A conventional bus user loss key path identification method and system
By constructing a user behavior state transition matrix and using various mathematical models to identify key paths of bus user churn, the problem of passenger flow loss in urban regular buses was solved, and bus route optimization and service quality improvement were achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TONGJI UNIV
- Filing Date
- 2023-06-19
- Publication Date
- 2026-07-21
AI Technical Summary
Existing technologies struggle to effectively identify the key pathways through which regular urban public transport users are lost, making it difficult for public transport managers to optimize bus routes and improve service quality in a targeted manner.
By constructing a user behavior state transition matrix and using K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model, the behavioral states of public transport users are identified, and the transition probabilities between states are calculated, thus identifying the key paths of public transport passenger loss.
It enables accurate identification of the reasons for bus user churn, allowing for targeted optimization of bus routes, reduction of operating costs, and improvement of service quality and efficiency. It is highly adaptable and applicable to the urban public transportation sector.
Smart Images
Figure CN116777715B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the technical field of identifying critical paths to passenger churn in urban public transportation, and particularly to a method and system for identifying critical paths to passenger churn in urban public transportation. Background Technology
[0002] In my country's urban public transportation sector, the continuous loss of regular bus passengers is a widespread problem in major cities. This loss is triggered by a combination of internal and external factors. Furthermore, the loss of regular bus users is a continuous process rather than a sudden event. During this process, the behavioral states of regular bus users often undergo a series of continuous shifts. Therefore, to effectively clarify the mechanism of regular bus passenger loss, it is necessary to develop a method that can identify the critical paths of regular bus user loss, so that urban public service managers can take targeted measures.
[0003] Existing technologies addressing this issue primarily fall into two categories. First, some existing studies use machine learning, neural networks, and decision trees based on historical passenger flow data to predict public transport passenger churn, but these methods often fail to effectively identify specific critical paths. Second, some methods collect user feedback through questionnaires or GPS trajectory analysis to understand the reasons for individual public transport user churn and their needs. While these methods can provide some qualitative analysis of public transport user churn, they require significant manpower and time investment.
[0004] Therefore, this invention proposes a method for identifying critical paths to passenger churn in conventional public transportation. This method first identifies the behavioral states of public transportation users within each time period. Then, based on these behavioral states, it establishes a user behavior state transition matrix between adjacent time periods. This matrix describes the transition probabilities of users between different states and reflects the travel behavior characteristics of public transportation users. By analyzing this matrix, the relationships between user behavior states are identified, and the churn probability corresponding to each state is calculated. Furthermore, the critical paths influencing passenger churn in public transportation can be identified.
[0005] The method of this invention can effectively identify key paths influencing public transportation passenger loss, providing public transportation managers with targeted management strategies. Compared with existing technologies, the method of this invention has the following advantages: First, based on the user behavior state transition matrix, the travel behavior characteristics of public transportation users can be quantified, avoiding the subjectivity and uncertainty of traditional survey methods. Second, the method is based on actual data analysis, possessing high accuracy and reliability. Third, the method of this invention can be adjusted according to specific circumstances, exhibiting a certain degree of adaptability and flexibility. Therefore, the method of this invention can provide an effective means to solve the problem of passenger loss in urban public transportation, possessing practicality and broad application prospects. In implementing this invention, various methods can be used to collect public transportation user travel data, such as utilizing GPS trajectory data and public transportation card swipe records. Furthermore, different methods can be used to analyze the state transition matrix, such as clustering and factor analysis. These methods can be selected and adjusted according to specific circumstances to obtain more accurate and reliable analysis results. Summary of the Invention
[0006] The purpose of this section is to outline some aspects of embodiments of the present invention and to briefly describe some preferred embodiments. Simplifications or omissions may be made in this section, as well as in the abstract and title of this application, to avoid obscuring the purpose of these documents; however, such simplifications or omissions should not be construed as limiting the scope of the invention.
[0007] In view of the above-mentioned problems, the present invention is proposed.
[0008] Therefore, the technical problem solved by this invention is: how to identify the key paths of bus user churn by establishing a user behavior state transition matrix, thereby revealing the behavioral change mechanism in the process of passenger flow loss in urban regular buses, and helping to solve the problem of passenger flow loss in the field of urban public transportation; and how to optimize bus routes in a targeted manner, reduce bus operating costs, and improve the service quality and efficiency of buses.
