Civil aviation logistics information query method based on big data analysis

Abstract

The present invention discloses a civil aviation logistics information query method based on big data analysis, and specifically relates to the technical field of aviation logistics information query. The specific steps include: constructing a window detection model to identify and judge the update status of logistics information, and outputting the type of abnormal update status, constructing a machine learning model to judge the spatial status of goods, and combining logistic regression analysis and time series analysis to predict the transportation nodes and time of the goods, and checking the goods according to the predicted transportation nodes and time. If the goods cannot be checked, the lost logistics information is supplemented by a graph neural network, and the supplemented logistics information is input into the window detection model to judge the spatial status of the goods and predict the transportation nodes and time of the goods, until the actual logistics status of the goods is checked and uploaded and updated, which effectively improves the search and inspection efficiency of goods with lost logistics information in air transportation and avoids the experience dependence of manual operation.

Description

Technical Field

[0001] The present invention relates to the technical field of aviation logistics information query, and more specifically, to a civil aviation logistics information query method based on big data analysis. Background Art

[0002] Civil aviation logistics information query usually relies on the tracking systems of various logistics links, including the transportation management system and warehouse management system used by airlines, third-party logistics companies, freight forwarders, etc. When the goods pass through the scanning nodes, such as airport cargo handling centers, loading and unloading platforms, sorting centers, etc., the transportation management system and warehouse management system update the status and location information of the goods. Users query the status of the goods by waybill number or flight number. If the logistics information is interrupted or lost, manual intervention is relied on for remediation. The goods are manually tracked according to the latest logistics information, and the relevant transportation nodes are contacted for confirmation, or the goods are relocated by querying warehouse or flight records and other information;

[0003] The current civil aviation logistics query relies on the scanning and manual input of logistics nodes. Usually, information can only be updated after the goods pass through a certain node, and it is difficult to achieve real-time updates. Especially during transit or transportation, there is a lag in the processing of logistics information loss, and it is impossible to respond quickly. The identification of abnormal events is generally passive, and it is necessary to wait until the logistics information is lost or interrupted for a period of time before it can be discovered. When encountering the problem of logistics information loss, it is usually necessary to rely on manual search and information supplement of the goods, which is time-consuming, labor-intensive and prone to errors. In large-scale logistics operations, the efficiency of manual intervention is very low. In complex scenarios, when goods cross multiple transportation chains or multi-country logistics, the risk of information loss increases. Existing methods are difficult to provide users with real-time and accurate logistics information queries when the logistics information of the goods is lost. At the same time, the judgment of the actual transportation status of the goods with lost logistics information depends on the personal experience of the operating workers, which is uncertain and biased.

[0004] In order to solve the above defects, a technical solution is now proposed. Summary of the invention

[0005] The purpose of the present invention is to provide a civil aviation logistics information query method based on big data analysis to solve the deficiencies in the background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: a civil aviation logistics information query method based on big data analysis, the specific steps include:

[0007] Collect, integrate and correct the logistics information of goods, and update the transportation status of goods in real time;

[0008] Build an empty window detection model to identify the update status of logistics information, perform logical judgment on the identified abnormal update status, and output the type of abnormal update status;

[0009] Based on the known logistics information, rule information, and spatial information combined with the type of abnormal update status, a machine learning model is built to determine the spatial status of the goods, and combined with logistic regression analysis and time series analysis to predict the transportation nodes and time of the goods;

[0010] Inspect the goods according to the predicted transportation nodes and time, and upload and update the actual logistics status of the goods after checking. If the goods cannot be inspected, the missing logistics information is supplemented through the graph neural network, and the supplemented logistics information is input into the empty window detection model for logical judgment;

[0011] Determine the spatial status of the goods and predict the transportation nodes and time of the goods until the actual logistics status of the goods is verified and the actual logistics status of the goods is uploaded and updated.

