Travel intention recognition method, travel strategy recommendation method and service platform

Abstract

The embodiment of the invention provides a travel intention recognition method, a travel strategy recommendation method and a service platform. The travel intention recognition method comprises the following steps: acquiring current travel information and current passenger information of a current travel, and historical travel information of a current passenger; extracting features from the current travel information, the current passenger information and the historical travel information, and combining the features to obtain to-be-matched combined features corresponding to the current travel in a plurality of travel scenes; matching the to-be-matched combined features of the current travel in each travel scene with the feature requirements of each travel scene, so as to determine the travel scene matched with the current travel from the plurality of travel scenes; wherein the feature requirements of the travel scene conform to the definition and features of the travel scene; and determining the travel intention of the current travel according to the travel scene matched with the current travel. According to the embodiment of the invention, the identification accuracy of the travel intention can be improved, and support is provided for improving the accuracy of travel strategy recommendation.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer Internet technology, and specifically to a travel intention recognition method, a travel strategy recommendation method, and a service platform. Background Art

[0002] Travel service platforms are online platforms that provide travel services to users, such as OTP (Online Travel Platform), OTA (Online Travel Agency) platforms, etc. Travel service platforms are designed to conveniently meet users' travel needs. With the continuous development of intelligent technology, travel service platforms not only provide ticketing services such as booking, purchasing, and changing tickets, but also involve diversified travel solutions such as hotel reservations, travel packages, and car rentals.

[0003] In order to improve users' travel experience, travel service platforms need to accurately identify users' travel intentions. Therefore, how to improve the accuracy of identifying travel intentions has become a technical problem that technical personnel in this field urgently need to solve. Summary of the invention

[0004] In view of this, the embodiments of the present application provide a travel intention recognition method, a travel strategy recommendation method and a service platform to improve the recognition accuracy of travel intention.

[0005] To achieve the above objectives, the embodiments of the present application provide the following technical solutions.

[0006] In a first aspect, an embodiment of the present application provides a method for identifying travel intention, including: Get the current trip information and current passenger information of the current trip, as well as the current passenger's historical travel information; Extracting and combining features from the current trip information, the current passenger information, and the historical travel information to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios; Matching the combined features to be matched of the current trip in each travel scenario with the feature requirements of each travel scenario respectively, so as to determine the travel scenario matched by the current trip from the multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; The travel intention of the current trip is determined according to the travel scenario matched by the current trip.

[0007] In a second aspect, an embodiment of the present application provides a travel strategy recommendation method, comprising: Acquire the travel intention of the current trip, wherein the travel intention of the current trip is determined based on the travel intention identification method described in the first aspect above; Determine the current passenger's travel preference information; A travel strategy for the current trip is recommended based on the travel intention of the current trip and the travel preference information of the current passengers.

[0008] In a third aspect, an embodiment of the present application provides a travel service platform, including: a travel intention recognition system; The travel intention recognition system comprises: A preprocessing module is used to extract and combine features from the current trip information and the current passenger information of the current trip, as well as the historical travel information of the current passenger, to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios; The travel intention recognition model is used to match the combined features to be matched in each travel scenario of the current trip with the feature requirements of each travel scenario, so as to determine the travel scenario that matches the current trip from multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; A post-processing module, used to determine the travel intention of the current trip according to the travel scenario matched by the current trip; The travel scenario knowledge base, as the background knowledge base of the travel intention recognition model, records the background knowledge of multiple travel scenarios, and the background knowledge of the travel scenarios at least includes the definition, characteristics and feature requirements of the travel scenarios.

[0009] In a fourth aspect, an embodiment of the present application provides a travel server, comprising a memory and a processor, wherein the memory stores computer-executable instructions, and the processor calls the computer-executable instructions to execute the travel intention identification method as described in the first aspect above, or the travel strategy recommendation method as described in the second aspect above.

[0010] In a fifth aspect, an embodiment of the present application provides a computer program product, comprising computer execution instructions, which, when executed, implement the travel intention identification method as described in the first aspect above, or the travel strategy recommendation method as described in the second aspect above.

[0011] The travel intention recognition method provided in the embodiment of the present application integrates multi-dimensional information such as current trip information, current passenger information, and historical travel information of the current passenger to match the current trip with the travel scenario; specifically, features are extracted and combined from the current trip information, the current passenger information, and the historical travel information to form the to-be-matched combined features corresponding to the current trip in multiple travel scenarios, so that the to-be-matched combined features corresponding to the travel scenarios of the current trip can reflect the various dimensional information and background of the current trip, avoid the possible deviation of the single dimensional information of the current trip, and ensure that the background of the current trip can be fully considered when matching the travel scenario of the current trip; then, based on the feature requirements of the travel scenario that are consistent with the definition and characteristics of the travel scenario, the to-be-matched combined features of the current trip in each travel scenario are matched with the feature requirements of each travel scenario respectively, so as to obtain the to-be-matched combined features of the current trip in multiple travel scenarios. The travel scenario that matches the current trip is determined from the travel scenarios, so as to accurately determine the travel scenario that matches the current trip; that is, each travel scenario has clear feature requirements, which are consistent with the specific definition and characteristics of the travel scenario. By comparing and matching the combined features to be matched in the current trip in the travel scenario with the feature requirements of the travel scenario, it is possible to more accurately judge whether the current trip meets the requirements of the travel scenario, thereby accurately identifying the travel scenario that matches the current trip from multiple travel scenarios; at the same time, the historical travel information of the current passenger is taken into account in the process of travel intention identification. Since the historical travel information of the current passenger helps to understand the historical travel habits and preferences of the current passenger, the travel scenario that matches the current trip can be made more accurate; furthermore, according to the travel scenario that matches the current trip, the travel intention of the current trip is determined, and the travel intention identification of the current trip can be completed. In summary, the embodiments of the present application can improve the accuracy of travel intention identification and provide support for improving the accuracy of travel strategy recommendations. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0013] Figure 1 A flowchart of a travel intention identification method provided in an embodiment of the present application.

[0014] Figure 2 The following are example pictures of some travel scenarios.

[0015] Figure 3 Example diagram of some of the feature types required for matching some travel scenarios.

[0016] Figure 4 The following are example diagrams of some feature requirements for some travel scenarios.

[0017] Figure 5 This is an example diagram of the composition of the travel scenario knowledge base.

[0018] Figure 6 This is an example diagram of the architecture of the travel service platform provided in an embodiment of the present application.

[0019] Figure 7 This is an example diagram of the verification rules for some travel scenarios recorded in the form of a verification list.

[0020] Figure 8 A flow chart of a travel strategy recommendation method provided in an embodiment of the present application.

[0021] Fig. 9 Another example diagram of the architecture of the travel service platform provided in an embodiment of the present application. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0023] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0024] In the embodiments of the present application, the user's travel intention refers to the motivation and reason for the user's travel, that is, the user's travel intention can describe the reason for the user's travel, such as the user's travel intention describes why the user travels. Identifying the user's travel intention is crucial for the travel service platform, because the user's travel intention can help the travel service platform better understand the user's travel needs, and then provide targeted travel services for the user, such as providing the user with targeted travel strategies.

[0025] Travel strategies refer to travel plans and / or travel additional services that are intelligently formulated and optimized based on the user's travel needs, travel preferences and other information. Specifically, formulating travel strategies may include formulating travel plans, such as itinerary routes when users book or purchase tickets, and recommending transportation schedules corresponding to itinerary routes (such as train schedule recommendations and flight recommendations for airplanes). Formulating travel strategies may not be limited to formulating travel plans, but may also include recommending travel additional services, such as recommending hotel reservation services, car rental services, departure and arrival shuttle services and other travel additional services.

[0026] After determining the user's travel intention, the travel service platform can formulate travel strategies around the user's travel intention and improve the user's travel experience. For example, when the user is booking a ticket, if there are no remaining tickets for the user's desired travel plan (for example, there are no remaining tickets for the train or flight the user expects), the travel service platform can provide alternative travel plans that meet the user's needs (such as alternative trains or flights) based on the user's travel intention; that is, when the user's desired travel plan cannot be met, the travel service platform does not simply inform the user that it is currently unable to book a ticket, nor does it provide simple itinerary transfer information, but instead identifies the travel intention and intelligently recommends alternative travel plans that meet the user's needs, thereby improving the user's travel experience.

[0027] For example, in addition to recommending travel plans to users when they book or purchase tickets, travel service platforms can also recommend other travel supplementary services that meet user needs based on their travel intentions, such as car rental services, hotel reservation services, etc., thereby building a comprehensive travel solution and improving users' travel experience and convenience.

[0028] For example, travel service platforms can help users promptly understand ticket changes, departure time adjustments, and even temporary weather, traffic and other factors affecting travel related to their travel based on their travel intentions through real-time notifications and push services, ensuring that their travel plans proceed smoothly.

[0029] It can be seen that identifying users' travel intentions is the basis for travel service platforms to provide users with accurate travel strategies. It is of great significance to improve user experience and reduce user decision-making costs. Therefore, optimizing the identification method of travel intentions and improving the accuracy of travel intention identification are of great technical significance.