[0009] To address the aforementioned technical problems, this invention provides the following technical solution: a method for identifying critical paths to passenger churn in conventional public transportation, comprising: constructing a behavioral state identification model for urban conventional public transportation users based on multiple mathematical models, identifying the behavioral state of public transportation users in each time period, establishing a user behavioral state transition matrix in adjacent time periods based on the behavioral state of public transportation users, analyzing the matrix to identify the relationship between user behavioral states, calculating the churn probability corresponding to each state, and identifying the critical paths to passenger churn in public transportation.
[0010] As a preferred embodiment of the critical path identification method for the loss of regular public transport users described in this invention, the user behavior state identification model includes: extracting the spatiotemporal features of the travel behavior of regular urban public transport users based on the card swiping data and urban POI data; modeling based on K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model; and extracting the optimal model.
[0011] As a preferred embodiment of the conventional public transport user churn critical path identification method described in this invention, the user behavior state identification model further includes:
[0012] The spatiotemporal characteristics of regular bus users' travel within the unit analysis period required for modeling are normalized. The normalized features are used as the feature set X for model training. The features of the same user in different analysis periods are used as different samples. Then, all regular bus user feature sets are input into the K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model for clustering to obtain the user behavior state recognition results.
[0013] As a preferred embodiment of the conventional public transport user churn critical path identification method described in this invention, the user behavior state identification model further includes using the Davies-Bouldin index, Homogeneity Score, and Completeness Score, which are indicators for evaluating the merits of clustering algorithms, to describe the performance of the K-means model, Gaussian mixture model, and Bayesian variational Gaussian mixture model, and selecting the model with the best performance in the clustering algorithm indicators to obtain the optimal user behavior state identification model.
[0014] The Davies-Bouldin index is a metric used to evaluate the performance of clustering algorithms, and is expressed as:
[0015]
[0016] Where n is the number of data points, c i It is the center of the i-th cluster, D i d(c) is the average distance from all data points in the i-th cluster to the center. i c j ) is the distance between the centers of the i-th cluster and the j-th cluster;
[0017] The Homogeneity Score is a metric used to evaluate the performance of clustering algorithms, expressed as:
[0018]
[0019] Where C is the true category label, K is the clustering result, H(C) represents the entropy of the category label, and H(C|K) represents the conditional entropy of the category label given the clustering result;
[0020] The Completeness Score is a metric used to evaluate the performance of clustering algorithms, and is expressed as follows:
[0021]
[0022] Where C is the true category label, K is the clustering result, H(K) represents the entropy of the clustering result, and H(K|C) represents the conditional entropy of the clustering result given the true category label.
[0023] As a preferred embodiment of the critical path identification method for the loss of conventional public transport users described in this invention, the user behavior state identification model further includes calculating the maximum and minimum values of each feature in the spatiotemporal feature set of conventional public transport users' travel.
[0024] The original data for the desired features is transformed to the range [0, 1] to obtain the normalized feature data.
[0025]
[0026] Where xk represents the k-th feature of the original data, This represents the maximum value of the k-th feature in the original data. This represents the minimum value of the k-th feature in the original data.
[0027] As a preferred embodiment of the conventional public transport user churn critical path identification method described in this invention, the transition matrix calculation method includes:
[0028] Based on the travel characteristics of regular public transport users within a unit time period, and using an optimal model, the travel behavior state of each user within each unit time period t is identified, and the different states of public transport users during their travel process are abstracted into state nodes s1, s2, ..., s z Then we can obtain the behavioral state of user u in time period t, Y(u,T) = s z , z∈1,...,Z, T∈1,...,T, where z is any behavioral state, Z is the maximum behavioral state, and T is the maximum time;
[0029] Based on the Markov property of the state transition process of public transport users' behavior, the transition probabilities between states within adjacent time intervals t-1 and t are calculated:
[0030]
[0031] The `card` function represents the number of elements in a finite set, specifically the number of regular public transport users in the set, s. z1 s represents the behavioral state of user u at time t-1. z2 This represents the behavioral state of user u at time t;
[0032] Obtain the transition probability between any consecutive unit time intervals between any behavioral states;
[0033] Calculate the mean of the transition probabilities across all consecutive time intervals:
[0034]
[0035] Based on the analysis results, a state transition matrix is constructed using the behavior of the state nodes, with the state nodes as rows and columns in the matrix and the transition probabilities as elements to fill the matrix.
