Air-rail intermodal trip chain identification method based on ticket data

Abstract

The application discloses a kind of air and rail intermodal travel chain identification method based on ticket data.The method constructs passenger continuous travel sequence by fusing aviation and railway heterogeneous ticket data, uses travel mode identification difference value to screen out iron-air or air-iron candidate pair;Introduce field station-city mapping space constraint to remove pseudo intermodal of different place transfer;For the same city candidate pair, the minimum transfer time is instantiated according to hub space distance and connection characteristics in three scenarios, and the minimum value of the absolute physical upper limit of the travel chain and the statistical threshold of the pure transfer group fitted by the Gaussian mixture model is determined to determine the maximum transfer time, form a differentiated dynamic effective transfer time threshold interval, retain the candidate pair whose transfer time falls within the interval, and output air and rail intermodal travel chain.The present application makes up for the lack of space constraints, poor adaptability of fixed threshold and other shortcomings of existing methods, has the advantages of high identification accuracy, low misjudgment rate, lightweight algorithm and easy engineering expansion.

Description

Technical Field

[0001] This invention relates to the field of integrated transportation management and data processing technology, specifically a method for identifying air-rail intermodal travel chains based on passenger ticket data. Background Technology

[0002] Air-rail intermodal transport is an important component of a comprehensive three-dimensional transportation system. Accurately identifying passengers' air-rail intermodal travel chains—that is, the complete journey in which passengers use both high-speed rail and air travel consecutively during an intercity trip—is the foundation and prerequisite for analyzing intermodal passenger travel behavior, optimizing intermodal travel plans, and improving the quality of intermodal services.

[0003] Currently, the identification of air-rail intermodal travel chains faces the following major technical challenges: First, data heterogeneity and fragmentation. Airline ticketing systems and railway ticketing systems are independent, with inconsistent data formats, storage standards, and passenger identification systems, making it difficult to link passenger itineraries across systems. Second, limited identification rules. Existing research often relies on fixed time thresholds for judgment, failing to consider the significant differences in transfer distances and traffic conditions between different cities and hubs, leading to a high misclassification rate. Third, lack of spatial constraints. Simply relying on time sorting cannot distinguish between "same-city transfers" and "intercity transfers," and the lack of spatial constraints results in many intercity transfer trips being misclassified as intermodal transport. Fourth, poor scalability. Faced with massive amounts of ticketing data, traditional methods based on rule bases or simple threshold scanning are computationally inefficient, making it difficult to achieve fast and accurate batch identification.

[0004] Therefore, there is an urgent need for a method that can integrate heterogeneous ticket data, introduce dual adaptive constraints of space and time, and achieve efficient and accurate identification of air-rail intermodal transport travel chains. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for identifying air-rail intermodal transport chains based on passenger ticket data. The aim is to achieve efficient and accurate identification of "air-rail" or "rail-air" intermodal transport chains by constructing a unified passenger travel sequence and introducing dual constraints of spatial distance threshold and differentiated time threshold, thus providing reliable data support for subsequent intermodal behavior analysis and service optimization.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for identifying air-rail intermodal transport travel chains based on passenger ticket data, comprising the following steps: Obtain passengers' travel records, including air travel records and rail travel records; Travel records must include at least the de-identified passenger's unique identifier, travel time, and origin and destination hub information; origin and destination hubs are collectively referred to as departure hubs and arrival hubs. Using the de-identified passenger's unique identifier as an index, all travel records of the same passenger are merged, sorted in ascending order by travel time, and a continuous travel sequence of the passenger is constructed. Traverse the continuous travel sequence of passengers, take two adjacent travel records as the initial candidate pairs, and use the difference between the two adjacent travel records to determine whether the corresponding initial candidate pairs are "rail to air" or "air to rail" candidate pairs. For each candidate pair, the transfer location is determined based on the station's geospatial information. If it is, the candidate pair is retained by the same-city constraint; otherwise, the candidate pair is directly eliminated. For each candidate pair that passes the same-city discrimination, calculate the transfer time and calculate a differentiated dynamic threshold range based on the transfer hub scenario type. Determine whether the transfer time falls within the range. If yes, the candidate pair is judged to have passed the valid transfer time discrimination and is retained. Otherwise, the candidate pair is directly removed. Candidate pairs that satisfy the same-city constraint and effective transfer time criteria are combined to generate and output the air-rail intermodal travel chain.

