Railway ticket recommendation method and system

Abstract

The invention discloses a railway passenger ticket recommendation method and system, and the method comprises the steps: building and updating a user portrait model and a railway passenger ticket feature model through the data information on a railway passenger ticket purchase platform; based on the user portrait model and the railway passenger ticket feature model, obtaining a matching degree between the user and each railway passenger ticket; and based on the matching degree between the user and each railway ticket, obtaining a recommended railway ticket list corresponding to each user, and pushing the recommended railway ticket list. According to the demands and preferences of the user, appropriate railway ticket recommendation is provided for the user, the satisfaction degree and purchase willingness of the user are improved, and meanwhile the efficiency and income of railway transportation are improved.

Description

Technical Field

[0001] The present invention relates to the field of rail transit technology, and in particular to a railway passenger ticket recommendation method and system. Background Art

[0002] With the development of railway transportation, the types and quantity of railway tickets are increasing. Users are faced with the problem of information overload when choosing and purchasing railway tickets. How to provide users with appropriate railway ticket recommendations based on their personalized needs and preferences is an important and challenging problem.

[0003] Traditional railway ticket recommendation systems usually use content-based recommendation methods or collaborative filtering-based recommendation methods. However, these recommendation methods have some defects; for example, content-based recommendation methods ignore the diversity and dynamics of users and cannot adapt to changes in user needs and preferences; collaborative filtering-based recommendation methods face the problems of data sparsity and cold start, and cannot provide effective recommendations for new users or new railway tickets; in addition, these recommendation methods do not take into account the different needs of users in different scenarios, such as travel purpose, time, budget and other factors. Summary of the invention

[0004] The present invention aims to provide a railway ticket recommendation method and system, which provides users with suitable railway ticket recommendations based on their needs and preferences, improves user satisfaction and purchase intention, and improves the efficiency and revenue of railway transportation.

[0005] To achieve the above object, an embodiment of the present invention provides a railway ticket recommendation method, comprising:

[0006] Using the data information on the railway ticket purchasing platform, the user portrait model and railway ticket feature model are constructed and updated respectively;

[0007] Based on the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is obtained;

[0008] Based on the matching degree between the user and each railway ticket, a list of recommended railway tickets corresponding to each user is obtained and pushed.

[0009] Furthermore, the data information on the railway ticket purchasing platform is used to build a user portrait model, including:

[0010] Preprocessing user data on the railway ticket purchasing platform; wherein the user data includes user registration information, user historical behavior information and user current information;

[0011] Extracting feature data from user registration information and user historical behavior information as static portrait features of the user; reducing the dimension of the static portrait features using a principal component analysis method to extract the features that best represent the static portrait and combine them into a low-dimensional vector, i.e., obtaining a static portrait vector of the user;

[0012] Extract feature data from the user's current information as the user's dynamic portrait feature; use the hidden Markov model to capture the temporal regularity between the user's dynamic portrait features and obtain the hidden state sequence of the user's dynamic portrait; perform feature extraction on each dynamic portrait feature according to the hidden state sequence and combine them into a low-dimensional vector, that is, obtain the user's dynamic portrait vector;

[0013] Among them, the user portrait model includes static portrait and dynamic portrait, and each user corresponds to a static portrait vector and a dynamic portrait vector.

[0014] Furthermore, the user registration information includes: gender, age, occupation, income, permanent residence, and preference questionnaire results;

[0015] The user's historical behavior information includes: browsing history, search history and purchase history;

[0016] The user's current information includes: travel purpose, time, budget and destination;

[0017] The user's current information refers to the current information provided or generated by the user on the railway ticket purchasing platform.

[0018] Furthermore, the data information on the railway ticket purchase platform is used to construct a railway ticket feature model, including:

[0019] The railway ticket feature model includes content features and context features of the railway ticket, and each railway ticket corresponds to a content feature vector and a context feature vector;

[0020] Preprocessing the railway ticket data on the railway ticket purchasing platform; wherein the railway ticket data includes ticket purchasing information and external information;

[0021] Extracting feature data from the ticket purchase information as content features of the railway ticket; reducing the dimension of the content features using a principal component analysis method to extract features that best represent the content features and combining them into a low-dimensional vector, i.e., obtaining a content feature vector of the railway ticket;

[0022] Extract feature data from the external information as context features of the railway ticket; use the frequent pattern mining method to capture the sequence regularity between each context feature of the railway ticket, and perform feature extraction on each context feature and combine it into a low-dimensional vector, that is, obtain the context feature vector of the railway ticket.