[0036] As a preferred embodiment of the conventional public transport user churn critical path identification method described in this invention, the critical path includes the following steps:
[0037] Calculate the user behavior state transition path graph model: Based on the user behavior state transition matrix, each state is used as a node in the transition path graph model, and the elements with a probability value greater than zero in the transition matrix are used as edges in the transition path graph model. The user behavior state transition path graph model is obtained based on the directed graph.
[0038] Selecting the critical path: Based on the path probability, select the path with the highest transition probability for each behavior state as the critical path, and the state with the highest transition probability as the exit state for that behavior state. The selected path is the critical path in the user behavior state transition probability matrix.
[0039] Extracting critical churn paths: Based on the obtained critical path graph, extract the behavioral states and paths required for each behavioral state to reach the churn state, and obtain the critical churn paths for each behavioral state.
[0040] Another objective of this invention is to provide a critical path identification system for bus user churn based on a user behavior state transition matrix. This system can identify critical paths for bus user churn by establishing a user behavior state transition matrix, revealing the behavioral change mechanism in the process of passenger flow churn in urban regular buses, solving the problem of passenger flow churn in the urban public transportation sector, and simultaneously optimizing bus routes, reducing bus operating costs, and improving the service quality and efficiency of buses.
[0041] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a critical path identification system for conventional public transportation user churn, comprising a data acquisition module, a data preprocessing module, a user behavior status identification module, a transition matrix construction module, a critical path analysis module, and a result display module;
[0042] The data acquisition module is used to collect user card swiping data and city POI data in the urban public transportation system.
[0043] The data preprocessing module is used to clean, integrate, and normalize the collected data.
[0044] The user behavior state recognition module is used to construct a user behavior state recognition model and recognize the behavior state of bus users in different time periods.
[0045] The transition matrix construction module is used to construct a user behavior state transition matrix within adjacent time periods based on the user behavior state.
[0046] The critical path analysis module is used to analyze the relationship between user behavior states through the user behavior state transition matrix, calculate the churn probability corresponding to each state, and identify the critical path of bus passenger flow loss.
[0047] The results display module is used to present the analysis results of the critical path of loss in the form of reports and charts.
[0048] A computer device includes a memory and a processor, wherein the memory stores a computer program, characterized in that the processor executes the computer program to implement the steps of a method and system for identifying critical paths to passenger churn in conventional public transportation.
[0049] A computer-readable storage medium having a computer program stored thereon, characterized in that, when the computer program is executed by a processor, it implements the steps of a method and system for identifying critical paths to passenger churn in conventional public transportation.
[0050] The beneficial effects of this invention are as follows: By providing a method for identifying critical paths leading to passenger churn in conventional public transportation, the causes and critical paths resulting from passenger loss can be effectively identified. Consequently, public transportation operators can optimize bus routes and improve service quality, thereby enhancing the overall service quality and efficiency of public transportation. Simultaneously, reducing passenger churn can lower public transportation operating costs. Furthermore, this method can be applied not only to public transportation but also to other urban public transportation sectors, such as subways and trams. This will further promote the development and improvement of urban public transportation services, encourage more people to use public transportation, reduce private car use, thereby reducing traffic congestion and environmental pollution, and promoting sustainable urban development. Therefore, this invention has significant social and economic implications and makes a positive contribution to the sustainable development of urban public transportation. Attached Figure Description
[0051] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:
[0052] Figure 1 A basic flowchart illustrating a method for identifying critical paths to passenger churn in conventional public transportation, provided as an embodiment of the present invention;
[0053] Figure 2 This is an overall structural diagram of a conventional public transport user churn critical path identification system provided in one embodiment of the present invention;
[0054] Figure 3 The feature set required for constructing a user behavior state recognition model for a conventional public transport user churn critical path identification method provided in an embodiment of the present invention;
[0055] Figure 4 A graph showing the change in the proportion of user transfers between different behavioral states within adjacent time periods, provided as an embodiment of the present invention, for a method for identifying critical paths of user churn in conventional public transportation.