[0007] Furthermore, the continuous travel sequence of passengers is represented as follows: , Indicates the passenger's number Travel records. , This represents the total number of trips taken by passengers within the statistical period; continuous travel sequences. Each travel record includes the mode of transport identifier, travel time, and origin and destination hub information.

[0008] Furthermore, the specific process of step S3 is as follows: Traverse the continuous travel sequences of passengers The two consecutive travel records are used as the initial candidate pairs, and the two consecutive travel records are calculated as the first... Trip record With the Trip record Difference in travel mode identification ,in Indicates the first The mode of transport identifier for each trip. Indicates the first The mode of transport identifier for each trip; like If so, the initial candidate pair is determined to be a "steel-to-short" candidate pair; if If so, the initial candidate pair is determined to be a "free-to-rail" candidate pair; if If the two adjacent travel records in the initial candidate pair are determined to be consecutive travels of the same mode of transportation, the initial candidate pair will be directly removed.

[0009] Furthermore, the specific process of step S4 is as follows: For each candidate pair, extract the first... Trip record Arrival station and the first Trip record The departure station is used to call a geographic information system or a preset station-city mapping table to determine the first... Trip record Arrival station and the first Trip record The system determines whether the departure stations are located in the same region. If they are in the same region, the system uses the same-city constraint to determine and retains the corresponding candidate pair. If they are not in the same region, the system determines that the candidate pair is invalid and removes it directly.

[0010] Furthermore, the specific process of step S5 is as follows: Step S5.1: Calculate transfer time: Calculate the transfer time of the candidate pair that passes the same-city discrimination. The first trip record and the first Transfer time between trips ; Step S5.2: Establish the effective transfer time threshold interval: Define the effective transfer time threshold interval as follows: ,in Identifiers indicating a transfer hub scenario. Indicates the minimum transfer time. Indicates the maximum transfer time; Step S5.3: Determine valid transfers: Based on the valid transfer time threshold interval, select the candidate pairs that have passed the same-city discrimination. The first trip record and the first Transfer time between trips The judgment is performed, and the judgment rule is: if If the candidate pair is determined by the effective transfer time, it is retained; otherwise, it is directly eliminated.

[0011] Furthermore, the minimum transfer time is instantiated based on the type of transfer hub scenario: According to the The arrival hub of the first trip record and the first Based on the spatial distance and connection characteristics between departure hubs in each trip record, the transfer hub scenario is divided into three typical types: integrated hub scenario, separated adjacent hub scenario, and separated distant hub scenario. The general calculation model for minimum transfer time is instantiated and calculated for each type. In the context of an integrated transportation hub, the calculation model for the minimum transfer time is as follows: ; In the formula, Indicating an integrated hub scenario Minimum transfer time; Indicating an integrated hub scenario The overall processing time is as follows; Indicating an integrated hub scenario Pure walking transfer time, This indicates the actual walking distance of the connecting corridors / passages in an integrated transportation hub scenario; This indicates the average walking speed of passengers carrying luggage; This indicates a preset, smaller safety buffer time; In the scenario of separate, adjacent hubs, the calculation model for minimum transfer time is as follows: ; In the formula, Represents a scenario of separate, neighboring hubs. Minimum transfer time; Represents a scenario of separate, neighboring hubs. Next, the The drop-off and exit times of each trip; Indicates the first The arrival hub of the first trip record and the first Spatial distance between departure hubs in each trip record; Represents a scenario of separate, neighboring hubs. Next, the The entry and security check times for each trip; Indicates dedicated connection line method The average operating speed; Indicates dedicated connection line method The corresponding low uncertainty coefficient; This indicates a preset medium safety buffer time; In the scenario of a separate, remote hub, the calculation model for the minimum transfer time is as follows: ; In the formula, Represents a separate, remote hub scenario. Minimum transfer time; Represents a separate, remote hub scenario. Next, the The drop-off and exit times of each trip; Represents a separate, remote hub scenario. Next, the The entry and security check times for each trip; Indicates the city's integrated transportation modes Average speed; Indicates the city's integrated transportation modes The corresponding high uncertainty coefficient; This indicates the preset extended safety buffer time.