[0023] Furthermore, the ticket purchase information includes: train number, departure place, destination, departure time, arrival time, seat type, and price;

[0024] The external information includes: seasons, holidays, delays, and passenger flow;

[0025] The external information refers to the external information generated or associated with the railway ticket on the railway ticket purchasing platform.

[0026] Furthermore, based on the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is obtained, including:

[0027] Based on the collaborative filtering algorithm, the implicit relationship between the user portrait model and the railway ticket feature model is learned by using the interest questionnaire when the user registers on the railway ticket purchase platform, as well as the railway ticket rating feedback data, search records and purchase records, and the implicit relationship is represented as a low-dimensional matrix;

[0028] According to the implicit relationship between the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is calculated and stored in the matching degree matrix;

[0029] Among them, the matching degree between the user and the railway ticket is calculated by the following formula:

[0030] Matching degree = static weight * static similarity + dynamic weight * dynamic similarity,

[0031] Among them, the static similarity represents the cosine similarity between the user's static portrait vector and the content feature vector of the railway ticket, the dynamic similarity represents the cosine similarity between the user's dynamic portrait vector and the context feature vector of the railway ticket, the static weight represents the weight ratio of the user's static portrait vector in the user portrait model, and the dynamic weight represents the weight ratio of the user's dynamic portrait vector in the user portrait model.

[0032] Furthermore, based on the matching degree between the user and each railway ticket, a list of recommended railway tickets corresponding to each user is obtained and pushed, including:

[0033] Based on the matching degree matrix, railway tickets that the user has not purchased and whose matching degree values ​​are higher than a predetermined threshold are collected to generate a candidate railway ticket set corresponding to the user;

[0034] The railway tickets in the candidate railway ticket set are sorted from high to low according to the matching value, the first N railway tickets are extracted to form a recommended railway ticket list corresponding to the user, and the recommended railway ticket list is displayed and pushed to the user.

[0035] Based on the same inventive concept, an embodiment of the present invention further provides a railway ticket recommendation system, comprising:

[0036] Establishing a unit for using data information on the railway ticket purchasing platform to construct and update the user portrait model and the railway ticket feature model respectively;

[0037] An obtaining unit, used for obtaining the matching degree between the user and each railway ticket based on the user portrait model and the railway ticket feature model;

[0038] The push unit is used to obtain a list of recommended railway tickets corresponding to each user and push the list based on the matching degree between the user and each railway ticket.

[0039] Based on the same inventive concept, an embodiment of the present invention also provides an electronic device, including: a memory, a processor; the processor is used to read and execute a computer program stored in the memory to implement the aforementioned railway ticket recommendation method.

[0040] Based on the same inventive concept, an embodiment of the present invention further provides a computer storage medium, in which computer executable instructions are stored, and when the computer executable instructions are executed, the aforementioned railway ticket recommendation method is implemented.

[0041] The technical effects and advantages of the present invention are as follows: the present invention can comprehensively consider various characteristics and needs of users, provide users with appropriate railway ticket recommendations, improve user satisfaction and purchase intention, and at the same time improve the efficiency and benefits of railway transportation; the present invention adopts an algorithm based on collaborative filtering, which can utilize the implicit relationship between users and railway tickets to improve the intelligence and flexibility of recommendations, and solve the cold start problem faced by collaborative filtering by means of questionnaires for new users; it also adopts a context-based recommendation method, which can adapt to different needs in different scenarios and improve the timeliness and practicality of recommendations.

[0042] Other features and advantages of the present invention will be described in the following description, and partly become apparent from the description, or understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained by the structures pointed out in the description, claims and drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 A flow chart of a railway ticket recommendation method according to an embodiment of the present invention;

[0045] Figure 2 A detailed flow chart of a railway ticket recommendation method according to an embodiment of the present invention;

[0046] Figure 3 This is a schematic diagram of the structure of a railway ticket recommendation system according to an embodiment of the present invention;

[0047] Figure 4 The figure is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0049] In order to solve the deficiencies of the prior art, the present invention discloses a railway ticket recommendation method, such as Figure 1 and Figure 2 As shown, the following steps are included:

[0050] Step S1: Using the data information on the railway ticket purchasing platform, the user portrait model and the railway ticket feature model are constructed and updated respectively.