[0012] Preferably, the construction logic of the empty window detection model is:

[0013] Define the logistics information update delay as , where The update time of logistics information. is the estimated time for the next logistics information update, and the geographic location deviation is , where is the radius of the earth, and the longitude and latitude coordinates of the current cargo are , the estimated longitude and latitude coordinates of the cargo are , is the latitude difference, and the expression is , is the longitude difference, and the expression is , the flight status deviation is , where is the actual status of the flight, is the expected status of the flight, where ,and ,but , the scanning state deviation is , where is the current scanning status, is the expected scan status, where ,and ,but , define an abnormal state function for the empty window detection model , where Update the delay threshold for the preset logistics information, is the preset geographic location deviation threshold.

[0014] Preferably, the method for performing logical judgment based on the identification result of the update status of the logistics information by the empty window detection model is:

[0015] When the empty window detection model outputs an abnormal status function of 1, the goods are marked as delayed in logistics information update; when the empty window detection model outputs an abnormal status function of 2, the goods are marked as geographical location offset; when the empty window detection model outputs an abnormal status function of 3, the goods are marked as flight abnormality; when the empty window detection model outputs an abnormal status function of 4, the goods are marked as incorrect scanning.

[0016] Preferably, the machine learning model consists of a logistic regression model based on logistic regression analysis and a time series model based on time series analysis. The logistic regression model is used to judge the spatial state of the cargo, including in the flight and on the ground. The time series model is used to predict the transportation nodes and time of the cargo.

[0017] Preferably, the method of using the logistic regression model to determine the spatial state of the cargo is:

[0018] The input features of the logistic regression model include logistics information, rule information, spatial information, and the type of abnormal update status. Logistics information includes the current cargo coordinates. , Expected cargo coordinates , the rule information is the average transportation time , Frequencies of transportation delays , spatial information includes flight number , Flight schedule , Current flight status , Flight Path , then the input features of the logistic regression model are ,in, is the type of abnormal update state, and the mathematical expression of the logistic regression model is , where is the output probability of the logistic regression model, indicating the probability of the cargo being lost on the flight. Indicates that the cargo is on the flight. Indicates that the cargo is on the ground. is the input feature of the logistic regression model, and n is a positive integer. is the weight parameter of the logistic regression model;

[0019] Preset classification threshold , the calculated With classification threshold For comparison, when Greater than the classification threshold When the cargo is judged to have lost its logistics information on the flight, Less than or equal to the classification threshold When the goods are on the ground, it is determined that the logistics information is lost.

[0020] Preferably, the method for constructing a time series model for predicting the transportation nodes and time of goods is:

[0021] The logistics information synchronized with timestamps and the output of the logistic regression model are used as input to build a time series model. The logistics information includes geographic location information, historical node information, and timestamps. The logistics information synchronized with timestamps and the output of the logistic regression model are converted into vector form. The vector form of the output of the logistic regression model is , the vector form of geographic location information is , for each historical node , the vector form of historical node information is , where the Kronecker delta function The function expression is , where i and j are two indexes representing different categories, i is the position of the current category, and i represents the node The index in all historical nodes, j represents the jth position in the vector, j is the index of the vector, k is the number of j, and k is a positive integer; the vector form of the timestamp is , where is the moment, then the input vector of the time series model is ;

[0022] After the input vector is trained with the time series model, the predicted transportation node and time are obtained.

[0023] Preferably, the method for completing the lost logistics information through the graph neural network is:

[0024] Represent logistics information in graphical form , is a collection of nodes, including airports, logistics sites, and sorting centers. It is a collection of transportation paths between nodes, including flight paths and ground transportation paths. It forms a node feature matrix based on the characteristics of the nodes, and forms an adjacency matrix based on the connection relationship of the logistics network. It receives information from the node's neighbor nodes through the message passing mechanism and updates the node's feature representation. It also performs information reasoning and completion on the nodes that have lost logistics information, and outputs the completion results of the nodes that have lost logistics information.