[0012] Furthermore, a mechanism combining an improved data-driven statistical method with travel chain theory constraints is employed to determine the maximum transfer time: For the corresponding transfer hub scenario, candidate pairs that meet the same-city constraint and whose transfer time is less than the preset absolute upper limit are extracted as samples to form the initial sample set; a priori filtering condition is set: samples whose transfer time is greater than the preset theoretical absolute upper limit are removed to obtain the cleaned effective sample set. The effective sample set was fitted using a Gaussian mixture model. Two Gaussian components were set to correspond to the pure transfer group and the composite stay group, respectively. The weight, mean and standard deviation of each Gaussian component were solved using the expectation-maximization algorithm. The Gaussian component with the smaller mean among the two Gaussian components is defined as the short-time transfer group, i.e., the target group, and a statistical threshold is calculated based on the parameters of the target group. Based on the theory of integrated transportation travel chain, an absolute physical upper limit is preset; the minimum value of the statistical threshold and the absolute physical upper limit is taken to obtain the maximum transfer time for the corresponding transfer hub scenario.

[0013] Furthermore, before using travel records, they are cleaned in a standardized manner, including standardizing date and time formats, standardizing station coding standards, and removing invalid and incomplete records.

[0014] An electronic device includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute a method for identifying air-rail intermodal transport travel chains based on passenger ticket data.

[0015] A non-volatile computer storage medium storing computer-executable instructions that execute a method for identifying air-rail intermodal travel chains based on passenger ticket data.

[0016] Compared with existing technologies, the present invention has the following advantages: (1) This invention constructs a progressive identification framework that integrates multi-source heterogeneous ticket data, spatial same-city constraints, and differentiated dynamic time threshold discrimination. Without the need for additional hardware deployment, it can completely restore the individual travel sequence of passengers and achieve high-precision, low-misjudgment batch identification of air-rail intermodal travel chains. Among them, the introduction of station-city mapping relationship fundamentally eliminates pseudo intermodal candidate pairs such as intercity transfers and eliminates structural misjudgments caused by simply relying on time sorting. The maximum transfer time calculation mechanism based on the dual fusion of Gaussian mixture model and physical upper limit of travel chain further achieves a balance between data-driven and traffic theory. This makes the identification results adaptable to the actual connection characteristics between different cities and different hubs, and also has unified engineering robustness. The overall solution can be stably extended to the rapid processing of tens of millions or even hundreds of millions of ticket data while maintaining low computational complexity.

[0017] (2) By introducing a spatial same-city discrimination mechanism of station-city mapping, this invention limits the transfer hub to the same prefecture-level administrative region and forcibly filters out pseudo intermodal transport candidate pairs whose origin and destination are not in the same city. This significantly improves the spatial authenticity of intermodal transport behavior judgment and avoids misjudging cross-city or even cross-province transfer trips as air-rail intermodal transport due to lack of spatial constraints, thereby greatly reducing the misjudgment rate and system noise.

[0018] (3) Based on the spatial distance and connection characteristics between transfer hubs, this invention divides the hub scenario into three types: integrated hub, separated neighboring hub, and separated distant hub. It also instantiates the minimum transfer time calculation model based on the actual operation characteristics of walking, dedicated line connection and urban integrated transportation. At the same time, it uses Gaussian mixture model to fit the probability density of historical ticket data. Based on the identification of pure transfer groups, it takes the smaller value of the absolute physical upper limit of the travel chain to generate an effective transfer time threshold range that is dynamically adjusted according to the physical conditions of the hub and the behavior of passengers. This effectively solves the problem of poor adaptability of traditional fixed thresholds in cities of different sizes, and takes into account the sensitivity and specificity of transfer judgment. Attached Figure Description

[0019] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0020] Example 1 like Figure 1 As shown, the present invention provides a technical solution: a method for identifying air-rail intermodal transport travel chains based on passenger ticket data, comprising the following steps: Step S1: Acquisition of multi-source heterogeneous ticket data: Obtain passengers' air travel records from the air ticket system and passengers' railway travel records from the railway ticket system; the travel records shall at least include the de-identified passenger unique identifier, travel time, and origin and destination hub (station) information; origin and destination hub (station) is the collective term for departure hub (station) and arrival hub (station); The desensitized passenger unique identifier refers to the original sensitive identification information in the air and railway ticketing system that can directly identify the identity of a specific natural person. After irreversible or one-way encryption privacy protection processing, it generates a string or number code that can uniquely distinguish different passengers and cannot be reversed to restore the passenger's real identity information. Before use, travel records obtained from the air and railway ticketing systems are cleaned in a standardized manner, including standardizing date and time formats, standardizing station coding standards, and removing invalid and incomplete records.