[0030] As an optional implementation, Figure 1An optional flow chart of the travel intention identification method provided in the embodiment of the present application is shown as an example, and the travel intention identification method can be applied to travel service platforms, such as OTP, OTA platform, etc. In an optional implementation, the travel service platform (such as OTP, OTA platform) may include a server on the service side, called a travel server (such as OTP server, OTA server), so that the travel intention identification method provided in the embodiment of the present application can be specifically applied to the travel server; in an optional implementation, the travel service platform (such as OTP, OTA platform) can be regarded as an online service platform formed by a server cluster formed by the travel server by setting a variety of software and hardware functional architectures.

[0031] Reference Figure 1 The travel intention identification method provided in the embodiment of the present application may include the following steps.

[0032] Step S110: Acquire the current itinerary information and current passenger information of the current itinerary, as well as the historical travel information of the current passenger.

[0033] In an optional implementation, the current itinerary is an ongoing or planned itinerary. For example, a user can log in to the travel service platform using a registered account of the travel service platform, so that the user can submit the current itinerary to the travel service platform through travel service requests such as ticketing requests and travel strategy recommendation requests. Among them, a ticketing request refers to any request initiated by a user when performing ticket-related operations, including but not limited to a booking request when the user books a ticket, a ticket inquiry request (such as a request to inquire about train tickets, a request to inquire about air tickets, etc.), a change request (to modify the travel time or flight of purchased train tickets, air tickets, etc.), etc. In a specific example, the itinerary submitted by the user through a ticket booking request can be the current itinerary. The travel strategy recommendation request is used by the user to request the travel service platform to recommend a travel strategy, such as a user requesting the travel service platform to recommend a travel plan and / or travel additional services for the current itinerary.

[0034] The current trip information includes multiple feature information related to the current trip, such as the departure place (such as departure city, departure station), destination (such as destination city, destination station), departure time (such as departure date and departure time), arrival time (such as arrival date and arrival time), travel time, transportation type (such as high-speed rail, motor train, express train, airplane, etc.), transportation schedule (such as train schedule, airplane flight, etc.), ticket price, seat type (such as second-class seat, first-class seat, business seat, etc. of high-speed rail), booking time, ticket type (such as student ticket, children's ticket, adult ticket, etc.), etc. In an optional implementation, the travel service platform can collect the current trip information of the current trip, such as based on the booking request or ticket inquiry request submitted by the user, collecting the information related to the trip in the booking request or ticket inquiry request as the current trip information.

[0035] It should be noted that the users referred to in the embodiments of the present application are users who register and use the travel service platform; the current passenger refers to the actual passenger of the current trip, such as the current passenger, and the current passenger may be the user himself or other people for whom the user has booked tickets; further, the current passenger may also be multiple people, including the user himself and other people for whom the user has booked tickets, or multiple other people for whom the user has booked tickets.

[0036] The current passenger information includes multiple characteristic information related to the current passenger, such as the name, age, identification (such as identification number) of the current passenger, the number of current passengers, etc. In an optional implementation, the travel service platform may collect the current passenger information of the current trip, such as based on a booking request or ticket inquiry request submitted by a user, collecting the passenger-related information in the booking request or ticket inquiry request as the current passenger information.

[0037] After determining the current passenger information, the embodiment of the present application can obtain the historical travel information of the current passenger based on the current passenger information; for example, based on the identity identifier of the current passenger in the current passenger information, the historical travel information of the current passenger is obtained from a historical travel database, wherein the historical travel database records the historical travel information of multiple passengers, such as the itinerary information and corresponding passenger information of multiple passengers' historical trips, and the historical trips are trips that have been carried out. In an optional implementation, the itinerary information and passenger information of the historical trips can be the itinerary information and passenger information of completed historical ticket orders, and the historical travel database can record the itinerary information and passenger information of historical ticket orders completed by multiple users (such as all users of the travel service platform).

[0038] In an optional implementation, the historical travel information of the current passenger may refer to the historical travel record of the current passenger, including the historical travel information corresponding to the historical itinerary of the current passenger, and the travel frequency corresponding to the historical itinerary of the current passenger. For example, the historical travel information corresponding to the historical itinerary of the current passenger may include but is not limited to the departure place, destination, departure time, arrival time, travel time, type of transportation, transportation schedule, fare, seat type, booking time, ticket type, etc. corresponding to the historical itinerary of the current passenger. For example, the travel frequency corresponding to the historical itinerary of the current passenger refers to the number of trips of the historical itinerary of the current passenger within a certain time range, including the travel frequency of the current passenger in the same historical itinerary, and further, it may also include the travel frequency of the current passenger in all historical itineraries; wherein, the same historical itinerary refers to a historical itinerary with the same departure place and destination and the same itinerary route. By counting the number of trips of the current passenger in the same historical itinerary within a certain time range, the travel frequency of the current passenger in the same historical itinerary can be obtained.

[0039] Step S120: extract and combine features from the current trip information, the current passenger information and the historical travel information of the current passenger to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios.

[0040] In the embodiments of the present application, the travel scenario can represent the travel intention, which is used to describe the user's travel purpose or travel motivation. The travel scenario can be understood as a classification of the user's travel needs and behavior patterns, including the major category dimension of the travel scenario type, and the dimension of at least one subcategory subdivided under the travel scenario type. Among them, the travel scenario type is a major classification of travel scenarios, indicating the main categories of travel scenarios, such as commuting, business travel, tourism, etc.; the subcategories under the travel scenario type are further subcategories of the travel scenario type, in order to more accurately describe the user's travel needs under the travel scenario type, such as the commuting travel scenario type can be further subdivided into weekly commuting and daily commuting subcategories. The subcategories subdivided under the travel scenario type can be multi-level subcategories, such as the first-level subcategories and second-level subcategories under the travel scenario type.

[0041] For ease of understanding, Figure 2 The following is an example diagram showing some travel scenarios. Figure 2 Some of the travel scenario types and subcategories, as well as the relevant definitions and characteristics of the examples can be referred to.

[0042] In the embodiment of the present application, multiple travel scenario types can be pre-set, and sub-categories can be subdivided under some or all travel scenario types; taking sub-categories under some travel scenario types as an example, sub-categories exist under some travel scenario types, but not under other travel scenario types, for example, Figure 2 As shown, Figure 2 In the example, there are subcategories under the travel scenario types such as commuting, tourism, and returning home, but there are no subcategories under the business travel travel scenario type. For travel scenario types with subcategories, a subcategories under a travel scenario type corresponds to a travel scenario, indicating a travel intention, and has corresponding definitions and characteristics; for example, the subcategories of nearby tours under the travel scenario type of tourism correspond to a travel scenario, and have corresponding definitions and characteristics; the subcategories of long-distance tours under the travel scenario type of tourism correspond to a travel scenario, and have corresponding definitions and characteristics, and so on. For travel scenario types without subcategories, a travel scenario type corresponds to a travel scenario, indicating a travel intention, and has corresponding definitions and characteristics; for example, the travel scenario type of business travel corresponds to a travel scenario, and has corresponding definitions and characteristics.

[0043] In other words, travel scenarios are specific situations that describe travel purposes and motivations, and are ways of expressing travel intentions; travel scenario types are types of travel scenarios (such as commuting, tourism, etc.), and subcategories under travel scenario types further refine the specific needs or details of travel (such as weekly commuting under commuting, long-distance travel under tourism, etc.). Therefore, travel scenarios can be expressed through travel scenario types, or a combination of travel scenario types and subcategories, that is, travel scenarios can include travel scenario types, and can also further include subcategories under travel scenario types to accurately express travel needs.

[0044] The travel intention identification method provided in the embodiment of the present application is mainly used to identify the travel intention of the current trip, specifically based on the current trip information, the current passenger information and the historical travel information of the current passenger, to determine the travel scene matching the current trip from multiple travel scenes (multiple travel scene types and subclasses), so as to determine the travel intention of the current trip based on the travel scene matching the current trip. In order to determine the travel scene matching the current trip from multiple travel scenes, the embodiment of the present application can extract features from the current trip information, the current passenger information, and the historical travel information of the current passenger and combine the features to obtain the combined features to be matched corresponding to the current trip in multiple travel scenes; the combined features to be matched corresponding to a travel scene of the current trip are used for the matching process between the current trip and the travel scene to determine whether the current trip matches the travel scene, so as to match the combined features to be matched corresponding to the current trip in multiple travel scenes, the embodiment of the present application can match the travel scene that matches the travel intention of the current trip from multiple travel scenes.

[0045] In an optional implementation, the embodiment of the present application can set the feature type required for matching each travel scenario, so that according to the feature type required for matching each travel scenario, features are extracted from the current trip information, the current passenger information, and the historical travel information of the current passenger, and the features are combined to obtain the combined features to be matched corresponding to the current trip in each travel scenario. In other words, the matching process of each travel scenario requires a specific feature type, so according to the feature type required for matching each travel scenario, the specific features required for matching each travel scenario are extracted from the current trip information, the current passenger information, and the historical travel information of the current passenger, and the features are combined, so that the combined features to be matched for the current trip in each travel scenario can be formed for matching with each travel scenario, so as to be used for matching with each travel scenario.