[0025] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0026] The present application predicts the logistics status of goods through the empty window detection model, thereby avoiding the problem that it is difficult to promptly discover the lost logistics information of goods. By constructing a machine learning model, the goods with lost logistics information are classified using logistic regression analysis, the spatial status of the goods is determined, and the transportation nodes and time of the goods with lost logistics information are predicted based on time series analysis. Manual inspection is performed based on the predicted results. If the manual inspection fails, the transportation status of the goods with lost logistics information is inferred and completed using a graph neural network. The result of the inference and completion is input into the empty window detection model, and logical judgment, classification and prediction are repeated until the actual transportation status of the goods with lost logistics information is obtained through manual inspection, and the actual transportation status is uploaded and updated for user query, which effectively solves the dependence of manual inspection of goods on the personal experience of operators, continuously updates the inspection scope through a cyclic update method, and narrows the inspection scope through a machine learning model, thereby improving the search efficiency of goods with lost logistics information. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0028] Figure 1 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0029] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0030] Example 1: Please refer to Figure 1 As shown, the present invention is a civil aviation logistics information query method based on big data analysis, and the specific steps include:

[0031] Collect, integrate and correct the logistics information of goods, and update the transportation status of goods in real time;

[0032] Build an empty window detection model to identify the update status of logistics information, perform logical judgment on the identified abnormal update status, and output the type of abnormal update status;

[0033] Based on the known logistics information, rule information, and spatial information combined with the type of abnormal update status, a machine learning model is built to determine the spatial status of the goods, and combined with logistic regression analysis and time series analysis to predict the transportation nodes and time of the goods;

[0034] Inspect the goods according to the predicted transportation nodes and time, and upload and update the actual logistics status of the goods after checking. If the goods cannot be inspected, the missing logistics information is supplemented through the graph neural network, and the supplemented logistics information is input into the empty window detection model for logical judgment;

[0035] Determine the spatial status of the goods and predict the transportation nodes and time of the goods until the actual logistics status of the goods is verified and the actual logistics status of the goods is uploaded and updated.

[0036] Collect, integrate and correct the logistics information of goods, and update the transportation status of goods in real time;

[0037] The logistics information of goods includes basic information of goods, shipping and receiving information, transportation route information, flight information, and transportation node information. Basic information of goods includes unique identification code, waybill number, category, weight, and size of goods. The unique identification code may be in the form of barcode, QR code, RFID tag, etc. Shipping and receiving information includes the address and contact information of the shipper and consignee. Transportation route information is the transportation route of goods from the place of shipment to the destination, including transit airports, stopover points, and sorting centers. Flight information includes the flight number, departure time, estimated arrival time, and flight dynamics of the goods. Flight dynamics include flight delays and flight cancellations. Transportation node information includes the storage and outbound time of goods at the transportation node, transportation vehicle number, and loading and unloading information. Loading information includes whether the vehicle is loaded, unloaded, or handed over.

[0038] Integrate logistics information data from different logistics nodes, equipment and systems, and classify and standardize logistics information data through the ETL process. The ETL process includes three steps: extraction, conversion and loading. It extracts logistics information of goods from different sources, performs format conversion and field standardization on data in different formats, and loads the converted data into a unified data warehouse or database. Logistics nodes include airports, storage centers, sorting centers, etc. Equipment includes barcode scanners, RFID readers, GPS devices, etc. Systems include airline transportation management systems, third-party logistics platform warehouse management systems, etc.

[0039] After the logistics information data is classified and standardized, the collected data is corrected to avoid errors, duplications and omissions;

[0040] Methods for correcting logistics information data include:

[0041] Remove duplicate logistics records, correct incorrect scans or data inputs through historical data comparison and manual review, synchronize the timestamps of all logistics information, and supplement missing data through historical data, transportation mode, and GPS information.

[0042] After obtaining the integrated and corrected logistics information data, the transportation status of the goods is updated in real time. When the goods pass through the logistics node, the logistics information of the goods passing through the node is automatically collected. For goods in ground transportation, the geographical location of the goods is obtained in real time through GPS equipment. The logistics information of the goods is analyzed and updated in real time by adopting a stream data processing framework. Commonly used stream data processing frameworks include Apache Kafka, Apache Flink, etc.