[0021] Step S2: Constructing a time-series-based individual passenger travel sequence: Using the anonymized passenger's unique identifier as an index, merge all travel records (air travel records and rail travel records) of the same passenger, sort them in ascending order according to travel time, and construct the passenger's continuous travel sequence. , Indicates the passenger's number Travel records. , This represents the total number of trips taken by passengers within the statistical period; continuous travel sequences. Each travel record includes a travel mode identifier (0=railway, 1=air), travel time, and origin and destination hubs (stations).

[0022] Step S3: Preliminary travel mode conversion judgment: Traverse the passenger's continuous travel sequence, take two adjacent travel records as the initial candidate pair, and use the difference between the two adjacent travel records to determine whether the corresponding initial candidate pair is a "rail to air" or "air to rail" candidate pair.

[0023] The specific process of step S3 is as follows: Traverse the continuous travel sequences of passengers The two consecutive travel records are used as the initial candidate pairs, and the two consecutive travel records are calculated as the first... Trip record With the Trip record Difference in travel mode identification ,in Indicates the first The mode of transport identifier for each trip. Indicates the first The mode of transport identifier for each trip; like If so, the initial candidate pair is determined to be a "steel-to-short" candidate pair; if If so, the initial candidate pair is determined to be a "free-to-rail" candidate pair; if If two consecutive trips in the initial candidate pair are determined to be consecutive trips of the same mode of transportation, the initial candidate pair will be directly removed and will not be included in subsequent judgments.

[0024] Step S4: Determine the same-city constraint for transfer based on spatial distance: For each candidate pair ("rail to air" or "air to rail" candidate pair), determine whether the transfer is located in the same area based on the station's geospatial information. If yes, the candidate pair is determined to pass the same-city constraint and is retained; otherwise, the candidate pair is directly eliminated.

[0025] The specific process of step S4 is as follows: For each candidate pair ("iron to air" or "air to iron" candidate pair), extract the first... Trip record Arrival station and the first Trip record The departure station is used to call a geographic information system or a preset station-city mapping table to determine the first... Trip record Arrival station and the first Trip record Whether the departure station is located in the same region (prefecture-level city or municipality); if it is located in the same region, it is determined by the same city constraint and the corresponding candidate pair is retained; if it is not located in the same region, it is determined as an invalid candidate pair (e.g., transit in different places) and is directly eliminated.

[0026] Step S5: Effective transfer time determination based on differentiated dynamic thresholds: For each candidate pair that passes the same-city determination, calculate the transfer time and calculate the differentiated dynamic threshold range according to the transfer hub scenario type (integrated hub, separate neighboring hub, separate distant hub). Determine whether the transfer time falls within the range. If yes, the candidate pair is determined to pass the effective transfer time determination and is retained. If not, the candidate pair is directly eliminated.

[0027] The specific process of step S5 is as follows: Step S5.1: Calculate transfer time: Calculate the transfer time of the candidate pair that passes the same-city discrimination. The first trip record and the first Transfer time between trips : ; In the formula, This indicates that among the candidate pairs determined by city-based criteria, the first... The planned arrival time for each trip; This indicates that among the candidate pairs determined by city-based criteria, the first... The planned departure time for each trip is recorded.

[0028] Step S5.2: Establish the effective transfer time threshold interval: Define the effective transfer time threshold interval as follows: ,in Identifiers indicating a transfer hub scenario. Indicates the minimum transfer time. Indicates the maximum transfer time; the identifier for a transfer hub scenario includes the transfer hub (station) pair, transfer method, and transfer distance; the definition of a transfer hub pair is: the first... The arrival hub (station) of the first trip record and the first A pair of hubs (stations) consisting of the departure hubs (stations) of each trip record.

[0029] Step S5.3: Determine valid transfers: Based on the valid transfer time threshold interval, select the candidate pairs that have passed the same-city discrimination. The first trip record and the first Transfer time between trips The judgment is performed, and the judgment rule is: if If the candidate pair is determined by the effective transfer time, it is retained; otherwise, it is directly eliminated.