[0046] As an optional implementation, for any travel scenario, the feature types required for travel scenario matching can be pre-set based on the definition and characteristics of the travel scenario, that is, the definition and characteristics of the travel scenario determine the types of features that the travel scenario needs to extract from the current trip information, the current passenger information, and the historical travel information, and can be pre-set. For multiple travel scenarios, some travel scenarios may need to extract and combine features from the current trip information, the current passenger information, and the historical travel information, while some travel scenarios may only need to extract and combine features from one or two of the current trip information, the current passenger information, and the historical travel information, depending on the specific definition and characteristics of the travel scenario.

[0047] In a further optional implementation, the current itinerary may correspond to at least one combined feature to be matched in any travel scenario, for example, the current itinerary corresponds to one or more combined features to be matched in a travel scenario, wherein a combined feature to be matched is composed of at least two features and can be regarded as a composite feature; for example, travel scenarios related to the travel scenario type of returning home may need to comprehensively consider features such as the identity, departure place, and destination of the current passenger to analyze whether the current passenger is traveling to and from his hometown, so that features such as the identity, departure place, and destination of the current passenger can be combined to form a combined feature to be matched; for another example, travel scenarios related to the travel scenario type of returning home may also need to comprehensively consider features such as departure date and arrival date to analyze whether the trip is during holidays. The number and specific form of the combined features to be matched corresponding to the current itinerary in any travel scenario can be determined based on the matching requirements of the travel scenario, and the embodiments of the present application are not limited thereto.

[0048] Of course, in other possible implementations, for any travel scenario, after extracting specific features required for travel scenario matching from current trip information, current passenger information and historical travel information, the embodiment of the present application may also support combining the extracted features as a whole into a combined feature to be matched, thereby obtaining the combined feature to be matched corresponding to the current trip in the travel scenario.

[0049] As an example, combining Figure 2As shown, for the weekly commuting subcategory under the commuting travel scenario type, based on the definition and characteristics of weekly commuting, the embodiment of the present application can preset the feature type required for matching the weekly commuting subcategory, so as to extract features corresponding to the feature type from the current trip information, the current passenger information and the historical travel information according to the feature type required for matching the weekly commuting subcategory, and combine them into one or more combined features to be matched; for example, features such as the departure place, destination, departure time, arrival time, travel time, fare, ticket type, etc. are extracted from the current trip information, features such as the age of the current passenger and the number of the current passengers are extracted from the current passenger information, and features corresponding to the current trip are extracted from the historical travel information. The features of historical trip information with the same itinerary route, such as the travel frequency and the identity of the passenger; based on the matching needs of the weekly commuting subcategory, these extracted features are combined into one or more to-be-matched combined features to obtain the to-be-matched combined features of the current trip in the weekly commuting travel scenario, so as to analyze whether the content of the to-be-matched combined features of the current trip in the weekly commuting matches the travel scenario of the weekly commuting, such as analyzing whether the content of the to-be-matched combined features of the current trip in the weekly commuting is consistent with the definition and characteristics of the weekly commuting. For example, it can be analyzed whether it is a regular round trip every week, whether the travel time matches the time corresponding to the longer travel time, whether the passenger is an adult, etc.

[0050] For another example, for the summer return home subcategory under the travel scenario type of return home, based on the definition and characteristics of summer return home, the embodiment of the present application can preset the feature type required for matching the summer return home subcategory, so as to extract features from the current itinerary information and the current passenger information and combine them into one or more combined features to be matched according to the feature type required for matching the summer return home subcategory, such as extracting the departure place, destination, departure time, fare, seat type, ticket type and other features from the current itinerary information, and extracting the current passenger's age, identity identification (such as identity number) and other features from the current passenger information, that is, at this time, features do not need to be extracted from the historical travel information; thus, Based on the matching needs of the summer return home subcategory, these extracted features are combined into one or more to-be-matched combined features to obtain the to-be-matched combined features of the current itinerary in the summer return home travel scenario, so as to analyze whether the content of the to-be-matched combined features of the current itinerary in the summer return home matches the travel scenario of the summer return home, such as analyzing whether the content of the to-be-matched combined features of the current itinerary in the summer return home is consistent with the definition and characteristics of the summer return home. For example, it can be analyzed whether the travel time is during the summer vacation, whether the current passenger is a student (analyzed by the age of the current passenger or whether the ticket type is a student ticket, etc.), whether the destination is the current passenger's hometown, etc.

[0051] In an optional implementation, the current passenger's native place can be determined through the current passenger's identity identification (such as an identity number). For example, the current passenger's identity number includes the geographic area code of the current passenger's native place. Specifically, the identity number is encoded according to certain rules, and specific numbers in the identity number (such as the first 6 digits) represent the geographic area code of the native place. Therefore, by analyzing whether the destination matches the native place corresponding to the current passenger's identity number, it is possible to analyze whether the current passenger is returning to his hometown.

[0052] For another example, for the subcategory of surrounding tours under the travel scenario type of tourism, based on the definition and characteristics of surrounding tours, the embodiment of the present application can preset the feature types required for matching the subcategory of surrounding tours, thereby extracting features from the current itinerary information and combining them into one or more combined features to be matched according to the feature types required for matching the subcategory of surrounding tours, such as extracting features such as departure place, destination, departure time, arrival time, travel time, fare, seat type, and ticket booking time from the current itinerary information; thereby, based on the matching requirements of the subcategory of surrounding tours, these extracted features are combined into one or more combined features to be matched, and the combined features to be matched of the current itinerary in the travel scenario of surrounding tours are obtained, so as to analyze whether the content of the combined features to be matched of the current itinerary in surrounding tours matches the travel scenario of surrounding tours, such as analyzing whether the content of the combined features to be matched of the current itinerary in surrounding tours is consistent with the definition and characteristics of surrounding tours. For example, it can be analyzed whether the travel time is a weekend or a short holiday, whether the destination is a nearby city or a nearby scenic spot, etc.

[0053] In an optional implementation, the embodiment of the present application can obtain the travel time by extracting the departure time and arrival time from the current itinerary information, and then compare the travel time with a pre-set calendar list to determine whether the current itinerary falls on weekends (non-working days) and holidays, where holidays refer to national statutory holidays, local holidays, etc.

[0054] It should be noted that, for any travel scenario (for example, for any travel scenario type and subclass), the feature type required for travel scenario matching can be pre-set based on the definition and characteristics of the travel scenario, thereby presetting the feature type required for travel scenario matching, wherein the feature types required for different travel scenario types and their subclasses may be different, and the feature type required for travel scenario matching can be flexibly set according to the definition and characteristics of the travel scenario; further, based on the preset feature type required for travel scenario matching, the embodiment of the present application can extract features corresponding to the feature type from the current trip information, the current passenger information and the historical travel information, and combine the features based on the matching requirements of the travel scenario to obtain one or more combined features to be matched corresponding to the current trip in the travel scenario, so as to realize the matching requirements based on the travel scenario, convert the feature information in the current trip information, the current passenger information and the historical travel information into the combined features to be matched corresponding to the current trip in the travel scenario, so as to provide support for matching a travel scenario that is consistent with the travel intention of the current trip from multiple travel scenarios.

[0055] For ease of understanding, Figure 3 An example diagram exemplarily shows some feature types required for matching of some travel scenarios, which can be used for reference.

[0056] Step S130: Match the combined features to be matched in each travel scenario of the current trip with the feature requirements of each travel scenario respectively, so as to determine the travel scenario that matches the current trip from multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario.

[0057] In the embodiments of the present application, each travel scenario has specific feature requirements, and the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; the feature requirements of the travel scenario can be understood as a detailed description of the features of the travel scenario, reflecting the definition and characteristics of the travel scenario. In other words, the feature requirements of the travel scenario are a specific description of the characteristic attributes of the travel scenario, including the feature matching criteria and feature threshold range of the travel scenario, which are the criteria for determining whether the current trip matches the travel scenario.

[0058] For ease of understanding, Figure 4 The following example diagrams exemplarily show some characteristic requirements of some travel scenarios, which can be used for reference. Figure 4 The diagram is only an example. The characteristic requirements of each travel scenario can be set based on the definition and characteristics of the travel scenario. The embodiments of the present application do not limit the specific form and content of the characteristic requirements of the travel scenario. As long as the characteristic requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario, they can be used as matching criteria for the travel scenario.

[0059] For any travel scenario, the embodiment of the present application can match the to-be-matched combined features of the current trip in the travel scenario with the feature requirements of the travel scenario, so as to determine whether the current trip matches the travel scenario, that is, whether the current trip belongs to the travel scenario; further, by matching the to-be-matched combined features of the current trip in each travel scenario with the feature requirements of each travel scenario, the travel scenario matched by the current trip among multiple travel scenarios can be determined, that is, the travel scenario to which the current trip belongs among multiple travel scenarios.