[0043] Input the updated logistics information data into the empty window detection model, perform abnormal detection on the logistics status according to the empty window detection model, identify the abnormal update status of the goods, perform logical judgment on the abnormal update status, and output the type of the abnormal update status;

[0044] The construction logic of the empty window detection model is:

[0045] Define the logistics information update delay as , where The update time of logistics information. is the estimated time for the next logistics information update, and the geographic location deviation is , where is the radius of the earth, and the longitude and latitude coordinates of the current cargo are , the estimated longitude and latitude coordinates of the cargo are , is the latitude difference, and the expression is , is the longitude difference, and the expression is , the flight status deviation is , where is the actual status of the flight, is the expected status of the flight, where ,and ,but , the scanning state deviation is , where is the current scanning status, is the expected scan status, where ,and ,but , define an abnormal state function for the empty window detection model , where Update the delay threshold for the preset logistics information, is a preset geographic location deviation threshold;

[0046] The update status of logistics information is logically judged based on the results of the empty window detection model. The judgment method is as follows:

[0047] When the empty window detection model outputs an abnormal status function of 1, the goods are marked as delayed in logistics information update; when the empty window detection model outputs an abnormal status function of 2, the goods are marked as geographical location offset; when the empty window detection model outputs an abnormal status function of 3, the goods are marked as flight abnormality; when the empty window detection model outputs an abnormal status function of 4, the goods are marked as incorrect scanning.

[0048] Based on the known logistics information, rule information, and spatial information combined with the type of abnormal update status, a machine learning model is built to determine the spatial status of the goods, and combined with logistic regression analysis and time series analysis to predict the transportation nodes and time of the goods;

[0049] The machine learning model consists of a logistic regression model based on logistic regression analysis and a time series model based on time series analysis. The logistic regression model is used to judge the spatial status of the cargo, including in the flight and on the ground. The time series model is used to predict the transportation nodes and time of the cargo.

[0050] Logistic regression analysis has strong explanatory power and performs well when inputting continuous and categorical variables. It is suitable for binary classification of goods on the ground or in flight. The method of using the logistic regression model to determine the spatial status of goods is as follows:

[0051] The input features of the logistic regression model include logistics information, rule information, spatial information, and the type of abnormal update status. Logistics information includes the current cargo coordinates. , Expected cargo coordinates , the rule information is the average transportation time , Frequencies of transportation delays , spatial information includes flight number , Flight schedule , Current flight status , Flight Path , then the input features of the logistic regression model are ,in, is the type of abnormal update state, and the mathematical expression of the logistic regression model is , where is the output probability of the logistic regression model, indicating the probability of the cargo being lost on the flight. Indicates that the cargo is on the flight. Indicates that the cargo is on the ground. is the input feature of the logistic regression model, and n is a positive integer. is the weight parameter of the logistic regression model;

[0052] Preset classification threshold , the calculated With classification threshold For comparison, when Greater than the classification threshold When the cargo is judged to have lost its logistics information on the flight, Less than or equal to the classification threshold When the goods are on the ground, it is determined that the logistics information is lost;

[0053] The method for constructing a time series model to predict the transportation nodes and time of goods is as follows:

[0054] The logistics information synchronized with timestamps and the output of the logistic regression model are used as input to build a time series model. The logistics information includes geographic location information, historical node information, and timestamps. The logistics information synchronized with timestamps and the output of the logistic regression model are converted into vector form. The vector form of the output of the logistic regression model is , the vector form of geographic location information is , for each historical node , the vector form of historical node information is , where the Kronecker delta function The function expression is , where i and j are two indexes representing different categories, i is the position of the current category, and i represents the node The index in all historical nodes, j represents the jth position in the vector, j is the index of the vector, k is the number of j, and k is a positive integer; the vector form of the timestamp is , where is the moment, then the input vector of the time series model is ;