[0056] Figure 5 This is a schematic diagram of the user behavior state probability transition matrix of a conventional public transportation user churn critical path identification method provided in an embodiment of the present invention;
[0057] Figure 6 This is a user behavior state transition critical path model diagram of a conventional public transportation user churn critical path identification method provided in one embodiment of the present invention;
[0058] Figure 7This is a schematic diagram of the critical churn path of a conventional bus user, provided as an embodiment of the present invention, for identifying the critical path of conventional bus user churn. Detailed Implementation
[0059] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.
[0060] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.
[0061] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.
[0062] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.
[0063] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.
[0064] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.
[0065] Example 1
[0066] Reference Figure 1 As an embodiment of the present invention, a method for identifying critical paths to passenger churn in conventional public transportation is provided, comprising:
[0067] A behavioral status recognition model for urban regular public transport users based on multiple mathematical models is constructed to identify the behavioral status of public transport users in different time periods.
[0068] Based on the behavioral state of public transport users, a user behavior state transition matrix is established within adjacent time periods.
[0069] By analyzing the matrix, the relationships between user behavior states are identified, and the churn probability corresponding to each state is calculated, thus identifying the key paths for public transport passenger loss.
[0070] S1: Construct a behavior status recognition model for urban regular public transport users based on multiple mathematical models to identify the behavior status of public transport users in different time periods;
[0071] Furthermore, the user behavior state recognition model includes:
[0072] The user behavior state recognition model includes extracting the spatiotemporal features of the travel behavior of regular urban public transport users based on the card swiping data of public transport users and urban POI data, and modeling based on K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model, and extracting the optimal model.
[0073] Furthermore, the user behavior state recognition model also includes,
[0074] The spatiotemporal characteristics of regular bus users' travel within the unit analysis period required for modeling are normalized. The normalized features are used as the feature set X for model training. The features of the same user in different analysis periods are used as different samples. Then, all regular bus user feature sets are input into the K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model for clustering to obtain the user behavior state recognition results.
[0075] Furthermore, the user behavior state recognition model also includes using the Davies-Bouldin index, Homogeneity Score and Completeness Score, which are indicators for evaluating the merits of clustering algorithms, to describe the performance of the K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model, and selecting the model with the best performance in the clustering algorithm indicators to obtain the optimal user behavior state recognition model.
[0076] The Davies-Bouldin index is a metric used to evaluate the performance of clustering algorithms, and is expressed as:
[0077]
[0078] Where n is the number of data points, c i It is the center of the i-th cluster, D i d(c) is the average distance from all data points in the i-th cluster to the center. i c j ) is the distance between the centers of the i-th cluster and the j-th cluster;
[0079] It should be noted that the DBI index means that among all clusters, the two least similar clusters are selected. The smaller the similarity between these two clusters, the smaller the DBI value, indicating a better clustering effect.
[0080] The Homogeneity Score, used to evaluate the performance of clustering algorithms, is expressed as:
[0081]
[0082] Where C is the true category label, K is the clustering result, H(C) represents the entropy of the category label, and H(C|K) represents the conditional entropy of the category label given the clustering result;
[0083] It should be noted that the meaning of Homogeneity Score is that if each cluster contains only data points belonging to the same category, then the value of Homogeneity Score is 1, indicating that the clustering effect is better.
[0084] The Completeness Score is a metric used to evaluate the performance of clustering algorithms, and is expressed as follows:
[0085]
[0086] Where C is the true category label, K is the clustering result, H(K) represents the entropy of the clustering result, and H(K|C) represents the conditional entropy of the clustering result given the true category label;
[0087] It should be noted that the Completeness Score means that if each category contains only data points belonging to the same cluster, then the Completeness Score is 1, indicating that the clustering effect is better.
[0088] Furthermore, the user behavior state recognition model also includes calculating the maximum and minimum values of each feature in the spatiotemporal feature set of regular public transport users' travel.