[0060] In an optional implementation, for any travel scenario, the embodiment of the present application may compare the to-be-matched combined features of the current trip in the travel scenario with the feature requirements of the travel scenario. If the to-be-matched combined features of the current trip in the travel scenario meet the feature requirements of the travel scenario, for example, the to-be-matched combined features of the current trip in the travel scenario are consistent with the feature requirements of the travel scenario or are within the feature threshold range of the feature requirements, then the current trip matches the travel scenario, that is, the current trip belongs to the travel scenario; if the to-be-matched combined features of the current trip in the travel scenario do not meet the feature requirements of the travel scenario, for example, the to-be-matched combined features of the current trip in the travel scenario are inconsistent with the feature requirements of the travel scenario or are not within the feature threshold range of the feature requirements, then the current trip does not match the travel scenario, that is, the current trip does not belong to the travel scenario.

[0061] In a further optional implementation, the embodiment of the present application may set up a travel scenario knowledge base and train a travel intention recognition model, thereby using the travel intention recognition model and the travel scenario knowledge base to match the combined features to be matched of the current trip in each travel scenario with the feature requirements of each travel scenario, and obtain a travel scenario that matches the current trip in multiple travel scenarios. For example, the combined features to be matched of the current trip in each travel scenario are respectively input into the travel intention recognition model, so that the travel intention recognition model matches the combined features to be matched of the current trip in each travel scenario with the feature requirements of each travel scenario based on the travel scenario knowledge base, so that the embodiment of the present application can obtain the travel scenario that matches the current trip output by the travel intention recognition model.

[0062] In the embodiment of the present application, the travel intention recognition model is a model used to recognize travel intentions, and the model forms include but are not limited to large models (such as natural language processing models), rule models, etc. The travel scenario knowledge base can be regarded as the background knowledge base of the travel intention recognition model, which records the background knowledge of multiple travel scenarios, where the background knowledge of a travel scenario includes but is not limited to the definition and characteristics of the travel scenario, and the characteristic requirements of the travel scenario; for ease of understanding, Figure 5The following is an example diagram showing the composition of the travel scenario knowledge base, which can be used for reference. The travel scenario knowledge base can be used to assist the travel intention recognition model in understanding the definition, characteristics and feature requirements of each travel scenario; further, the combined features to be matched in each travel scenario of the current trip are input into the travel intention recognition model, and the travel intention recognition model can determine the travel scenario that matches the current trip from multiple travel scenarios based on the records in the travel scenario knowledge base.

[0063] As an optional implementation, taking the travel intention recognition model as a large model for identifying travel intentions as an example, the large model refers to a model that can process a large amount of data and perform automatic feature learning and prediction. The large model can use a natural language processing model or other deep learning model. In an embodiment of the present application, the travel intention recognition model can be regarded as an inference engine or knowledge inference system for travel intentions. By providing a travel scenario knowledge base to the travel intention recognition model, the travel intention recognition model can understand the background knowledge related to the travel scenario, including but not limited to understanding the definition, characteristics, and feature requirements of each travel scenario (such as each travel scenario type and subclass). Furthermore, the travel scenario knowledge base can be continuously improved and updated according to actual conditions, such as by improving and updating the background knowledge of travel scenarios recorded in the travel scenario knowledge base through expert input, historical data analysis, etc.

[0064] In an optional implementation, the travel intention recognition model in the form of a large model can understand the background knowledge of the travel scenarios recorded in the travel scenario knowledge base through pattern learning or reasoning algorithms, so that the travel intention recognition model has the reasoning ability to match the travel scenario for the current trip. Furthermore, the combination features to be matched in each travel scenario of the current trip are input into the travel intention recognition model in the form of a large model, and the travel intention recognition model can match the combination features to be matched in the travel scenario of the current trip with the feature requirements of the travel scenario based on the background knowledge of the travel scenario recorded in the travel scenario knowledge base; if the travel intention recognition model determines that the combination features to be matched in a certain travel scenario of the current trip matches the feature requirements of the travel scenario, the travel intention recognition model can determine that the travel scenario is the travel scenario that matches the current trip, and thus output the matching travel scenario.

[0065] As an optional implementation, taking the travel intention recognition model as an example of a rule model for identifying travel intention, the rule model relies on an explicit set of rules for matching. Specifically, the rule model is composed of matching rules for various travel scenarios, wherein the matching rules for travel scenarios are formed based on the background knowledge of travel scenarios recorded in the travel scenario knowledge base, that is, the matching rules for travel scenarios correspond to the background knowledge of travel scenarios; for ease of understanding, as an example, taking the background knowledge of travel scenarios including the characteristic requirements of travel scenarios as an example, the rule model may include the following simplified matching rules for daily commuting: if the travel time is during the morning rush hour and the frequency is every weekday, then the daily commuting is matched.

[0066] As an optional implementation, after constructing the travel scenario knowledge base, the embodiment of the present application can construct matching rules for each travel scenario based on the background knowledge of each travel scenario in the travel scenario knowledge base (a travel scenario has one or more matching rules) to form a travel intention recognition model in the form of a rule model. Furthermore, the rule model checks whether the combined features to be matched in the travel scenario of the current trip match the feature requirements of the travel scenario through the matching rules of the travel scenario, so as to confirm the travel scenario matched by the current trip; for example, the matching rule can be a rule structure of conditions and conclusions, so that the combined features to be matched in the travel scenario of the current trip are used as input, and the feature requirements of the travel scenario are used as conditions. When the input meets the conditions, the travel intention recognition model in the form of a rule model can draw the conclusion that the current trip matches the travel scenario.

[0067] Of course, the big model and the rule model are only optional model forms of the travel intention recognition model. The embodiments of the present application can also support travel intention recognition models in other model forms, not limited to the big model and the rule model.

[0068] Step S140: Determine the travel intention of the current trip according to the travel scenario matched by the current trip.

[0069] As an optional implementation, the current itinerary may match one or more travel scenarios, for example, the current itinerary matches one or more travel scenarios. Specifically, the definitions and characteristics of some travel scenarios may overlap or overlap (for example, travel scenarios related to returning home and traveling may both occur on holidays), so that multiple features of the current itinerary (such as departure time, departure place, destination, etc.) may simultaneously meet different travel scenarios, so the current itinerary may match more than one travel scenario.

[0070] For example, taking the current itinerary of a user going home to visit relatives during holidays, and his hometown is a tourist city, the combined features to be matched in the travel scenarios related to returning home of the current itinerary are similar to the combined features to be matched in the travel scenarios related to tourism. Therefore, when the travel intention recognition model determines the travel scenarios that match the current itinerary, the current itinerary may match both the travel scenarios related to returning home and the travel scenarios related to tourism.

[0071] In other words, the intersection of the definition and characteristics of travel scenarios makes it possible for the current trip to have multiple matches in more than one travel scenario, which is in line with the matching logic of actual travel scenarios.

[0072] In an optional implementation, if the current trip matches only one travel scenario, the travel scenario matching the current trip can be directly used as the travel intention of the current trip.

[0073] In an optional implementation, if the current trip matches multiple travel scenarios, the embodiment of the present application can determine the travel scenario with the highest priority matching the current trip as the travel intention of the current trip based on the priorities of the various travel scenarios matching the current trip.

[0074] As an optional implementation, the priority of travel scenarios can be determined based on the matched ratio of travel scenarios. The matched ratio of a travel scenario represents the ratio of the number of matches of the travel scenario to the total number of matches of multiple travel scenarios. Thus, the travel scenario with the highest matched ratio among the travel scenarios matching the current trip has the highest priority. That is to say, the matched ratio of travel scenarios refers to the proportion of the number of travel scenarios that have been matched in all travel scenarios; for example, if the total number of matches of all travel scenarios is 100 times, that is, the travel service platform has identified the travel intentions of multiple trips, and the travel scenarios that are matched by multiple trips as a whole are 100 times, further, assuming that the number of matches of daily commuting travel scenarios is 50 times, that is, the travel service platform has identified the travel intentions of multiple trips, and the travel scenarios of daily commuting are matched 50 times, then the matched ratio of daily commuting travel scenarios is 50 / 100=50%; the relative priority of travel scenarios can be determined by the proportion of the number of matches of travel scenarios to the total number of matches of all travel scenarios, so that when the current trip matches more than one travel scenario, the travel scenario with the highest priority matched by the current trip is taken as the travel intention of the current trip, that is, the travel scenario with the highest matching ratio among the travel scenarios matched by the current trip is the travel intention of the current trip.