[0030] Among them, the minimum transfer time refers to the time a passenger can travel from the first transfer point under normal conditions. The arrival hub (station) of the first trip record to the first The shortest necessary time required to reach the departure hub (station) for each trip record. The general calculation model for minimum transfer time is: ; In the formula, Indicates the corresponding transfer hub scenario Next, the The time of disembarkation and exit from the station for each trip (for high-speed rail, this includes the time from when the train comes to a complete stop, when the doors open, and when walking to the exit gate; for air travel, this includes the time from taxiing, when the cabin doors open, and when walking to the arrival hall). Indicates the corresponding transfer hub scenario Next, the The arrival hub (station) of the first trip record and the first The transfer time between departure hubs (stations) recorded for each trip. , Indicates the transfer method used. The average speed; Indicates the first The arrival hub (station) of the first trip record and the first The spatial distance between departure hubs (stations) in each trip record. This represents the uncertainty coefficient, reflecting factors such as waiting and congestion. Indicates the corresponding transfer hub scenario Next, the The time spent entering the station and going through security for each trip (for high-speed rail, this includes the time spent walking to the entrance, security checkpoint, and ticket gate; for air travel, this includes the time spent walking from the terminal entrance to the check-in, security checkpoint, and boarding gate). This indicates a safety buffer period, used to cope with random factors such as earlier delays or abnormal delays.

[0031] Specifically, the minimum transfer time is instantiated based on the type of transfer hub scenario: According to the The arrival hub (station) of the first trip record and the first Based on the spatial distance and connection characteristics between departure hubs (stations) in each trip record, the transfer hub scenario is divided into three typical types: integrated hub scenario, separated adjacent hub scenario, and separated distant hub scenario. The general calculation model for minimum transfer time is instantiated and calculated for each type. 1. Integrated hub scenario; Scene characteristics: The arrival hub (station) of the first trip record and the first The departure hubs (stations) for each trip are highly integrated in physical space, directly connected by pedestrian corridors, escalators, etc., without relying on external urban transportation; In an integrated transportation hub scenario, transfer time is simplified to pure walking time; the first Passenger drop-off and exit times recorded for this trip With the Entry and security check times for each trip The process is simplified to a single integrated processing time within the hub. In the context of an integrated transportation hub, the calculation model for the minimum transfer time is as follows: ; In the formula, Indicating an integrated hub scenario Minimum transfer time; Indicating an integrated hub scenario The overall processing time is the sum of empirical values ​​corresponding to the scale of the airport / high-speed rail station (e.g., 20-30 minutes for large hubs). Indicating an integrated hub scenario Pure walking transfer time, Indicates the actual walking distance of the connecting corridor / passage; This represents the average walking speed of passengers carrying luggage (usually taken as 1.0-1.2 m / s), in which case the uncertainty coefficient is considered to be 0; This indicates a relatively short safety buffer time. Since the transfer environment is highly controllable, it is usually taken as an empirical value (such as 10-15 minutes).

[0032] 2. Separate neighboring hub scenarios; Scene characteristics: The arrival hub (station) of the first trip record and the first The departure hubs (stations) of each trip are physically separate but close in distance (usually ≤15km), mainly relying on dedicated connecting lines such as maglev express lines, airport APM, and intercity short-distance railways, and the travel time is highly certain. In the scenario of separate, adjacent hubs, the calculation model for minimum transfer time is as follows: ; In the formula, Represents a scenario of separate, neighboring hubs. Minimum transfer time; Represents a scenario of separate, neighboring hubs. Next, the The drop-off and exit times for each trip must be based on the first... The arrival hub (station) of the first trip record and the first The building scale level of the departure hub (station) for each trip record is taken from a preset experience value library; Indicates the first The arrival hub (station) of the first trip record and the first The spatial distance between departure hubs (stations) for each trip; Represents a scenario of separate, neighboring hubs. Next, the The entry and security check times for each trip need to be based on the first trip. The arrival hub (station) of the first trip record and the first The building scale level of the departure hub (station) for each trip record is taken from a preset experience value library; Indicates dedicated connection line method The average operating speed is directly adopted from the timetable operating speed of rail transit or dedicated lines; Indicates dedicated connection line method The corresponding low uncertainty coefficient, due to the independent right-of-way and frequent service of dedicated bus routes, usually takes a very small value (such as 0.05-0.10), mainly covering the time passengers wait for the next bus. This indicates a preset medium safety buffer time, used to cover minor delays caused by intervals between connecting trains, typically 15-20 minutes.