[0051] In some specific embodiments, the data information on the railway ticket purchasing platform is used to construct a user portrait model, including:

[0052] The user portrait model is a model that describes user characteristics and needs. The user portrait model includes static portraits and dynamic portraits. Each user corresponds to a static portrait vector and a dynamic portrait vector.

[0053] Among them, static portraits refer to characteristics that reflect users' inherent attributes and long-term preferences, such as gender, age, occupation, income, permanent residence, etc.; dynamic portraits refer to characteristics that reflect users' current needs and short-term preferences, such as travel purpose, time, budget, destination, etc.

[0054] Preprocess user data on the railway ticket purchasing platform to improve data quality and availability.

[0055] Among them, the preprocessing includes: data cleaning, data conversion and data normalization and other operations, and the user data includes user registration information, user historical behavior information and user current information.

[0056] User registration information includes: gender, age, occupation, income, permanent residence, and preference questionnaire results; user historical behavior information includes: browsing history, search history, and purchase history; user current information includes: travel purpose, time, budget, and destination; among them, user current information refers to the current information provided or generated by the user on the railway ticket purchase platform.

[0057] Extract feature data from user registration information and user historical behavior information as static profile features of users and convert them into numerical variables; for example, you can use the railway ticket information that users have browsed or purchased, such as train number, departure place, destination, departure time, arrival time, seat type, price, etc., to reflect users' long-term preferences.

[0058] The static portrait features are reduced in dimension using a principal component analysis method to extract the features of the static portrait that best represent the static portrait and combine them into a low-dimensional vector, that is, to obtain the user's static portrait vector.

[0059] Extract feature data from the user's current information as the user's dynamic portrait features; use the hidden Markov model to capture the temporal regularity between the user's dynamic portrait features and obtain the hidden state sequence of the user's dynamic portrait; perform feature extraction on each dynamic portrait feature according to the hidden state sequence and combine them into a low-dimensional vector, that is, obtain the user's dynamic portrait vector.

[0060] After obtaining a static portrait vector and a dynamic portrait vector corresponding to each user, they are stored in a user portrait database.

[0061] In some specific embodiments, the railway ticket feature model is constructed by using the data information on the railway ticket purchase platform, including:

[0062] The railway ticket feature model is a model that represents the attributes and value of railway tickets. The railway ticket feature model includes: content features and context features; each railway ticket corresponds to a content feature vector and a context feature vector.

[0063] Among them, content features refer to features that reflect the attributes and value of the railway ticket itself, such as train number, departure place, destination, departure time, arrival time, seat type, price, etc.; context features refer to features that reflect the correlation between the railway ticket and the external environment, such as season, holidays, weather, traffic conditions, etc.

[0064] The railway ticket data on the railway ticket purchasing platform is preprocessed to improve the quality and availability of the data; wherein the preprocessing includes operations such as data cleaning, data conversion and data normalization, and the railway ticket data includes ticket purchase information and external information.

[0065] Ticket purchase information includes: train number, departure place, destination, departure time, arrival time, seat type, and price; external information includes: season, holidays, delays, and passenger flow; among them, external information refers to external information generated or associated with railway tickets on the railway ticket purchase platform.

[0066] Extract characteristic data from the ticket purchase information as content characteristics of the railway ticket; use the principal component analysis method to reduce the dimension of the content characteristics to extract the features that best represent the content characteristics and combine them into a low-dimensional vector, that is, obtain the content characteristic vector of the railway ticket.

[0067] Extract feature data from the external information as context features of the railway ticket; use the frequent pattern mining method to capture the sequence regularity between each context feature of the railway ticket, and perform feature extraction on each context feature and combine it into a low-dimensional vector, that is, obtain the context feature vector of the railway ticket.

[0068] After obtaining a content feature vector and a context feature vector corresponding to each railway ticket, they are stored in a railway ticket feature database.