[0089] The original data for the desired features is transformed to the range [0, 1] to obtain the normalized feature data.
[0090]
[0091] Where xk represents the k-th feature of the original data, This represents the maximum value of the k-th feature in the original data. This represents the minimum value of the k-th feature in the original data.
[0092] S2: Based on the behavioral state of public transport users, establish a user behavior state transition matrix within adjacent time periods;
[0093] Furthermore, the method for calculating the transition matrix includes:
[0094] Based on the travel characteristics of regular public transport users within a unit time period, and using an optimal model, the travel behavior state of each user within each unit time period t is identified, and the different states of public transport users during their travel process are abstracted into state nodes s1, s2, ..., s z Then we can obtain the behavioral state of user u in time period t, Y(u,t) = s z , z∈1,...,Z, t∈1,...,T, where z is any behavioral state, Z is the maximum behavioral state, and T is the maximum time;
[0095] Based on the Markov property of the state transition process of public transport users' behavior, the transition probabilities between states within adjacent time intervals t-1 and t are calculated:
[0096]
[0097] The `card` function represents the number of elements in a finite set, specifically the number of regular public transport users in the set, s. z1 s represents the behavioral state of user u at time t-1. z2 This represents the behavioral state of user u at time t;
[0098] Obtain the transition probability between any consecutive unit time intervals between any behavioral states;
[0099] Calculate the mean of the transition probabilities across all consecutive time intervals:
[0100]
[0101] Based on the analysis results, a user behavior state transition matrix is constructed, with state nodes as rows and columns in the matrix and transition probabilities as elements to fill the matrix.
[0102] S3: By analyzing the matrix, the relationships between user behavior states are identified, and the churn probability corresponding to each state is calculated, thus identifying the key paths of bus passenger churn.
[0103] Furthermore, the critical path includes the following steps:
[0104] Calculate the user behavior state transition path graph model: Based on the user behavior state transition matrix, each state is used as a node in the transition path graph model, and the elements with a probability value greater than zero in the transition matrix are used as edges in the transition path graph model. The user behavior state transition path graph model is obtained based on the directed graph.
[0105] Selecting the critical path: Based on the path probability, select the path with the highest transition probability for each behavior state as the critical path, and the state with the highest transition probability as the exit state for that behavior state. The selected path is the critical path in the user behavior state transition probability matrix.
[0106] Extracting critical churn paths: Based on the obtained critical path graph, extract the behavioral states and paths required for each behavioral state to reach the churn state, and obtain the critical churn paths for each behavioral state.
[0107] Example 2
[0108] The second embodiment of the present invention differs from the first embodiment in that:
[0109] If the aforementioned functions are implemented as 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 invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0110] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.
[0111] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.
[0112] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0113] Example 3
[0114] Reference Figure 2 This is the third embodiment of the present invention. This embodiment provides a public transport user churn critical path identification system based on user behavior state transition matrix, including: a data acquisition module, a data preprocessing module, a user behavior state identification module, a transition matrix construction module, a critical path analysis module, and a result display module;
[0115] The data acquisition module is used to collect user card swiping data and city POI data in the urban public transportation system.
[0116] The data preprocessing module is used to clean, integrate, and normalize the collected data.
[0117] The user behavior state recognition module is used to construct a user behavior state recognition model and recognize the behavior state of bus users in different time periods.
[0118] The transition matrix construction module is used to construct a user behavior state transition matrix within adjacent time periods based on the user behavior state.
[0119] The critical path analysis module is used to analyze the relationship between user behavior states through the user behavior state transition matrix, calculate the churn probability corresponding to each state, and identify the critical path of bus passenger flow loss.
[0120] The results display module is used to present the analysis results of the critical path of loss in the form of reports and charts.
[0121] Example 4
[0122] Reference Figures 3-7 This is the fourth embodiment of the present invention, which provides a method and system for identifying critical paths to the loss of conventional public transport users. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through economic benefit calculation and experiments.
[0123] The following is an embodiment of the present invention, which provides a method and system for identifying critical paths to passenger churn in conventional public transportation, including:
[0124] In this embodiment, the selected research data are regular bus card swiping data in Shenzhen from January 2014 to October 2015, and Shenzhen's POI data in 2015.