[0055] After the input vector is trained with the time series model, the predicted transportation node and time are obtained. The goods are inspected according to the predicted transportation node and time. After the actual logistics status of the goods is inspected, the actual logistics status of the goods is uploaded and updated. If the goods cannot be inspected, the missing logistics information is supplemented through the graph neural network, and the supplemented logistics information is input into the empty window detection model for logical judgment;

[0056] Represent logistics information in graphical form , is a collection of nodes, including airports, logistics sites, and sorting centers. It is a collection of transportation paths between nodes, including flight paths and ground transportation paths. A node feature matrix is ​​formed according to the characteristics of the nodes, and an adjacency matrix is ​​formed according to the connection relationship of the logistics network. Information is received from the neighboring nodes of the node through the message passing mechanism and the feature representation of the node is updated. Information reasoning and completion are performed on the nodes with lost logistics information, and the completion results of the nodes with lost logistics information are output. According to the completion results, the empty window detection model is input for logical judgment, and the type of abnormal update status is output. The spatial status of the goods is judged through the machine learning model, and the transportation nodes and time of the goods are predicted in combination with logistic regression analysis and time series analysis. The goods are inspected according to the predicted transportation nodes and time until the actual logistics status of the goods with lost logistics information is obtained, and the actual logistics status of the goods is uploaded and updated.

[0057] The present application predicts the logistics status of goods through the empty window detection model, thereby avoiding the problem that it is difficult to promptly discover the lost logistics information of goods. By constructing a machine learning model, the goods with lost logistics information are classified using logistic regression analysis, the spatial status of the goods is determined, and the transportation nodes and time of the goods with lost logistics information are predicted based on time series analysis. Manual inspection is performed based on the predicted results. If the manual inspection fails, the transportation status of the goods with lost logistics information is inferred and completed using a graph neural network. The result of the inference and completion is input into the empty window detection model, and logical judgment, classification and prediction are repeated until the actual transportation status of the goods with lost logistics information is obtained through manual inspection, and the actual transportation status is uploaded and updated for user query, which effectively solves the dependence of manual inspection of goods on the personal experience of operators, continuously updates the inspection scope through a cyclic update method, and narrows the inspection scope through a machine learning model, thereby improving the search efficiency of goods with lost logistics information.

[0058] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0059] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0060] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0061] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0062] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0063] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0064] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A civil aviation logistics information query method based on big data analysis, characterized in that: The specific steps include: Collect, integrate and correct the logistics information of goods, and update the transportation status of goods in real time; Build an empty window detection model to identify the update status of logistics information, perform logical judgment on the identified abnormal update status, and output the type of abnormal update status; Based on the known logistics information, rule information, and spatial information combined with the type of abnormal update status, a machine learning model is built to determine the spatial status of the goods, and combined with logistic regression analysis and time series analysis to predict the transportation nodes and time of the goods; Inspect the goods according to the predicted transportation nodes and time, and upload and update the actual logistics status of the goods after checking. If the goods cannot be inspected, the missing logistics information is supplemented through the graph neural network, and the supplemented logistics information is input into the empty window detection model for logical judgment; Determine the spatial status of the goods and predict the transportation nodes and time of the goods until the actual logistics status of the goods is checked and the actual logistics status of the goods is uploaded and updated; The construction logic of the empty window detection model is: Define the logistics information update delay as ΔT = T up -T ex , where T up is the update time of logistics information, T ex is the estimated time for the next logistics information update, and the geographic location deviation is Where R is the radius of the earth, and the longitude and latitude coordinates of the current cargo are The estimated longitude and latitude coordinates of the cargo are is the latitude difference, and the expression is Δλ is the longitude difference, the expression is Δλ=λ2-λ1, and the flight status deviation is ΔF=F st -F ex , where F st is the actual status of the flight, F ex is the expected status of the flight, where and but The scanning state deviation is ΔS = S sc -S ex , where S sc is the current scanning status, S ex is the expected scan status, where and but Define an abnormal state function for the empty window detection model Where, T th is the preset logistics information update delay threshold, L th is the preset geographic location deviation threshold.