[0075] The travel intention recognition method provided in the embodiment of the present application integrates multi-dimensional information such as current trip information, current passenger information, and historical travel information of the current passenger to match the current trip with the travel scenario; specifically, features are extracted and combined from the current trip information, the current passenger information, and the historical travel information to form the to-be-matched combined features corresponding to the current trip in multiple travel scenarios, so that the to-be-matched combined features corresponding to the travel scenarios of the current trip can reflect the various dimensional information and background of the current trip, avoid the possible deviation of the single dimensional information of the current trip, and ensure that the background of the current trip can be fully considered when matching the travel scenario of the current trip; then, based on the feature requirements of the travel scenario that are consistent with the definition and characteristics of the travel scenario, the to-be-matched combined features of the current trip in each travel scenario are matched with the feature requirements of each travel scenario respectively, so as to obtain the to-be-matched combined features of the current trip in multiple travel scenarios. The travel scenario that matches the current trip is determined from the travel scenarios, so as to accurately determine the travel scenario that matches the current trip; that is, each travel scenario has clear feature requirements, which are consistent with the specific definition and characteristics of the travel scenario. By comparing and matching the combined features to be matched in the current trip in the travel scenario with the feature requirements of the travel scenario, it is possible to more accurately judge whether the current trip meets the requirements of the travel scenario, thereby accurately identifying the travel scenario that matches the current trip from multiple travel scenarios; at the same time, the historical travel information of the current passenger is taken into account in the process of travel intention identification. Since the historical travel information of the current passenger helps to understand the historical travel habits and preferences of the current passenger, the travel scenario that matches the current trip can be made more accurate; furthermore, according to the travel scenario that matches the current trip, the travel intention of the current trip is determined, and the travel intention identification of the current trip can be completed. In summary, the embodiments of the present application can improve the accuracy of travel intention identification and provide support for improving the accuracy of travel strategy recommendations.

[0076] As an implementation example, Figure 6 The following is an exemplary diagram showing the architecture of the travel service platform provided in the embodiment of the present application. Figure 6 As shown, the travel service platform may include: a ticketing system 610 , a historical travel database 620 , and a travel intention recognition system 630 .

[0077] The ticketing system 610 is used to provide ticketing services such as ticket booking and ticket inquiry for users; the ticketing system 610 can transmit the current itinerary information and current passenger information of the current itinerary in the user's ticketing request to the travel intention recognition system 630. In an optional implementation example, when a user books a ticket through the ticketing system 610, if there are no remaining tickets for the user's desired travel plan (for example, there are no remaining tickets for the user's desired train or flight), the ticketing system 610 can transmit the current itinerary information and current passenger information of the current itinerary in the ticketing request to the travel intention recognition system 630, so that the travel intention recognition system 630 can recognize the travel intention of the current itinerary and provide support for the user to recommend alternative travel plans.

[0078] The historical travel database 620 records the historical travel information of multiple passengers; after the travel intention identification system 630 obtains the current trip information and current passenger information transmitted by the ticketing system 610, it can retrieve the current passenger's historical travel information from the historical travel database 620 based on the current passenger information; thus, the travel intention identification system 630 can obtain the current trip information, the current passenger information and the current passenger's historical travel information to identify the travel intention of the current trip.

[0079] The travel intention recognition system 630 may include: a pre-processing module 631 , a travel intention recognition model 632 , a post-processing module 633 , and a travel scenario knowledge base 634 .

[0080] The preprocessing module 631 is used to extract and combine features from the current trip information, the current passenger information and the historical travel information to obtain the combined features to be matched corresponding to the current trip in multiple travel scenarios; The travel intention recognition model 632 is used to match the combined features to be matched in each travel scenario of the current trip with the feature requirements of each travel scenario, so as to determine the travel scenario that matches the current trip from multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; A post-processing module 633 is used to determine the travel intention of the current trip according to the travel scenario matched by the current trip; The travel scenario knowledge base 634, as a background knowledge base of the travel intention recognition model, records the background knowledge of multiple travel scenarios. The background knowledge of travel scenarios at least includes the definition, characteristics and feature requirements of the travel scenarios.

[0081] The relevant contents of the pre-processing module, travel intention recognition model, post-processing module, and travel scenario knowledge base in the travel intention recognition system 630 can refer to the description of the corresponding parts of the previous text, which will not be expanded here. In an optional implementation, the travel intention recognition model can be a large model or a rule model, and the relevant contents of the travel intention recognition model in the form of a large model or a rule model can refer to the description of the corresponding parts of the previous text, which will not be expanded here.

[0082] Further, combined with Figure 6 As shown, the travel service platform may also include: a proportion data recording module 640, which is used to record the matching ratio of each travel scenario; thus, when the travel intention recognition model 632 determines that there are more than one travel scenarios that match the current trip, the post-processing module 633 can determine the travel scenario with the highest matching ratio among the travel scenarios that match the current trip as the travel intention of the current trip based on the matching ratio of each travel scenario recorded by the proportion data recording module. In other optional implementations, if the travel intention recognition model 632 determines that there is only one travel scenario that matches the current trip, the post-processing module 633 can directly determine the travel scenario that matches the current trip as the travel intention of the current trip.

[0083] In a further optional implementation, the proportion data recording module can be regarded as part of the market proportion data system of the travel service platform; the market proportion data system is a large-scale analysis system based on global data of the travel service platform, which aims to identify group characteristics and travel trends by counting and analyzing the behavior patterns of all users. In other words, the market proportion data system is a macro analysis tool that provides reference data from an overall level, such as collecting ticket booking data of all users, and extracting characteristic information such as departure place, destination, travel time, seat type, etc., so as to analyze the overall travel trend (such as popular routes, ticket demand in different time periods, etc.). Further, the market proportion data system can also collect and analyze user data of all users (such as age group, occupational distribution, regional preference) and group characteristics related to travel. Through the macro analysis of the full amount of data by the market proportion data system, the travel rules of specific time, specific area, and specific population can be determined. In the embodiment of the present application, the matched proportion of each travel scene is a type of data collected and analyzed by the market proportion data system, which can be recorded in the proportion data recording module in the market proportion data system.

[0084] Further, combined with Figure 6As shown, the travel service platform may also include a travel intention storage module 650. The travel intention storage module 650, as the travel intention storage layer of the travel service platform, may be used to store the travel intention data of each trip; in an optional implementation, the travel intention data of the trip may at least include the travel intention of the trip, and may further include the OD (Origin / Destination) pair corresponding to the trip. For example, for multiple trips submitted to the travel service platform by multiple users, the travel service platform may use the travel intention identification method provided in the embodiment of the present application to determine the travel intention of each trip, thereby forming the travel intention of each trip, and the departure and destination (i.e., OD pair) corresponding to each trip, and storing the travel intention data of each trip in the travel intention storage module.

[0085] In a further optional implementation, the travel intention data of a trip stored in the travel intention storage module may expire over time, that is, the travel intention data of a trip may lose timeliness over time. For example, the travel intention related to returning home during holidays may expire after the holidays. Therefore, the expired travel intention data in the travel intention storage module needs to be cleaned up regularly to maintain the efficient operation of the travel intention storage module. Furthermore, when the travel intention deviates greatly from the user's actual travel behavior, the determined travel intention may become invalid. For example, the user originally planned to return home, but eventually changed the ticket to a non-hometown location. In this case, there is a deviation between the travel intention and the user's actual travel behavior, and the travel intention becomes invalid; the travel intention data corresponding to the invalid travel intention can be marked as invalid data.

[0086] Further, combined with Figure 6 As shown, the travel service platform may also include: a verification module 660, which is used to verify, according to the verification list, whether the travel scenario matched by the current trip output by the travel intention recognition model 632 meets the verification rules corresponding to the travel scenario; if not, a warning prompt message is triggered. The output result of the travel intention recognition model 632 (the travel scenario matched by the current trip) can be verified by the verification module 660, and then the travel intention recognition model can be optimized based on the verification result, which can improve the reliability and accuracy of travel intention recognition. As an example, the verification module can set the verification rules corresponding to each travel scenario respectively through the verification list to verify the output result of the travel intention recognition model, and the verification rules of the travel scenario may be partially similar to the rules when the travel intention recognition model matches the travel scenario for the current trip.

[0087] For example, for any travel scenario, the check list may specifically record the check items corresponding to the current trip information, current passenger information, and historical travel information of the travel scenario. Figure 7An example diagram showing the verification rules for some travel scenarios recorded in the form of a verification list is provided for reference.

[0088] Furthermore, the background knowledge of travel scenarios recorded in the travel scenario knowledge base may include, in addition to the definition, characteristics and feature requirements of travel scenarios, extended knowledge points and logical rules for matching travel scenarios, so as to help the travel intention recognition model more accurately determine the travel scenarios that match the current trip in specific tasks. For example, the extended knowledge points and logical rules that can be extended by the travel scenario knowledge base for matching travel scenarios include but are not limited to: The identity number rule is used to analyze the current passenger's identity number to determine the hometown of origin; for example, the travel scenario knowledge base can expand the knowledge points and logic rules of the geographical location corresponding to the geographic area code, where the identity number contains the geographic area code of the hometown; the identity number rule is used to determine whether the travel intention of the current trip is a travel scenario related to returning home; The tourist city rule is used to identify whether the destination is a tourist city to assist in determining travel intentions, such as determining whether the travel intention is for the purpose of tourism; thus, the travel scenario knowledge base can expand knowledge points and logical rules such as the tourist city list and its corresponding scenic spot information; the tourist city rule is used to determine whether the travel intention of the current itinerary is a travel scenario related to tourism. If the destination of the current itinerary is in the tourist city list, or the destination is associated with the scenic spot corresponding to the tourist city, it can be determined that the travel intention of the current itinerary meets the travel scenario related to tourism; further, combined with holiday data, it can be determined whether the travel scenario related to tourism is holiday tourism.

