to obtain the preliminary risk identifier of the international traveler.

7. The intelligent early warning method for health abnormalities of international travelers as described in claim 1, characterized in that, The process of acquiring the international traveler's travel trajectory information and performing spatiotemporal context fusion analysis on the preliminary risk identifier and the travel trajectory information to obtain the international traveler's comprehensive risk assessment result includes: Obtain the travel itinerary information of the international traveler, and add timestamps to the travel itinerary information to obtain the travel nodes of the travel itinerary information; The preliminary risk identifier is correlated and matched with the travel trajectory information to obtain the initial risk value of the travel trajectory information; The environmental risk factors of the region where the travel node is located are spatiotemporally aggregated to obtain the regional risk coefficient of the travel node; The dynamic risk value of the travel node is calculated based on the preset spatiotemporal attenuation effect and regional risk transmission. By integrating and analyzing the dynamic risk values, a comprehensive risk assessment result for the international traveler is obtained.

8. The intelligent early warning method for health abnormalities of international travelers as described in claim 7, characterized in that, The formula for calculating the dynamic risk value is as follows: ； In the formula, Indicates the first The dynamic risk value of each of the aforementioned travel nodes. Indicates the first The dynamic risk value of each of the aforementioned travel nodes. Represents the natural constant. This represents the preset time decay coefficient. This represents the time interval between adjacent travel nodes in the travel node. Indicates the first The regional risk coefficient of each of the aforementioned travel nodes. This represents the preset spatial conductivity coefficient. This indicates the spatial correlation between adjacent travel nodes in the travel node.

9. The intelligent early warning method for health abnormalities of international travelers as described in claim 1, characterized in that, The step of generating a warning signal for the international traveler based on the comprehensive risk assessment results includes: The comprehensive risk assessment result is compared with the preset early warning level threshold at multiple levels to obtain the target early warning level of the comprehensive risk assessment result; Based on the target warning level, a preset warning content template is invoked to generate basic warning information for the comprehensive risk assessment result; The key information from the comprehensive risk assessment results and the basic early warning information are combined into complete early warning information based on the comprehensive risk assessment results; The complete early warning information is encapsulated to obtain the early warning signal for the international traveler.

10. An intelligent early warning system for health abnormalities in international travelers, characterized in that, The system includes: The data acquisition and preprocessing module is used to collect multi-source health data of international travelers and to standardize the multi-source health data to obtain structured health data of the international travelers. The multidimensional feature fusion module is used to extract the basic features of the structured health data based on a preset health anomaly knowledge base, and to perform multidimensional fusion on the basic features to obtain the multidimensional feature vector of the basic features. The risk level assessment module is used to assess the risk level of the multidimensional feature vector based on a preset health risk threshold, and obtain the preliminary risk identifier of the international traveler. The spatiotemporal fusion analysis module is used to acquire the travel trajectory information of the international traveler, and to perform spatiotemporal context fusion analysis on the preliminary risk identifier and the travel trajectory information to obtain the comprehensive risk assessment result of the international traveler; The early warning signal generation module is used to generate an early warning signal for the international traveler based on the comprehensive risk assessment results.

Images