2. The civil aviation logistics information query method based on big data analysis according to claim 1 is characterized in that: The method for making logical judgments based on the results of the empty window detection model on the updated status of logistics information is as follows: When the empty window detection model outputs an abnormal status function of 1, the goods are marked as delayed in logistics information update; when the empty window detection model outputs an abnormal status function of 2, the goods are marked as geographical location offset; when the empty window detection model outputs an abnormal status function of 3, the goods are marked as flight abnormality; when the empty window detection model outputs an abnormal status function of 4, the goods are marked as incorrect scanning.

3. The civil aviation logistics information query method based on big data analysis according to claim 1 is characterized in that: The machine learning model consists of a logistic regression model based on logistic regression analysis and a time series model based on time series analysis. The logistic regression model is used to judge the spatial status of the cargo, which includes in the flight and on the ground. The time series model is used to predict the transportation nodes and time of the cargo.

4. The civil aviation logistics information query method based on big data analysis according to claim 3 is characterized in that: The method of using the logistic regression model to determine the spatial status of goods is: The input features of the logistic regression model include logistics information, rule information, spatial information and the type of abnormal update status. The logistics information includes the current cargo coordinates Fe1 and the expected cargo coordinates Fe2. The rule information is the average transportation time Fe3 and the transportation delay frequency Fe4. The spatial information includes the flight number Fe5, the flight schedule time Fe6, the current flight status Fe7 and the flight path Fe8. The input features of the logistic regression model are X = [Fe1, Fe2, Fe3, Fe4, Fe5, Fe6, Fe7, Fe8, A], where A is the type of abnormal update status. The mathematical expression of the logistic regression model is: Where P(Y=1|X) is the output probability of the logistic regression model, which indicates the probability of the cargo being lost on the flight. Y=1 indicates that the cargo is on the flight, Y=0 indicates that the cargo is on the ground, and [X1, X2, ...X n ] is the input feature of the logistic regression model, and n is a positive integer, [ω0, ω1, ...ω n ] is the weight parameter of the logistic regression model; Preset classification threshold P th , the calculated P(Y=1|X) and the classification threshold P th Compare, when P(Y=1|X) is greater than the classification threshold P th When P(Y=1|X) is less than or equal to the classification threshold P th When the goods are on the ground, it is determined that the logistics information is lost.

5. The civil aviation logistics information query method based on big data analysis according to claim 4 is characterized in that: The method for constructing a time series model to predict the transportation nodes and time of goods is as follows: The logistics information synchronized with the timestamp and the output of the logistic regression model are used as input to build a time series model. The logistics information includes geographic location information, historical node information, and timestamps. The logistics information synchronized with the timestamp and the output of the logistic regression model are converted into vector form. The vector form of the output of the logistic regression model is Q out =[P(Y=1|X)], the vector form of geographic location information is Q geo = [longitude, latitude], for each historical node c i , the vector form of historical node information is Among them, Kronecker delta function δ ij The function expression is In the formula, i and j are two indexes representing different categories, i is the position of the current category, and i represents node c i The index in all historical nodes, j represents the jth position in the vector, j is the index of the vector, k is the number of j, and k is a positive integer; the vector form of the timestamp is In the formula, Time is the moment, so the input vector of the time series model is After the input vector is trained with the time series model, the predicted transportation node and time are obtained.

6. The civil aviation logistics information query method based on big data analysis according to claim 1 is characterized in that: The method of completing the missing logistics information through graph neural network is: The logistics information is represented in the form of a graph G = (V, E), where V is a set of nodes, including airports, logistics sites, and sorting centers; E is a set of transportation paths between nodes, including flight paths and ground transportation paths. A node feature matrix is ​​formed according to the characteristics of the nodes, and an adjacency matrix is ​​formed according to the connection relationship of the logistics network. Information is received from the node's neighbor nodes through the message passing mechanism and the node's feature representation is updated. Information reasoning is performed on the nodes that have lost logistics information, and the completion results of the nodes that have lost logistics information are output.

Images