[0069] Step S2: Based on the user portrait model and the railway ticket feature model, obtain the matching degree between the user and each railway ticket; including:

[0070] Based on a collaborative filtering algorithm based on matrix decomposition, the implicit relationship between the user portrait model and the railway ticket feature model (i.e., the two models) is learned by using the interest questionnaire when the user registers on the railway ticket purchase platform, as well as the rating feedback data, search records and purchase records of the railway ticket, and the implicit relationship is expressed as a low-dimensional matrix. The collaborative filtering algorithm can also optimize the parameters of the two models by minimizing the mean square error between the feature matrix and the low-dimensional matrix.

[0071] According to the implicit relationship between the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is calculated and stored in the matching degree matrix; wherein, the matching degree between the user and the railway ticket is calculated by the following formula:

[0072] Matching degree = static weight * static similarity + dynamic weight * dynamic similarity,

[0073] Among them, the static similarity represents the cosine similarity between the user's static portrait vector and the content feature vector of the railway ticket, the dynamic similarity represents the cosine similarity between the user's dynamic portrait vector and the context feature vector of the railway ticket, the static weight represents the weight ratio of the user's static portrait vector in the user portrait model, and the dynamic weight represents the weight ratio of the user's dynamic portrait vector in the user portrait model.

[0074] Step S3: based on the matching degree between the user and each railway ticket, obtain a list of recommended railway tickets corresponding to each user and push it; including:

[0075] Based on the matching degree matrix, railway tickets that the user has not purchased and whose matching degree values ​​are higher than a predetermined threshold are collected to generate a candidate railway ticket set corresponding to the user;

[0076] The railway tickets in the candidate railway ticket set are sorted from high to low according to the matching value, the first N railway tickets are extracted to form a recommended railway ticket list corresponding to the user, and the recommended railway ticket list is displayed and pushed to the user.

[0077] Among them, the recommended railway passenger ticket list = Top-N,

[0078] Here, N represents the length of the recommendation list, which can be determined according to user needs or platform policies.

[0079] Based on the same inventive concept, the present invention also provides a railway ticket recommendation system, such as Figure 3 As shown, including:

[0080] Establishing a unit for using data information on the railway ticket purchasing platform to construct and update the user portrait model and the railway ticket feature model respectively;

[0081] An obtaining unit, used for obtaining the matching degree between the user and each railway ticket based on the user portrait model and the railway ticket feature model;

[0082] The push unit is used to obtain a list of recommended railway tickets corresponding to each user and push the list based on the matching degree between the user and each railway ticket.

[0083] Regarding the system in the above embodiment, the specific manner in which each unit module performs operations has been described in detail in the embodiment of the method, and will not be elaborated here.

[0084] Based on the same inventive concept, an embodiment of the present invention further provides an electronic device, whose structure is as follows: Figure 4 As shown, it includes: a memory and a processor, wherein the processor is used to read and execute the computer program stored in the memory to implement the aforementioned railway ticket recommendation method.

[0085] Based on the same inventive concept, an embodiment of the present invention further provides a computer storage medium, in which computer executable instructions are stored, and when the computer executable instructions are executed, the aforementioned railway ticket recommendation method is implemented.

[0086] Finally, it should be noted that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, it is still possible for those skilled in the art to modify the technical solutions described in the aforementioned embodiments or to make equivalent substitutions for some of the technical features therein. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A railway ticket recommendation method, characterized in that: include: Using the data information on the railway ticket purchasing platform, the user portrait model and railway ticket feature model are constructed and updated respectively; Based on the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is obtained; Based on the matching degree between the user and each railway ticket, a list of recommended railway tickets corresponding to each user is obtained and pushed.

2. A railway ticket recommendation method according to claim 1, characterized in that: Using the data information on the railway ticket purchasing platform, we build a user portrait model, including: Preprocessing user data on the railway ticket purchasing platform; wherein the user data includes user registration information, user historical behavior information and user current information; Extracting feature data from user registration information and user historical behavior information as static portrait features of the user; reducing the dimension of the static portrait features using a principal component analysis method to extract the features that best represent the static portrait and combine them into a low-dimensional vector, i.e., obtaining a static portrait vector of the user; Extract feature data from the user's current information as the user's dynamic portrait feature; use the hidden Markov model to capture the temporal regularity between the user's dynamic portrait features and obtain the hidden state sequence of the user's dynamic portrait; perform feature extraction on each dynamic portrait feature according to the hidden state sequence and combine them into a low-dimensional vector, that is, obtain the user's dynamic portrait vector; Among them, the user portrait model includes static portrait and dynamic portrait, and each user corresponds to a static portrait vector and a dynamic portrait vector.