[0089] It should be noted that the extended knowledge points and logical rules of the travel scenario knowledge base can be determined according to actual conditions, and the above is only an example; for example, based on the definition, characteristics and matching requirements of specific travel scenarios, the travel scenario knowledge base is extended to the extended knowledge points and logical rules required for specific travel scenarios. The embodiments of the present application do not limit this.

[0090] In a further optional implementation, after determining the travel intention of the current trip, the travel service platform can recommend a travel strategy based on the travel intention of the current trip. As an optional implementation, Figure 8 An optional flow chart of a travel strategy recommendation method provided in an embodiment of the present application is exemplarily shown. The travel strategy recommendation method can be applied to a travel service platform, such as a travel server.

[0091] Reference Figure 8 The travel strategy recommendation method provided in the embodiment of the present application may include the following steps.

[0092] Step S810: Obtain the travel intention of the current trip.

[0093] In an optional implementation, the current itinerary can be submitted to the travel service platform through travel service requests such as ticketing requests and travel strategy recommendation requests, so that the travel service platform can use the travel intention identification method provided in the embodiment of the present application to determine the travel intention of the current itinerary. The relevant content of identifying the travel intention of the current itinerary can be referred to the previous description and will not be expanded here.

[0094] In an optional implementation, the embodiment of the present application can directly determine the travel intention of the current trip after obtaining the travel service request. For example, if the travel service request is a travel strategy recommendation request that directly requests a travel strategy, the travel intention of the current trip can be directly determined. For example, the embodiment of the present application can directly determine the travel intention of the current trip without limiting the type and specific form of the travel service request.

[0095] In other optional implementations, the embodiments of the present application can determine the travel intention of the current trip after obtaining the travel service request and satisfying the preset travel intention identification conditions. For example, when the travel service request is a ticket request, the embodiments of the present application can determine the travel intention of the current trip when the ticket status corresponding to the ticket request satisfies the preset travel intention identification conditions.

[0096] For example, taking the travel service request as a booking request in a ticketing request, the booking request is used to book a ticket for the desired travel plan (such as a desired train or a desired flight) for the current itinerary. If there are no remaining tickets for the desired travel plan for the current itinerary, it is deemed to meet the preset travel intention identification conditions, and the embodiment of the present application can determine the travel intention of the current itinerary.

[0097] Step S820: Determine the travel preference information of the current passenger.

[0098] The travel preference information of the current passenger refers to the habits and tendencies of the current passenger when traveling. In an optional implementation, the travel preference information of the current passenger can be the personal travel preference information of the current passenger, or the group travel preference information of the group to which the current passenger belongs. That is to say, the embodiment of the present application can use the personal travel preference information of the current passenger as the travel preference information of the current passenger; or use the group travel preference information of the group to which the current passenger belongs as the travel preference information of the current passenger.

[0099] The personal travel preference information of the current passenger can be determined based on the current passenger’s historical travel information, and / or personalized settings, and / or the current passenger’s operation records on the travel service platform, and be regarded as the current passenger’s personalized travel preference; for example, the personal travel preference information of the current passenger is determined based on the current passenger’s historical travel records, the current passenger’s active choices (such as the preference settings filled in by the current passenger on the travel service platform), the current passenger’s behavioral operations on the travel service platform (such as frequent inquiries about a certain type of transportation), etc.

[0100] In an optional implementation, the travel service platform can record the personal travel preference information of each passenger through a personal travel preference database, including the passenger information of each passenger and the corresponding personal travel preference information; thus, the embodiment of the present application can obtain the personal travel preference information of the current passenger from the personal travel preference database based on the identity of the current passenger. In the optional implementation example, the personal travel preference information can be regarded as part of the personal portrait.

[0101] The travel preference information of the group to which the current passenger belongs refers to the general travel preference information of the group to which the current passenger belongs, that is, the travel preference information of the group is the general travel preference characteristics of a specific group, rather than the personalized characteristics of a single individual. The travel preference information of the group as a group preference does not change with the short-term behavior of the individual; in an optional implementation, the embodiment of the present application can determine the travel preference information of each group based on the massive group data related to travel accumulated by the travel service platform.

[0102] In an optional implementation, the travel service platform can record the travel preference information of each group of people through a travel preference database, including the population characteristics of each group of people and the corresponding travel preference information of the population; thus, the embodiment of the present application can determine the population characteristics corresponding to the passenger information of the current passenger based on the passenger information of the current passenger, and thus determine the population corresponding to the population characteristics as the population to which the current passenger belongs. For example, the population can be classified according to dimensions such as age, occupation, and region, so as to divide different groups of people; then, the travel preference information of the population to which the current passenger belongs is obtained from the travel preference database.

[0103] For example, the current passenger's travel preference information (not limited to the current passenger's personal travel preference information or group travel preference information) may include at least one of the following: Travel time preferences, such as departure time preference, arrival time preference, etc.; Arrival time overflow sensitivity, for example, if the vehicle fails to arrive at the destination on time, the user's sensitivity to the timeout of the vehicle's arrival at the destination; Departure time overflow sensitivity, for example, if the vehicle does not depart from the departure station on time, the user's sensitivity to the timeout of the vehicle from the departure station; Fare preferences, such as how sensitive users are to fare overspill; Comfort preference, that is, the user's comfort demand for transportation, such as transportation preference, seat preference, etc.; Trip time overflow sensitivity, such as the user's sensitivity to trip timeouts; Departure anxiety refers to whether users are easily anxious about departure. If the anxiety level is high, users tend to arrive at the departure site longer in advance, while if the anxiety level is low, users tend to arrive at the departure site close to the departure time.

[0104] Of course, the above travel preference information is only an example, and the embodiment of the present application does not limit the specific form of the travel preference information, as long as it can express the habits and tendencies of passengers when traveling.

[0105] Step S830: recommend a travel strategy for the current trip based on the travel intention of the current trip and the travel preference information of the current passenger.

[0106] After determining the travel intention of the current trip and the travel preference information of the current passenger, the embodiment of the present application can recommend travel strategies based on the travel intention of the current trip (such as returning home, business travel, tourism, etc.) and travel preference information. For example, based on the travel intention of the current trip, recommend travel plans and / or travel additional services that meet the travel preferences of the current passenger.

[0107] For example, when the travel intention of the current trip is a commuting-related travel scenario, the embodiment of the present application can combine the current passenger's travel preference information to recommend travel plans that meet commuting needs from multiple aspects such as transportation, time arrangements, transfer plans, and seat types; further, it can also provide additional services such as shuttle services for commuting-related travel scenarios, commuting discount plans (such as high-speed rail monthly tickets, subway monthly cards, etc.) to enhance the commuting experience and convenience.

[0108] For example, when the travel intention of the current itinerary is to return home during a major holiday (such as returning home for the Spring Festival), the embodiment of the present application can combine the current passenger's travel preference information to recommend travel plans that meet the needs of returning home during a major holiday (such as returning home for the Spring Festival) in terms of transportation, travel time, ticketing priority, etc.; further, additional services such as transfer guidance and itinerary reminders can also be provided to ensure that passengers can return home smoothly and improve travel convenience and comfort.

[0109] For example, when the travel intention of the current itinerary is for business travel, the embodiment of the present application can combine the current passenger's travel preference information to recommend travel plans that meet business travel needs from multiple aspects such as efficiency-first transportation, flexible time arrangements, and more comfortable seat selection; further, it can also provide additional services such as itinerary planning, rapid security checks, and transportation connections to improve the travel efficiency and experience of business travelers.

[0110] For example, when the travel intention of the current itinerary is a tourism-related travel scenario, the embodiment of the present application can combine the current passenger's travel preference information to recommend travel plans that meet tourism needs from multiple aspects such as suitable leisure transportation and flexible ticketing options; further, it can also provide additional services such as attraction information introduction, hotel reservations, and transportation connections to help passengers optimize their itinerary arrangements and enhance the overall travel experience.

[0111] As an implementation example, Fig. 9 Another example diagram of the architecture of the travel service platform provided by the embodiment of the present application is shown in FIG. Figure 6 and Fig. 9 As shown, in Figure 6 Based on the example, Fig. 9 The travel service platform shown may also include: a travel strategy recommendation system 910 .

[0112] The travel strategy recommendation system 910 is used to obtain the travel intention of the current trip and determine the travel preference information of the current passengers; based on the travel intention of the current trip and the travel preference information of the current passengers, recommend the travel strategy of the current trip.

[0113] In an optional implementation, the travel intention of the current trip can be stored in the travel intention storage module 650, and the travel strategy recommendation system 910 can obtain the travel intention of the current trip from the travel intention storage module 650.