[0125] First, a continuous analysis period of 6 months was selected. Based on this 6-month analysis period, spatiotemporal features of urban regular public transport users' travel behavior required for user modeling were extracted. The total number of users was 15.5 million, and the total sample size for user behavior state identification was 141 million. For detailed features, please refer to [link to details]. Figure 3 .
[0126] After normalizing the obtained feature values, they are used as model inputs. K-means model, Gaussian mixture model (GMM) model and Bayesian variational Gaussian mixture model (BVIGMM) model are used to model the travel behavior of urban regular public transport users.
[0127] After the model was built, the Davies-Bouldin index, Homogeneity Score, and Completeness Score were used to describe the model's performance in order to obtain the optimal model. The performance of the K-means model, Gaussian Mixture Model (GMM) model, and Bayesian Variational Gaussian Mixture Model (BVIGMM) is shown in Table 1. It can be found that the BVIGMM model has higher values on the evaluation metrics than the K-means model and the Gaussian Mixture Model (GMM). Therefore, the model trained by the BVIGMM model was selected as the optimal model and input into the calculation of the user behavior state transition probability matrix.
[0128] Table 1. Training results of K-means model, Gaussian Mixture Model (GMM) model, and Bayesian Variational Gaussian Mixture Model (BVIGMM).
[0129]
[0130]
[0131] The trained urban public transport user travel behavior status recognition model is used to determine the user behavior status of each user within each unit time period. Here, user behavior status is set into a total of 7 categories, including "new entrant" and "churned user," and the changing trend of a user's behavior status category between two adjacent time periods can be obtained. See the appendix for details. Figure 4 .
[0132] Based on the above analysis, it can be calculated that Figure 4 The transition probability between any two consecutive time periods is calculated. The average of these transition probabilities across different consecutive time periods yields the user behavior state probability transition matrix for a 6-month analysis period, as shown below. Figure 5 As shown. Here, without loss of generality, any duration can be chosen as the length of the unit analysis period, therefore Figure 5 The results of the user behavior state probability transition matrices for analysis periods of January and March are also presented. The similarity of the three matrices indicates that the length of the analysis period does not have a significant impact on the user behavior state probability transition matrices.
[0133] Based on the obtained probability transition matrix, a critical transition path graph model of user behavior state is calculated, as detailed in the attached figure. Figure 6 As shown, the critical transition path graph models for user behavior states are illustrated, considering only the maximum transition probability and the maximum and second-largest transition probabilities, respectively. Based on this, the critical churn paths from each initial behavior state to the user churn state can be separated, such as... Figure 7 As shown.
[0134] The results show that this method can not only quantify the travel behavior characteristics of public transport users, avoiding the subjectivity and uncertainty of traditional survey methods, but also be adjusted according to specific circumstances, thus possessing a certain degree of adaptability and flexibility.
[0135] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.
Claims
1. A method for identifying critical paths to passenger churn in conventional public transportation, characterized in that: include, A behavioral status recognition model for urban regular public transport users based on multiple mathematical models is constructed to identify the behavioral status of public transport users in different time periods. Based on the behavioral state of public transport users, a user behavior state transition matrix is established within adjacent time periods. By analyzing the matrix, the relationships between user behavior states are identified, the churn probability corresponding to each state is calculated, and the key paths of bus passenger flow loss are identified. The method for calculating the user behavior state transition matrix includes the following specific steps: Based on the travel characteristics of regular public transport users within a given time period, and using the optimal model, we can identify each user's travel patterns within each time period. t Within the context of travel behavior, different states during a public transport user's journey are abstracted into state nodes. ， Then the user can be obtained u exist t Behavioral state during a period of time in For any behavioral state, It is the largest behavioral state. For the maximum time; Based on the Markov property of the public transport user behavior state transition process, the calculation of adjacent unit time periods is performed. t-1 and t Transition probabilities between states: in, The function represents the number of elements in a finite set, here referring to the number of regular public transport users in the set. Indicates user u The behavioral state at time t-1 Indicates user u The behavioral state at time t; Obtain the transition probability between any consecutive unit time intervals between any behavioral states; Calculate the mean of the transition probabilities across all consecutive time intervals: Based on the analysis results, a user behavior state transition matrix is constructed, with state nodes as rows and columns in the matrix and transition probabilities as elements to fill the matrix. The critical path includes the following specific steps. Calculate the user behavior state transition path graph model. Based on the user behavior state transition matrix, each state is used as a node in the transition path graph model, and the elements with a probability value greater than zero in the transition matrix are used as edges in the transition path graph model. The user behavior state transition path graph model is obtained based on the directed graph. Select the critical path: Based on the path probability, select the path with the highest transition probability for each behavior state as the critical path, and the state with the highest transition probability as the transition state for that behavior state. The selected path is the critical path in the user behavior state transition probability matrix. Extract key churn paths. Based on the obtained key path graph, extract the behavioral states and paths required for each behavioral state to reach the churn state, and obtain the key churn paths for each behavioral state.