[0033] 3. Separate, remote hub scenarios; Scene characteristics: The arrival hub (station) of the first trip record and the first The departure hubs (stations) of each trip are distributed in different areas of the city, with long spatial distances (usually more than 20 kilometers). Transfers require crossing the main urban area of ​​the city, and the trip is highly dependent on regular public transportation (subway, bus, ride-hailing). The travel time is easily affected by morning and evening rush hours and traffic congestion, and the uncertainty is extremely high. In the scenario of a separate, remote hub, the calculation model for the minimum transfer time is as follows: ; In the formula, Represents a separate, remote hub scenario. Minimum transfer time; Represents a separate, remote hub scenario. Next, the The drop-off and exit times of each trip; Represents a separate, remote hub scenario. Next, the The entry and security check times for each trip; Indicates the city's integrated transportation modes Average speed (if the transfer is mainly via subway, take the average travel speed of the city's subway network (about 30-35 km / h); if the transfer is mainly via road traffic, historical traffic data should be used to take the average travel speed (about 20-25 km / h)). Indicates the city's integrated transportation modes The corresponding high uncertainty coefficient needs to be significantly amplified to absorb the time fluctuations caused by morning and evening rush hours or complex road conditions (e.g., 0.2-0.3 for subway transfers, and 0.4-0.6 for road buses / ride-hailing services). This indicates a pre-set extended safety buffer time to address random factors that may occur during long-distance intercity transfers, such as getting lost, long walking distances between transfer points, and delays while waiting for transportation. It can be set to 30-45 minutes or even longer.

[0034] Among them, the maximum transfer time This is used to distinguish between passengers' "effective connecting transfer behavior" and "long-term stopover behavior" (such as business, tourism, or overnight stays in transit cities, which are not part of the connecting destination). To ensure that the identification model can adapt to the time distribution characteristics of different hubs and has strong engineering robustness, this invention adopts a mechanism that combines an improved data-driven statistical method with travel chain theory constraints for determination. The specific steps are as follows: 1. For a specific transfer hub scenario, extract all connections that satisfy the same-city constraint and have a transfer time. Candidate pairs are used as samples to form the initial sample set; to avoid interference from extremely long-tailed dirty data in model fitting, a priori filtering condition is set: samples with transfer times greater than a preset theoretical absolute upper limit (e.g., 24 hours) are removed, resulting in a cleaned effective sample set. ; Given the continuous and overlapping characteristics of passenger transfer times in the overall distribution, a Gaussian mixture model (GMM), which can characterize a multi-peaked distribution, is selected for fitting; assuming an effective sample set... Following a Gaussian mixture distribution, set the number of clusters. (Representing "pure transfer groups" and "compound stay groups" respectively), Gaussian mixture model (GMM) probability density function Defined as: ; In the formula, Indicates the first The weights of the Gaussian components, and satisfying ; , They represent the first Gaussian distribution of Gaussian components The mean and variance of the Gaussian component are individual independent Gaussian (normal) distributions that make up the entire mixture distribution. Each Gaussian component contains weight, mean, and variance. The first Gaussian component corresponds to the pure transfer group, and the second Gaussian component corresponds to the composite stay group. The Expectation-Maximization (EM) algorithm is used on the effective sample set. By performing fitting, the optimal Gaussian mixture model parameters can be obtained. , Indicates the first The algorithm calculates the posterior probability of a sample belonging to each group in the expectation step and updates the parameters in the maximization step until the log-likelihood function converges. By comparing the means of the "pure transfer group" and the "combined stay group", the group with the smaller mean is defined as the "short-time transfer group", i.e., the target group. , Based on target group Parameter calculation statistical threshold ,in This represents the critical value of the standard normal distribution, at a confidence level of [missing information]. hour, This means that it covers approximately 95% of the group's transfer behavior; When the effective sample size of the transfer hub scenario When the number of data points is less than the set fitting threshold (e.g., 100), the statistical threshold of mature transfer hub scenarios with similar scale and transfer distance is used as a substitute to ensure the availability of the algorithm during the cold start phase.

[0035] 2. Based on the theory of integrated transportation travel chains, when a passenger's dwell time at a transfer hub exceeds a certain physical limit, the primary focus of their travel destination will shift. Therefore, it is necessary to set an absolute physical upper limit supported by business logic. (e.g., 24 hours) serves as a hard constraint. This constraint acts as a bottom-line fallback mechanism to ensure that the identification results do not violate the basic principles of traffic physics when there are systematic biases or sudden anomalies in the data.