[0114] In a further optional implementation, the travel service platform may also include: a travel preference database 920, which records travel preference information; the travel preference database 920 may be a personal travel preference database that records personal travel preference information of each passenger, or a crowd travel preference database that records crowd travel preference information of each crowd. The embodiment of the present application may detect the travel preference information of the current passenger from the travel preference database 920, such as detecting the crowd travel preference information of the crowd to which the current passenger belongs, and thus input the detected travel preference information of the current passenger into the travel strategy recommendation system 910, so that the travel strategy recommendation system 910 recommends the travel strategy of the current trip in combination with the travel intention of the current trip and the travel preference information of the current passenger.

[0115] In a further optional implementation, the travel strategy recommendation system 910 may include a strategy management component for managing travel strategy recommendation rules (such as recommendation algorithms), trigger conditions, etc.

[0116] In a further optional implementation, taking the recommendation of travel plans as an example, the recommendation rules of the travel strategy recommendation system (such as the recommendation algorithm) can be implemented based on the recall strategy and the scoring strategy. The recall strategy and the scoring strategy are respectively responsible for the two key stages in the travel strategy recommendation process: the preliminary screening stage (recall stage) and the precise sorting stage (scoring stage). That is, the recall strategy and the scoring strategy are important components of the recommendation algorithm (used to recommend travel strategies), and their purpose is to recommend efficient and accurate travel plans for passengers.

[0117] As an optional implementation, the purpose of the recall strategy is to quickly filter out candidate sets related to travel demand in order to reduce the amount of subsequent calculations and improve the efficiency of recommendations. The candidate set can include multiple candidate travel plans. The recall strategy has the following characteristics: a large range, a larger candidate set will be filtered out during the recall (for example, the train number that matches the destination is selected from all train number information); fast screening, focusing on filtering out travel plans that obviously do not meet travel needs; coarse-grained screening, not particularly finely sorting candidate travel plans, only ensuring that these candidate travel plans are related to travel needs. For example, in the travel scenario of returning home during the Spring Festival, the recall strategy can filter out the train number that matches the destination of the trip from all train ticket information at the travel time related to returning home during the Spring Festival.

[0118] In a more specific optional implementation, the recall strategy may include at least one of the following sub-strategies: The rejection sub-strategy eliminates travel plans that do not meet the passenger's preferences. For example, if the passenger prefers to only accept high-speed trains, ordinary trains will be directly eliminated. If the passenger prefers not to accept evening trains, evening trains will be eliminated. The cross-city sub-strategy recommends travel plans that require cross-city transfers; The time-limited sub-strategy recalls travel plans within a specified time period; The price restriction sub-strategy eliminates travel plans that exceed the passenger's budget.

[0119] That is to say, in the case of recommending travel plans, the core purpose of the recall strategy is to quickly find candidate travel plans that may meet travel needs (travel needs can be determined by a combination of travel intentions and travel preferences) and provide input for the subsequent scoring strategy stage.

[0120] As an optional implementation, the scoring strategy is to accurately score and sort the candidate set generated by the recall strategy based on multiple dimensions to select the recommended travel plan. The scoring strategy has the following characteristics: small scope, only for the candidate set provided by the recall strategy; precise calculation, detailed evaluation and sorting of the candidate travel plans in the candidate set; fine-grained sorting, combining the passenger's specific travel preference information (such as time, price, comfort) to determine the priority.

[0121] As an example, the scoring dimension of the recall strategy may include at least one of the following: Departure time: give priority to recommending trains that meet passengers' time preferences; Price: Based on the passenger's fare preference, priority is given to recommending trains that meet the budget; Time consumption: recommend travel plans with shorter total time consumption; Booking success rate: In the ticket booking scenario, priority is given to recommending trains with higher booking success rates.

[0122] That is to say, in the case of recommending travel plans, the core purpose of the scoring strategy is to select the recommended travel plan from the candidate set of the recall strategy to ensure the accuracy of the recommended travel plan.

[0123] In a further optional implementation, the travel service platform may also include a test component 930 for testing the impact of different versions of travel strategies on user behavior using a control experiment test method (such as an AB test method), such as the impact of different versions of travel strategies on user click-through rates and order conversion rates, so as to optimize the recommendation logic of the travel strategy. In an optional implementation, the test component may use a control experiment test method such as an AB test to test the impact of different versions of travel strategies on user behavior; specifically, an AB test is a control experiment test method used to compare the differences between two or more versions of travel strategies, thereby evaluating the impact of different versions of travel strategies on user behavior or system goals (such as click-through rates, order conversion rates, etc.) to optimize the recommendation logic of the travel strategy.

[0124] In an optional implementation, different versions of travel strategies may include a first version of the travel strategy and a second version of the travel strategy. In an embodiment of the present application, the second version of the travel strategy is determined based on travel intention and travel preference information, and the first version of the travel strategy is the original version of the travel strategy.

[0125] For example, the first version of the travel strategy can be regarded as the travel strategy of version A in the AB test, that is, the original version of the travel strategy, and the original version of the travel strategy is not determined by the technical solution provided in the embodiment of the present application; the second version of the travel strategy can be regarded as the travel strategy of version B in the AB test, that is, the new version of the travel strategy, and the new version of the travel strategy is determined by the technical solution provided in the embodiment of the present application, that is, the second version of the travel strategy is determined in combination with travel intention and travel preference.

[0126] During the testing process, the embodiments of the present application can define key indicators of the test, such as key indicators such as improving user click-through rate, increasing order conversion rate, optimizing user retention rate, etc.; then, randomly divide the users into two groups, such as randomly divide the users into Group A and Group B, Group A experiences version A, and Group B experiences version B, and ensure that the distribution of the two groups of users is random and uniform; the test component collects user behavior data corresponding to version A and version B, such as how many times the user clicks on the recommended content, how many orders the user completes, etc., so as to compare the performance of Group A and Group B on key indicators, determine the version with better key indicator performance, and optimize the recommendation logic.

[0127] In a further optional implementation, the travel service platform may also include a data collection and analysis system 940 for collecting and analyzing user behavior data through data collection to optimize the performance of the travel strategy recommendation system and the travel strategy recommended by the travel strategy recommendation system.

[0128] Specifically, data tracking refers to presetting data collection points in the key processes of the system or user behaviors, and recording user behaviors, operations, and feedback through tracking points.

[0129] In an optional implementation, the dimensions of data embedding can be test dimensions and population dimensions.

[0130] The data points in the test dimension are used to evaluate the effectiveness of different versions of travel strategies (such as travel strategy version A and travel strategy version B in AB testing); the data points can be recall data and link data; among them, the recall data can include coverage rate, indicating whether the recalled candidate solutions can meet user needs, such as whether the recalled candidate solutions include the user's expected destination or departure time; link data can include seat occupation success rate and ticket issuance rate, etc. The seat occupation success rate is used to count the success rate of users occupying the required seats / seats in scenarios such as holiday ticket booking (such as Spring Festival travel scenarios), and the ticket issuance rate is used to count the proportion of users who actually complete ticket issuance.

[0131] Data points in the crowd dimension are used to analyze whether the recommended travel strategies meet the needs of different crowds based on their preferences, so as to evaluate the effectiveness of the recommendation. The data points can be attribution data, such as the user's final decision data, including factors that affect the user's decision (such as time, price, comfort, etc.).

[0132] Data analysis can be performed by collecting user behavior data through data embedding. The specific indicators of data analysis may include but are not limited to: Reach rate: whether the recommended travel strategy is seen or paid attention to by users; Callback rate: whether the user has taken further actions on the recommended travel strategy (such as clicking, viewing details, etc.); Acceptance rate: whether the user accepts the recommended travel strategy; Ticket issuance rate refers to the proportion of users who actually complete ticket issuance.

[0133] In a further optional implementation, in order to measure whether the recommendation causes interference to the user, the data tracking and analysis system can collect and analyze the interruption data to measure whether the travel strategy recommendation system causes recommendation interference to the user; wherein, the interruption data refers to the data that disturbs the normal operation or decision of the user due to the recommendation or intervention behavior of the travel strategy recommendation system, reflecting the negative feedback or unexpected behavior of the user due to irrelevant or inappropriate recommendations. In other words, the interruption data includes the behavior data generated by the user due to the interference of the recommended content.

[0134] As an optional implementation, the relevant factors of the disturbance data are not limited to closing recommendation behavior and rejection rate, wherein closing recommendation behavior is the behavior of users actively closing recommended content, and the rejection rate is the proportion of users explicitly rejecting recommended content. In other words, closing recommendation behavior is a negative reaction behavior of users to recommended content, which is manifested as the operation of users actively closing recommended content; the rejection rate is the proportion of users' behavior of explicitly rejecting recommended content to the number of recommended display times. Furthermore, the embodiment of the present application can determine the disturbance rate corresponding to the disturbance data based on closing recommendation behavior and rejection rate, wherein the disturbance rate is an indicator calculated from the disturbance data, which is used to measure the proportion of recommended content causing interference to users; for example, the disturbance rate is determined by the number of closing recommendation behavior and rejecting recommendation behavior, wherein the number of rejecting recommendation behavior can be determined based on the rejection rate, for example, the disturbance rate = (number of closed recommendations + number of rejected recommendations) / number of recommended content displays; if the disturbance rate is high, it means that the recommended content is more disturbing to the user, and the recommendation logic needs to be optimized.