2. The method for identifying critical paths to passenger churn in conventional public transportation as described in claim 1, characterized in that: The user behavior state recognition model includes extracting the spatiotemporal features of the travel behavior of regular urban public transport users based on the card swiping data of public transport users and urban POI data, and modeling based on K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model, and extracting the optimal model.
3. The method for identifying critical paths to passenger churn in conventional public transportation as described in claim 2, characterized in that: The user behavior state recognition model also includes normalizing the spatiotemporal features of regular bus users' travel within the unit analysis period required for modeling, using the normalized features as the feature set X for model training, using the features of the same user in different analysis periods as different samples, and then inputting all regular bus user feature sets into the K-means model, Gaussian mixture model and Bayesian variational Gaussian mixture model for clustering to obtain the user behavior state recognition result.
4. The method for identifying critical paths to passenger churn in conventional public transportation as described in claim 3, characterized in that: The user behavior state recognition model further includes using the Davies-Bouldin index, Homogeneity Score, and Completeness Score—indices for evaluating the merits of clustering algorithms—to describe the performance of the K-means model, Gaussian mixture model, and Bayesian variational Gaussian mixture model. The model with the best performance in the clustering algorithm indicators is selected to obtain the optimal user behavior state recognition model. The Davies-Bouldin index is a metric used to evaluate the performance of clustering algorithms, and is expressed as: Where n is the number of data points. It is the center of the i-th cluster. It is the average distance from all data points in the i-th cluster to the center. It is the distance between the centers of the i-th cluster and the j-th cluster; The Homogeneity Score is a metric used to evaluate the performance of clustering algorithms, and is expressed as: Where C is the true category label, and K is the clustering result. The entropy of the category label. The conditional entropy represents the category labels given the clustering results. The Completeness Score is a metric used to evaluate the performance of clustering algorithms, and is expressed as: Where C is the true category label, and K is the clustering result. Entropy represents the clustering result. The conditional entropy represents the clustering result given the true class labels.
5. The method for identifying critical paths to passenger churn in conventional public transportation as described in claim 4, characterized in that: The user behavior state recognition model also includes calculating the maximum and minimum values of each feature in the spatiotemporal feature set of regular public transport users' travel. Transform the raw data of the desired features to The range is used to obtain normalized feature data. , represented as: in, This represents the k-th feature of the original data. This represents the maximum value of the k-th feature in the original data. This represents the minimum value of the k-th feature in the original data.
6. A system employing the critical path identification method for conventional public transport user churn as described in any one of claims 1 to 5, characterized in that: It includes a data acquisition module, a data preprocessing module, a user behavior status recognition module, a transition matrix construction module, a critical path analysis module, and a results display module; The data acquisition module is used to collect user card swiping data and city POI data in the urban public transportation system; The data preprocessing module is used to clean, integrate, and normalize the collected data; The user behavior state recognition module is used to construct a user behavior state recognition model and identify the behavior state of bus users in different time periods. The transition matrix construction module is used to construct a user behavior state transition matrix within adjacent time periods based on the user behavior state. The critical path analysis module is used to analyze the relationship between user behavior states through the user behavior state transition matrix, calculate the churn probability corresponding to each state, and identify the critical paths for public transport passenger loss. The results display module is used to present the analysis results of the critical path of loss in the form of reports and charts.
7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 5.
8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 5.