[0036] 3. Statistical thresholds and absolute physical upper limits The minimum value is taken to achieve double filtering and obtain the maximum transfer time for the corresponding transfer hub scenario: ; Ensure that the final determined maximum transfer time is strictly greater than the minimum transfer time, i.e., satisfy the following conditions. ,in The preset minimum time interval (can be 30 minutes or 60 minutes).

[0037] Step S6, Generation and Output of Intermodal Travel Chains: Combine candidate pairs that meet the same-city constraint and effective transfer time criteria to generate and output an air-rail intermodal travel chain. The output results should include at least: passenger unique identifier, intermodal type (rail to air / air to rail), transportation details, transfer time, and the city where the transfer hub is located.

[0038] A second embodiment of the present invention also provides an electronic device, including a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, wherein the memory is used to store a set of program code, and the processor is used to call the program code stored in the memory to execute a method for identifying air-rail intermodal transport travel chains based on passenger ticket data.

[0039] A third embodiment of the present invention also provides a non-volatile computer storage medium storing computer-executable instructions that execute a method for identifying air-rail intermodal travel chains based on passenger ticket data.

[0040] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for identifying air-rail intermodal transport travel chains based on passenger ticket data, characterized in that, The steps include the following: Step S1: Obtain the passenger's travel records, including air travel records and rail travel records; Travel records must include at least the de-identified passenger's unique identifier, travel time, and origin and destination hub information; origin and destination hubs are collectively referred to as departure hubs and arrival hubs. Step S2: Using the de-identified passenger unique identifier as an index, merge all travel records of the same passenger, sort them in ascending order according to travel time, and construct the passenger's continuous travel sequence; Step S3: Traverse the passenger's consecutive travel sequence, take two adjacent travel records as the initial candidate pair, and use the difference between the two adjacent travel records to determine whether the corresponding initial candidate pair is a "rail to air" or "air to rail" candidate pair; Step S4: For each candidate pair, determine whether the transfer is located in the same area based on the station's geospatial information. If yes, retain the candidate pair by passing the same-city constraint; otherwise, directly remove the candidate pair. Step S5: For each candidate pair that passes the same-city discrimination, calculate the transfer time and calculate the differentiated dynamic threshold range according to the transfer hub scenario type. Determine whether the transfer time falls within the range. If yes, it is determined that the candidate pair has passed the valid transfer time discrimination and is retained. Otherwise, the candidate pair is directly removed. Step S6: Combine candidate pairs that satisfy the same-city constraint and effective transfer time criteria to generate and output the air-rail intermodal travel chain; The specific process of step S5 is as follows: Step S5.1: Calculate transfer time: Calculate the transfer time of the candidate pair that passes the same-city discrimination. The first trip record and the first Transfer time between trips ; Step S5.2: Establish the effective transfer time threshold interval: Define the effective transfer time threshold interval as follows: ,in Identifiers indicating a transfer hub scenario. Indicates the minimum transfer time. Indicates the maximum transfer time; Step S5.3: Determine valid transfers: Based on the valid transfer time threshold interval, select the candidate pairs that have passed the same-city discrimination. The first trip record and the first Transfer time between trips The judgment is performed, and the judgment rule is: if If the candidate pair is determined by the effective transfer time, it is retained; otherwise, it is directly eliminated.

2. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 1, characterized in that: Passengers' consecutive travel sequences are represented as , Indicates the passenger's number Travel records. , This represents the total number of trips taken by passengers within the statistical period; continuous travel sequences. Each travel record includes the mode of transport identifier, travel time, and origin and destination hub information.

3. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 2, characterized in that: The specific process of step S3 is as follows: Traverse the continuous travel sequences of passengers The two consecutive travel records are used as the initial candidate pairs, and the two consecutive travel records are calculated as the first... Trip record With the Trip record Difference in travel mode identification ,in Indicates the first The mode of transport identifier for each trip. Indicates the first The mode of transport identifier for each trip; like If so, the initial candidate pair is determined to be a "steel to short" candidate pair; if If so, the initial candidate pair is determined to be a "free-to-rail" candidate pair; if If the two adjacent travel records in the initial candidate pair are determined to be consecutive travels of the same mode of transportation, the initial candidate pair will be directly removed.

4. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 3, characterized in that: The specific process of step S4 is as follows: For each candidate pair, extract the first... Trip record Arrival station and the first Trip record The departure station is used to call a geographic information system or a preset station-city mapping table to determine the first... Trip record Arrival station and the first Trip record Are the departure stations located in the same area? If they are located in the same region, the corresponding candidate pair is retained based on the same-city constraint; if they are not located in the same region, they are deemed invalid candidate pairs and are directly eliminated.

5. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 4, characterized in that: Instantiate the minimum transfer time based on the type of transfer hub scenario: According to the The arrival hub of the first trip record and the first Based on the spatial distance and connection characteristics between departure hubs in each trip record, the transfer hub scenario is divided into three typical types: integrated hub scenario, separated adjacent hub scenario, and separated distant hub scenario. The general calculation model for minimum transfer time is instantiated and calculated for each type. In the context of an integrated transportation hub, the calculation model for the minimum transfer time is as follows: ； In the formula, Indicating an integrated hub scenario Minimum transfer time; Indicating an integrated hub scenario The overall processing time is as follows; Indicating an integrated hub scenario Pure walking transfer time, This indicates the actual walking distance of the connecting corridors / passages in an integrated transportation hub scenario; This indicates the average walking speed of passengers carrying luggage; This indicates a preset, smaller safety buffer time; In the scenario of separate, adjacent hubs, the calculation model for minimum transfer time is as follows: ； In the formula, Represents a scenario of separate, neighboring hubs. Minimum transfer time; Represents a scenario of separate, neighboring hubs. Next, the The drop-off and exit times of each trip; Indicates the first The arrival hub of the first trip record and the first Spatial distance between departure hubs in each trip record; Represents a scenario of separate, neighboring hubs. Next, the The entry and security check times for each trip; Indicates dedicated connection line method The average operating speed; Indicates dedicated connection line method The corresponding low uncertainty coefficient; This indicates a preset medium safety buffer time; In the scenario of a separate, remote hub, the calculation model for the minimum transfer time is as follows: ； In the formula, Represents a separate, remote hub scenario. Minimum transfer time; Represents a separate, remote hub scenario. Next, the The drop-off and exit times of each trip; Represents a separate, remote hub scenario. Next, the The entry and security check times for each trip; Indicates the city's integrated transportation modes Average speed; Indicates the city's integrated transportation modes The corresponding high uncertainty coefficient; This indicates the preset extended safety buffer time.

6. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 5, characterized in that: The maximum transfer time is determined by a mechanism that combines an improved data-driven statistical method with travel chain theory constraints. For the corresponding transfer hub scenario, candidate pairs that meet the same-city constraint and whose transfer time is less than the preset absolute upper limit are extracted as samples to form the initial sample set; a priori filtering condition is set: samples whose transfer time is greater than the preset theoretical absolute upper limit are removed to obtain the cleaned effective sample set. The effective sample set was fitted using a Gaussian mixture model. Two Gaussian components were set to correspond to the pure transfer group and the composite stay group, respectively. The weight, mean and standard deviation of each Gaussian component were solved using the expectation-maximization algorithm. The Gaussian component with the smaller mean among the two Gaussian components is defined as the short-time transfer group, i.e., the target group, and a statistical threshold is calculated based on the parameters of the target group. Based on the theory of integrated transportation travel chains, an absolute physical upper limit is preset; The maximum transfer time for the corresponding transfer hub scenario is obtained by taking the minimum value of the statistical threshold and the absolute physical upper limit.

7. The method for identifying air-rail intermodal transport travel chains based on passenger ticket data according to claim 6, characterized in that: Before use, travel records are cleaned in a standardized manner, including standardizing date and time formats, standardizing station coding standards, and removing invalid and incomplete records.

8. An electronic device, characterized in that, The system includes a processor, a memory, and a bus, wherein the processor and the memory are connected via the bus, the memory is used to store a set of program codes, and the processor is used to call the program codes stored in the memory to execute the air-rail intermodal transport travel chain identification method based on passenger ticket data as described in any one of claims 1-7.

9. A non-volatile computer storage medium storing computer-executable instructions, characterized in that, The computer can execute instructions to perform the air-rail intermodal travel chain identification method based on passenger ticket data as described in any one of claims 1-7.

Images