3. A railway ticket recommendation method according to claim 2, characterized in that: The user registration information includes: gender, age, occupation, income, permanent residence, and preference questionnaire results; The user's historical behavior information includes: browsing history, search history and purchase history; The user's current information includes: travel purpose, time, budget and destination; The user's current information refers to the current information provided or generated by the user on the railway ticket purchasing platform.

4. A railway ticket recommendation method according to claim 2, characterized in that: Using the data information on the railway ticket purchasing platform, a railway ticket feature model is constructed, including: The railway ticket feature model includes content features and context features of the railway ticket, and each railway ticket corresponds to a content feature vector and a context feature vector; Preprocessing the railway ticket data on the railway ticket purchasing platform; wherein the railway ticket data includes ticket purchasing information and external information; Extracting feature data from the ticket purchase information as content features of the railway ticket; reducing the dimension of the content features using a principal component analysis method to extract features that best represent the content features and combining them into a low-dimensional vector, i.e., obtaining a content feature vector of the railway ticket; Extract feature data from the external information as context features of the railway ticket; use the frequent pattern mining method to capture the sequence regularity between each context feature of the railway ticket, and perform feature extraction on each context feature and combine it into a low-dimensional vector, that is, obtain the context feature vector of the railway ticket.

5. A railway ticket recommendation method according to claim 4, characterized in that: The ticket purchase information includes: train number, departure place, destination, departure time, arrival time, seat type, and price; The external information includes: seasons, holidays, delays, and passenger flow; The external information refers to the external information generated or associated with the railway ticket on the railway ticket purchasing platform.

6. A railway ticket recommendation method according to claim 1, characterized in that: Based on the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is obtained, including: Based on the collaborative filtering algorithm, the implicit relationship between the user portrait model and the railway ticket feature model is learned by using the interest questionnaire when the user registers on the railway ticket purchase platform, as well as the railway ticket rating feedback data, search records and purchase records, and the implicit relationship is represented as a low-dimensional matrix; According to the implicit relationship between the user portrait model and the railway ticket feature model, the matching degree between the user and each railway ticket is calculated and stored in the matching degree matrix; Among them, the matching degree between the user and the railway ticket is calculated by the following formula: Matching degree = static weight * static similarity + dynamic weight * dynamic similarity, Among them, the static similarity represents the cosine similarity between the user's static portrait vector and the content feature vector of the railway ticket, the dynamic similarity represents the cosine similarity between the user's dynamic portrait vector and the context feature vector of the railway ticket, the static weight represents the weight ratio of the user's static portrait vector in the user portrait model, and the dynamic weight represents the weight ratio of the user's dynamic portrait vector in the user portrait model.

7. A railway ticket recommendation method according to claim 1 or 6, characterized in that: Based on the matching degree between the user and each railway ticket, a list of recommended railway tickets corresponding to each user is obtained and pushed, including: Based on the matching degree matrix, railway tickets that the user has not purchased and whose matching degree values ​​are higher than a predetermined threshold are collected to generate a candidate railway ticket set corresponding to the user; The railway tickets in the candidate railway ticket set are sorted from high to low according to the matching value, the first N railway tickets are extracted to form a recommended railway ticket list corresponding to the user, and the recommended railway ticket list is displayed and pushed to the user.

8. A railway ticket recommendation system, characterized in that: include: Establishing a unit for using data information on the railway ticket purchasing platform to construct and update the user portrait model and the railway ticket feature model respectively; An obtaining unit, used for obtaining the matching degree between the user and each railway ticket based on the user portrait model and the railway ticket feature model; The push unit is used to obtain a list of recommended railway tickets corresponding to each user and push the list based on the matching degree between the user and each railway ticket.

9. An electronic device, characterized in that: include: Memory, processor; The processor is used to read and execute the computer program stored in the memory to implement a railway ticket recommendation method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-executable instructions, which, when executed, implement a railway ticket recommendation method as described in any one of claims 1-7.

Images