[0135] In a further optional implementation, the embodiment of the present application also provides a travel server, which is regarded as a server or server cluster set up in the travel service platform; in an optional implementation, the travel server may include a memory and a processor, the memory stores computer execution instructions, and the processor calls the computer execution instructions stored in the memory to execute the travel intention identification method provided by the embodiment of the present application, or the travel strategy recommendation method provided by the embodiment of the present application.

[0136] In a further optional implementation, the embodiment of the present application also provides a storage medium, which stores computer execution instructions. When the computer execution instructions are executed (for example, when the computer execution instructions are executed by a processor), the travel intention identification method provided by the embodiment of the present application, or the travel strategy recommendation method provided by the embodiment of the present application is implemented.

[0137] In a further optional implementation, the embodiment of the present application also provides a computer program product, including computer execution instructions. When the computer execution instructions are executed (for example, when the computer execution instructions are executed by a processor), the travel intention identification method provided by the embodiment of the present application, or the travel strategy recommendation method provided by the embodiment of the present application is implemented.

[0138] The above describes multiple implementation schemes provided by the embodiments of the present application. The various optional methods introduced in each implementation scheme can be combined and cross-referenced with each other without conflict, thereby extending a variety of possible implementation schemes, which can all be considered as implementation schemes disclosed and open in the embodiments of the present application.

[0139] Although the embodiments of the present application are disclosed above, the present application is not limited thereto. Any person skilled in the art may make various changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be subject to the scope defined by the claims.

Claims

1. A travel intention recognition method, characterized in that: include: Get the current trip information and current passenger information of the current trip, as well as the current passenger's historical travel information; Extracting and combining features from the current trip information, the current passenger information, and the historical travel information to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios; Matching the combined features to be matched of the current trip in each travel scenario with the feature requirements of each travel scenario respectively, so as to determine the travel scenario matched by the current trip from the multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; A travel intention of the current trip is determined according to a travel scenario matched by the current trip.

2. The method according to claim 1, characterized in that The matching of the combined features to be matched of the current trip in each travel scenario with the feature requirements of each travel scenario respectively to determine the travel scenario matching the current trip from the multiple travel scenarios includes: Inputting the to-be-matched combined features of the current trip in each travel scenario into the travel intention recognition model respectively, so that the travel intention recognition model can match the to-be-matched combined features of the current trip in each travel scenario with the feature requirements of each travel scenario respectively based on the travel scenario knowledge base; Obtaining a travel scenario matching the current trip output by the travel intention recognition model; Among them, the travel scenario knowledge base is a background knowledge base of the travel intention recognition model, which records the background knowledge of the multiple travel scenarios. The background knowledge of the travel scenarios at least includes the definition, characteristics and feature requirements of the travel scenarios.

3. The method according to claim 2, characterized in that The travel intention recognition model is a large model, which understands the background knowledge of multiple travel scenarios recorded in the travel scenario knowledge base through pattern learning or reasoning algorithm, so as to have the reasoning ability to match the travel scenario for the current trip; or, The travel intention recognition model is a rule model, including matching rules for various travel scenarios; the matching rules for the travel scenarios are formed based on the background knowledge of the travel scenarios recorded in the travel scenario knowledge base, and are used to check whether the combined features to be matched in the current trip in the travel scenario match the feature requirements of the travel scenario.

4. The method according to any one of claims 1 to 3, characterized in that: The extracting and combining features from the current trip information, the current passenger information and the historical travel information to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios respectively includes: For any travel scenario, extract features from the current trip information, the current passenger information, and the historical travel information based on the feature type required for matching the travel scenario; For any travel scenario, based on the matching requirements of the travel scenario, the extracted features are combined into at least one to-be-matched combined feature corresponding to the current trip in the travel scenario, or the extracted features are combined as a whole into one to-be-matched combined feature corresponding to the current trip in the travel scenario.

5. The method according to any one of claims 1 to 3, characterized in that: Determining the travel intention of the current trip according to the travel scenario matched by the current trip includes: If the current trip matches one travel scenario, the travel scenario matching the current trip is used as the travel intention of the current trip; If there are multiple travel scenarios matched by the current trip, according to the priorities of the travel scenarios matched by the current trip, determining the travel scenario with the highest priority matched by the current trip as the travel intention of the current trip; Among them, the priority of the travel scenario is determined based on the matched ratio of the travel scenario. The matched ratio of a travel scenario represents the ratio of the number of matches of the travel scenario to the total number of matches of the multiple travel scenarios. The travel scenario with the highest matched ratio among the travel scenarios matching the current trip is the travel scenario with the highest priority.

6. The method according to any one of claims 2-3, characterized in that: The method further comprises: According to the verification list, verify whether the travel scenario matched by the current trip output by the travel intention recognition model meets the verification rules corresponding to the travel scenario; if not, trigger a warning prompt message; Among them, for any travel scenario, the verification list records the verification items corresponding to the current trip information, current passenger information and historical travel information of the travel scenario respectively; The multiple travel scenarios include multiple travel scenario types and subcategories under the travel scenario types; wherein some of the multiple travel scenario types have subcategories; for travel scenario types with subcategories, one subcategory under one travel scenario type corresponds to one travel scenario; for travel scenario types without subcategories, one travel scenario type corresponds to one travel scenario; The background knowledge of the travel scenario also includes: extended knowledge points and logic rules for matching the travel scenario.

7. A travel strategy recommendation method, characterized in that: include: Acquire the travel intention of the current trip, wherein the travel intention of the current trip is determined based on the travel intention identification method according to any one of claims 1 to 6; Determine the current passenger's travel preference information; A travel strategy for the current trip is recommended based on the travel intention of the current trip and the travel preference information of the current passengers.

8. The travel strategy recommendation method according to claim 7, characterized in that: Before executing the step of obtaining the travel intention of the current trip, the method further includes: Obtaining a travel service request, the travel service request being used to request a travel service for the current trip; the travel service request including a ticket request or a travel strategy recommendation request, wherein when the ticket status corresponding to the ticket request satisfies a preset travel intention identification condition, entering the step of obtaining the travel intention of the current trip; The method further comprises: A controlled experimental testing method is used to test the impact of different versions of travel strategies on user behavior in order to optimize the recommendation logic of travel strategies. Among them, different versions of travel strategies include a first version of travel strategy and a second version of travel strategy. The second version of travel strategy is determined based on travel intention and travel preference information, and the first version of travel strategy is the original version of travel strategy.

9. A travel service platform, characterized in that: include: Travel intention recognition system; The travel intention recognition system comprises: A preprocessing module is used to extract and combine features from the current trip information and the current passenger information of the current trip, as well as the historical travel information of the current passenger, to obtain the to-be-matched combined features corresponding to the current trip in multiple travel scenarios; The travel intention recognition model is used to match the combined features to be matched in each travel scenario of the current trip with the feature requirements of each travel scenario, so as to determine the travel scenario that matches the current trip from multiple travel scenarios; wherein the feature requirements of the travel scenario are consistent with the definition and characteristics of the travel scenario; A post-processing module, used to determine the travel intention of the current trip according to the travel scenario matched by the current trip; The travel scenario knowledge base, as the background knowledge base of the travel intention recognition model, records the background knowledge of multiple travel scenarios, and the background knowledge of the travel scenarios at least includes the definition, characteristics and feature requirements of the travel scenarios.

10. The travel service platform according to claim 9, characterized in that: Also includes: Travel strategy recommendation system, used to obtain the travel intention of the current trip and determine the travel preference information of the current passengers; Recommend travel strategies for the current trip based on the travel intention of the current trip and the travel preference information of the current passengers; A travel preference database, recording travel preference information; The travel preference database includes a personal travel preference database that records personal travel preference information of each passenger, or a crowd travel preference database that records crowd travel preference information of each crowd; A test component, used to test the effects of different versions of travel strategies on user behavior using a controlled experimental test method, so as to optimize the recommendation logic of the travel strategy; wherein the different versions of travel strategies include a first version of the travel strategy and a second version of the travel strategy; The second version of the travel strategy is determined based on the travel intention and travel preference information, and the first version of the travel strategy is the original version of the travel strategy; The data collection and analysis system is used to collect and analyze user behavior data through data collection to optimize the performance of the travel strategy recommendation system and the travel strategies recommended by the travel strategy recommendation system.

11. A travel server, characterized in that: It includes a memory and a processor, the memory stores computer-executable instructions, and the processor calls the computer-executable instructions to execute the travel intention identification method as described in any one of claims 1-6, or the travel strategy recommendation method as described in any one of claims 7-8.

12. A computer program product, characterized in that It includes computer-executable instructions, which, when executed, implement the travel intention identification method as described in any one of claims 1-6, or the travel strategy recommendation method as described in any one of claims 7-8.

Images